AT91RM9200
AT91RM9200
AT91RM9200
1
Description The AT91RM9200 is a complete system-on-chip built around the ARM920T ARM Thumb pro-
cessor. It incorporates a rich set of system and application peripherals and standard interfaces
in order to provide a single-chip solution for a wide range of compute-intensive applications
that require maximum functionality at minimum power consumption at lowest cost.
The AT91RM9200 incorporates a high-speed on-chip SRAM workspace, and a low-latency
External Bus Interface (EBI) for seamless connection to whatever configuration of off-chip
memories and memory-mapped peripherals is required by the application. The EBI incorpo-
rates controllers for synchronous DRAM (SDRAM), Burst Flash and Static memories and
features specific circuitry facilitating the interface for SmartMedia, CompactFlash and NAND
Flash.
The Advanced Interrupt Controller (AIC) enhances the interrupt handling performance of the
ARM920T processor by providing multiple vectored, prioritized interrupt sources and reducing
the time taken to transfer to an interrupt handler.
The Peripheral Data Controller (PDC) provides DMA channels for all the serial peripherals,
enabling them to transfer data to or from on- and off-chip memories without processor inter-
vention. This reduces the processor overhead when dealing with transfers of continuous data
streams.The AT91RM9200 benefits from a new generation of PDC which includes dual point-
ers that simplify significantly buffer chaining.
The set of Parallel I/O (PIO) controllers multiplex the peripheral input/output lines with general-
purpose data I/Os for maximum flexibility in device configuration. An input change interrupt,
open drain capability and programmable pull-up resistor is included on each line.
The Power Management Controller (PMC) keeps system power consumption to a minimum by
selectively enabling/disabling the processor and various peripherals under software control. It
uses an enhanced clock generator to provide a selection of clock signals including a slow
clock (32 kHz) to optimize power consumption and performance at all times.
The AT91RM9200 integrates a wide range of standard interfaces including USB 2.0 Full
Speed Host and Device and Ethernet 10/100 Base-T Media Access Controller (MAC), which
provides connection to a extensive range of external peripheral devices and a widely used net-
working layer. In addition, it provides an extensive set of peripherals that operate in
accordance with several industry standards, such as those used in audio, telecom, Flash
Card, infrared and Smart Card applications.
To complete the offer, the AT91RM9200 benefits from the integration of a wide range of debug
features including JTAG-ICE, a dedicated UART debug channel (DBGU) and an embedded
real time trace. This enables the development and debug of all applications, especially those
with real-time constraints.
2 AT91RM9200
1768BATARM08/03
AT91RM9200
Block Diagram Bold arrows ( ) indicate master-to-slave dependency.
TST0-TST1 Reset
and
NRST Test ARM920T Core TSYNC
TCLK
ICE
PIO
JTAGSEL ETM
TDI TPS0 - TPS2
TDO JTAG
Scan Instruction Cache MMU Data Cache TPK0 - TPK15
TMS 16K bytes 16K bytes
TCK
NTRST
BMS
D0-D15
FIQ A0/NBS0
Address
PIO
PIO
XOUT32 A23-A24
A25/CFRNW
Static NCS4/CFCS
DRXD DBGU NCS5/CFCE1
Memory
NCS6/CFCE2
DTXD PDC Controller NCS7
PIO
D16-D31
PIOA/PIOB/PIOC/PIOD
Controller HDMA
Transceiver
DMA FIFO HDPA
DDM HDPB
DDP USB Device
DMA FIFO
ETXCK-ERXCK-EREFCK
ETXEN-ETXER
MCCK
ECRS-ECOL
MCCDA
MCDA0-MCDA3 ERXER-ERXDV
MCI ERX0-ERX3
MCCDB Ethernet MAC 10/100
MCDB0-MCDB3 ETX0-ETX3
PDC EMDC
EMDIO
RXD0 EF100
TXD0 APB TF0
SCK0 USART0 TK0
RTS0 TD0
CTS0 PDC SSC0 RD0
RK0
RXD1 RF0
TXD1 PDC
SCK1 TF1
RTS1 TK1
CTS1 USART1 TD1
DSR1
PIO
SSC1 RD1
PIO
PIO
DTR1 RK1
DCD1 RF1
RI1 PDC PDC
TF2
RXD2 TK2
TXD2 TD2
SCK2 USART2 SSC2 RD2
RTS2 RK2
CTS2 PDC RF2
PDC
RXD3
TXD3 TCLK0
Timer Counter TCLK1
SCK3 USART3
RTS3 TCLK2
CTS3 PDC TIOA0
TC0 TIOB0
TIOA1
NPCS0 TC1 TIOB1
NPCS1 TIOA2
NPCS2 TC2 TIOB2
NPCS3 SPI
MISO TCLK3
MOSI Timer Counter TCLK4
SPCK PDC TCLK5
TIOA3
TC3 TIOB3
TWD TIOA4
TWI TC4 TIOB4
TWCK TIOA5
TC5 TIOB5
3
1768BATARM08/03
Key Features This section presents the key features of each block.
4 AT91RM9200
1768BATARM08/03
AT91RM9200
Reset Controller Two reset input lines (NRST and NTRST) providing, respectively:
Initialization of the User Interface registers (defined in the user interface of each
peripheral) and:
Sample the signals needed at bootup
Compel the processor to fetch the next instruction at address zero.
Initialization of the embedded ICE TAP controller.
5
1768BATARM08/03
Abort generation in case of misalignment
Remap command
Provides remapping of an internal SRAM in place of the boot NVM
6 AT91RM9200
1768BATARM08/03
AT91RM9200
Peripheral Data Generates transfers to/from peripherals such as DBGU, USART, SSC, SPI and MCI
Controller Twenty channels
One Master Clock cycle needed for a transfer from memory to peripheral
Two Master Clock cycles needed for a transfer from peripheral to memory
Advanced Controls the interrupt lines (nIRQ and nFIQ) of an ARM Processor
Interrupt Thirty-two individually maskable and vectored interrupt sources
Controller Source 0 is reserved for the Fast Interrupt Input (FIQ)
Source 1 is reserved for system peripherals (ST, RTC, PMC, DBGU)
Source 2 to Source 31 control thirty embedded peripheral interrupts or external
interrupts
Programmable Edge-triggered or Level-sensitive Internal Sources
Programmable Positive/Negative Edge-triggered or High/Low Level-sensitive
External Sources
8-level Priority Controller
Drives the Normal Interrupt of the processor
Handles priority of the interrupt sources 1 to 31
Higher priority interrupts can be served during service of lower priority interrupt
Vectoring
Optimizes Interrupt Service Routine Branch and Execution
7
1768BATARM08/03
One 32-bit Vector Register per interrupt source
Interrupt Vector Register reads the corresponding current Interrupt Vector
Protect Mode
Easy debugging by preventing automatic operations
Fast Forcing
Permits redirecting any normal interrupt source on the Fast Interrupt of the
processor
General Interrupt Mask
Provides processor synchronization on events without triggering an interrupt
8 AT91RM9200
1768BATARM08/03
AT91RM9200
Two-pin UART
Implemented features are 100% compatible with the standard Atmel USART
Independent receiver and transmitter with a common programmable Baud Rate
Generator
Even, Odd, Mark or Space Parity Generation
Parity, Framing and Overrun Error Detection
Automatic Echo, Local Loopback and Remote Loopback Channel Modes
Interrupt generation
Support for two PDC channels with connection to receiver and transmitter
Debug Communication Channel Support
Offers visibility of COMMRX and COMMTX signals from the ARM Processor
Interrupt generation
Chip ID Registers
Identification of the device revision, sizes of the embedded memories, set of
peripherals
USB Host Port Compliance with Open HCI Rev 1.0 specification
Compliance with USB V2.0 Full-speed and Low-speed Specification
Supports both Low-speed 1.5 Mbps and Full-speed 12 Mbps USB devices
Root hub integrated with two downstream USB ports
Two embedded USB transceivers
Supports power management
Operates as a master on the Memory Controller
USB Device Port USB V2.0 full-speed compliant, 12 Mbits per second
Embedded USB V2.0 full-speed transceiver
Embedded dual-port RAM for endpoints
Suspend/Resume logic
Ping-pong mode (two memory banks) for isochronous and bulk endpoints
Six general-purpose endpoints
Endpoint 0, Endpoint 3: 8 bytes, no ping-pong mode
Endpoint 1, Endpoint 2: 64 bytes, ping-pong mode
Endpoint 4, Endpoint 5: 256 bytes, ping-pong mode
9
1768BATARM08/03
Ethernet MAC Compatibility with IEEE Standard 802.3
10 and 100 Mbits per second data throughput capability
Full- and half-duplex operation
MII or RMII interface to the physical layer
Register interface to address, status and control registers
DMA interface, operating as a master on the Memory Controller
Interrupt generation to signal receive and transmit completion
28-byte transmit and 28-byte receive FIFOs
Automatic pad and CRC generation on transmitted frames
Address checking logic to recognize four 48-bit addresses
Supports promiscuous mode where all valid frames are copied to memory
Supports physical layer management through MDIO interface control of alarm and update
time/calendar data in
10 AT91RM9200
1768BATARM08/03
AT91RM9200
Serial Provides serial synchronous communication links used in audio and telecom applications
Synchronous Contains an independent receiver and transmitter and a common clock divider
Controller Interfaced with two PDC channels (DMA access) to reduce processor overhead
Offers a configurable frame sync and data length
Receiver and transmitter can be programmed to start automatically or on detection of
different event on the frame sync signal
Receiver and transmitter include a data signal, a clock signal and a frame synchronization
signal
11
1768BATARM08/03
Supports two slots
One slot for one MultiMediaCard bus (up to 30 cards) or one SD Memory Card
Support for stream, block and multi-block data read and write
Connection to a Peripheral Data Controller channel
Minimizes processor intervention for large buffer transfers
12 AT91RM9200
1768BATARM08/03
AT91RM9200
Power Supplies The AT91RM9200 has five types of power supply pins:
VDDCORE pins. They power the core, including processor, memories and peripherals;
voltage ranges from 1.65V to 1.95V, 1.8V nominal.
VDDIOM pins. They power the External Bus Interface I/O lines; voltage ranges from 1.65V
to 3.6V, 1.8V, 3V or 3.3V nominal.
VDDIOP pins. They power the Peripheral I/O lines and the USB transceivers; voltage
ranges from 1.65V to 3.6V, 1.8V, 3V or 3.3V nominal.(1)
VDDPLL pins. They power the PLL cells; voltage ranges from 1.65V to 1.95V, 1.8V
nominal.
VDDOSC pin. They power both oscillators; voltage ranges from 1.65V to 1.95V, 1.8V
nominal.
Note: 1. Powering VDDIOP with a voltage lower than 3V prevents any use of the USB Host and
Device Ports. This also affects the operation of the Trace Port.
The double power supplies VDDIOM and VDDIOP are identified in Table 1 on page 14 and
Table 2 on page 16. These supplies enable the user to power the device differently for inter-
facing with memories and for interfacing with peripherals.
Ground pins are common to all power supplies, except VDDPLL and VDDOSC pins. For these
pins, GNDPLL and GNDOSC are provided, respectively.
13
1768BATARM08/03
208-lead PQFP Package Pinout
Table 1. AT91RM9200 Pinout for 208-lead PQFP Package
Pin Pin Pin Pin
Number Signal Name Number Signal Name Number Signal Name Number Signal Name
1 PC24 37 VDDPLL 73 PA27 109 TMS
14 AT91RM9200
1768BATARM08/03
AT91RM9200
157 104
208
53
1 52
15
1768BATARM08/03
256-ball BGA Package Pinout
Table 2. AT91RM9200 Pinout for 256-ball BGA Package
Pin Signal Name Pin Signal Name Pin Signal Name Pin Signal Name
16 AT91RM9200
1768BATARM08/03
AT91RM9200
NWR1/NBS1/
L1 GND N2 A5 P13 D15 T7
CFIOR
NWR3/NBS3/
M1 VDDIOM N14 PC31 R8 U2 GND
CFIOW
17
1768BATARM08/03
Mechanical Figure 3 on page 18 shows the orientation of the 256-ball BGA Package.
Overview of the A detailed mechanical description is given in the section Mechanical Characteristics.
256-ball BGA
Package Figure 3. 256-ball BGA Pinout (Top View)
18 AT91RM9200
1768BATARM08/03
AT91RM9200
19
1768BATARM08/03
PIO Controller B Multiplexing
Table 4. Multiplexing on PIO Controller B
PIO Controller B Application Usage
Reset
I/O Line Peripheral A Peripheral B State Function Comments
PB0 TF0 RTS3 I/O
PB1 TK0 CTS3 I/O
PB2 TD0 SCK3 I/O
PB3 RD0 MCDA1 I/O
PB4 RK0 MCDA2 I/O
PB5 RF0 MCDA3 I/O
PB6 TF1 TIOA3 I/O
PB7 TK1 TIOB3 I/O
PB8 TD1 TIOA4 I/O
PB9 RD1 TIOB4 I/O
PB10 RK1 TIOA5 I/O
PB11 RF1 TIOB5 I/O
PB12 TF2 ETX2 I/O
PB13 TK2 ETX3 I/O
PB14 TD2 ETXER I/O
PB15 RD2 ERX2 I/O
PB16 RK2 ERX3 I/O
PB17 RF2 ERXDV I/O
PB18 RI1 ECOL I/O
PB19 DTR1 ERXCK I/O
PB20 TXD1 I/O
PB21 RXD1 I/O
PB22 SCK1 I/O
PB23 DCD1 I/O
PB24 CTS1 I/O
PB25 DSR1 EF100 I/O
PB26 RTS1 I/O
PB27 PCK0 I/O
PB28 FIQ I/O
PB29 IRQ0 I/O
20 AT91RM9200
1768BATARM08/03
AT91RM9200
21
1768BATARM08/03
PIO Controller D Multiplexing
The PIO Controller D multiplexes pure output signals on peripheral A connections, in particular from the EMAC RMII inter-
face and the ETM Port on the peripheral B connections.
The PIO Controller D is available only in the 256-ball BGA package option of the AT91RM9200.
22 AT91RM9200
1768BATARM08/03
AT91RM9200
23
1768BATARM08/03
Table 7. Pin Description List (Continued)
Active
Pin Name Function Type Level Comments
Memory Controller
BMS Boot Mode Select Input
Debug Unit
DRXD Debug Receive Data Input Debug Receive Data
DTXD Debug Transmit Data Output Debug Transmit Data
AIC
IRQ0 - IRQ6 External Interrupt Inputs Input
FIQ Fast Interrupt Input Input
PIO
PA0 - PA31 Parallel IO Controller A I/O Pulled-up input at reset
PB0 - PB29 Parallel IO Controller B I/O Pulled-up input at reset
PC0 - PC31 Parallel IO Controller C I/O Pulled-up input at reset
PD0 - PD27 Parallel IO Controller D I/O Pulled-up input at reset
EBI
D0 - D15 Data Bus I/O Pulled-up input at reset
D16 - D31 Data Bus I/O Pulled-up input at reset
A0 - A25 Address Bus Output 0 at reset
SMC
NCS0 - NCS7 Chip Select Lines Output Low 1 at reset
NWR0 - NWR3 Write Signal Output Low 1 at reset
NOE Output Enable Output Low 1 at reset
NRD Read Signal Output Low 1 at reset
NUB Upper Byte Select Output Low 1 at reset
NLB Lower Byte Select Output Low 1 at reset
NWE Write Enable Output Low 1 at reset
NBS0 - NBS3 Byte Mask Signal Output Low 1 at reset
EBI for CompactFlash Support
CFCE1 - CFCE2 CompactFlash Chip Enable Output Low
CFOE CompactFlash Output Enable Output Low
CFWE CompactFlash Write Enable Output Low
CFIOR CompactFlash IO Read Output Low
CFIOW CompactFlash IO Write Output Low
24 AT91RM9200
1768BATARM08/03
AT91RM9200
25
1768BATARM08/03
Table 7. Pin Description List (Continued)
Active
Pin Name Function Type Level Comments
DTR1 Data Terminal Ready Output
DCD1 Data Carrier Detect Input
RI1 Ring Indicator Input
USB Device Port
DDM USB Device Port Data - Analog
DDP USB Device Port Data + Analog
USB Host Port
HDMA USB Host Port A Data - Analog
HDPA USB Host Port A Data + Analog
HDMB USB Host Port B Data - Analog
HDPB USB Host Port B Data + Analog
Ethernet MAC
EREFCK Reference Clock Input RMII only
ETXCK Transmit Clock Input MII only
ERXCK Receive Clock Input MII only
ETXEN Transmit Enable Output
ETX0 - ETX3 Transmit Data Output ETX0 - ETX1 only in RMII
ETXER Transmit Coding Error Output MII only
ERXDV Receive Data Valid Input MII only
ECRSDV Carrier Sense and Data Valid Input RMII only
ERX0 - ERX3 Receive Data Input ERX0 - ERX1 only in RMII
ERXER Receive Error Input
ECRS Carrier Sense Input MII only
ECOL Collision Detected Input MII only
EMDC Management Data Clock Output
EMDIO Management Data Input/Output I/O
EF100 Force 100 Mbits/sec. Output High RMII only
Synchronous Serial Controller
TD0 - TD2 Transmit Data Output
RD0 - RD2 Receive Data Input
TK0 - TK2 Transmit Clock I/O
RK0 - RK2 Receive Clock I/O
TF0 - TF2 Transmit Frame Sync I/O
RF0 - RF2 Receive Frame Sync I/O
26 AT91RM9200
1768BATARM08/03
AT91RM9200
27
1768BATARM08/03
Peripheral The AT91RM9200 embeds a wide range of peripherals. Table 8 defines the peripheral identifi-
Identifiers ers of the AT91RM9200. A peripheral identifier is required for the control of the peripheral
interrupt with the Advanced Interrupt Controller and for the control of the peripheral clock with
the Power Management Controller.
1 SYSIRQ
6 US0 USART 0
7 US1 USART 1
8 US2 USART 2
9 US3 USART 3
17 TC0 Timer/Counter 0
18 TC1 Timer/Counter 1
19 TC2 Timer/Counter 2
20 TC3 Timer/Counter 3
21 TC4 Timer/Counter 4
22 TC5 Timer/Counter 5
28 AT91RM9200
1768BATARM08/03
AT91RM9200
System Interrupt The System Interrupt is the wired-OR of the interrupt signals coming from:
the Memory Controller
the Debug Unit
the System Timer
the Real-Time Clock
the Power Management Controller
The clock of these peripherals cannot be controlled and the Peripheral ID 1 can only be used
within the Advanced Interrupt Controller.
External Interrupts All external interrupt signals, i.e., the Fast Interrupt signal FIQ or the Interrupt signals IRQ0 to
IRQ6, use a dedicated Peripheral ID. However, there is no clock control associated with these
peripheral IDs.
29
1768BATARM08/03
Product Memory Mapping
A first level of address decoding is performed by the Memory Controller, i.e., by the implemen-
tation of the Advanced System Bus (ASB) with additional features.
Decoding splits the 4G bytes of address space into 16 areas of 256M bytes. The areas 1 to 8
are directed to the EBI that associates these areas to the external chip selects NC0 to NCS7.
The area 0 is reserved for the addressing of the internal memories, and a second level of
decoding provides 1M bytes of internal memory area. The area 15 is reserved for the peripher-
als and provides access to the Advanced Peripheral Bus (APB).
Other areas are unused and performing an access within them provides an abort to the master
requesting such an access.
0xEFFF FFFF
0xF000 0000
256M Bytes Peripherals
0xFFFF FFFF
30 AT91RM9200
1768BATARM08/03
AT91RM9200
Internal RAM The AT91RM9200 integrates a high-speed, 16-Kbyte internal SRAM. After reset and until the
Remap Command is performed, the SRAM is only accessible at address 0x20 0000. After
Remap, the SRAM is also available at address 0x0.
Internal ROM The AT91RM9200 integrates a 128-Kbyte Internal ROM. At any time, the ROM is mapped at
address 0x10 0000. It is also accessible at address 0x0 after reset and before the Remap
Command if the BMS is tied high during reset.
USB Host Port The AT91RM9200 integrates a USB Host Port Open Host Controller Interface (OHCI). The
registers of this interface are directly accessible on the ASB Bus and are mapped like a stan-
dard internal memory at address 0x30 0000.
0x0FFF FFFF
31
1768BATARM08/03
Peripheral Mapping
0xFFFF F000
0xFFFF F1FF
0xFFFF F200
0xFFFF F3FF
0xFFFF F400
0xFFFF F5FF
0xFFFF F600
0xFFFF F7FF
0xFFFF F800
0xFFFF F9FF
0xFFFF FA00
0xFFFF FBFF
0xFFFF FC00
PMC Power Management Controller 256 bytes/64 registers
0xFFFF FCFF
0xFFFF FD00
ST System Timer 256 bytes/64 registers
0xFFFF FDFF
0xFFFF FE00
RTC Real-time Clock 256 bytes/64 registers
0xFFFF FEFF
0xFFFF FF00
MC Memory Controller 256 bytes/64 registers
0xFFFF FFFF
32 AT91RM9200
1768BATARM08/03
AT91RM9200
User Peripherals The User Peripherals are mapped in the upper 256M bytes of the address space, between the
Mapping addresses 0xFFFA 0000and 0xFFFE 3FFF. Each peripheral has a 16-Kbyte address space.
Reserved
0xFFF9 FFFF
0xFFFA 0000
TC0, TC1, TC2 Timer/Counter 0, 1 and 2 16K Bytes
0xFFFA 3FFF
0xFFFA 4000
TC3, TC4, TC5 Timer/Counter 3, 4 and 5 16K Bytes
0xFFFA 7FFF
0xFFFA 8000
Reserved
0xFFFA FFFF
0xFFFB 0000
UDP USB Device Port 16K Bytes
0xFFFB 3FFF
0xFFFB 4000
MCI Multimedia Card Interface 16K Bytes
0xFFFB 7FFF
0xFFFB 8000
TWI Two-Wire Interface 16K Bytes
0xFFFB BFFF
0xFFFB C000
EMAC Ethernet MAC 16K Bytes
0xFFFB FFFF
0xFFFC 0000
USART0 Universal Synchronous Asynchronous 16K Bytes
Receiver Transmitter 0
0xFFFC 3FFF
0xFFFC 4000
USART1 Universal Synchronous Asynchronous 16K Bytes
Receiver Transmitter 1
0xFFFC 7FFF
0xFFFC 8000
USART2 Universal Synchronous Asynchronous 16K Bytes
Receiver Transmitter 2
0xFFFC BFFF
0xFFFC C000
USART3 Universal Synchronous Asynchronous 16K Bytes
Receiver Transmitter 3
0xFFFC FFFF
0xFFFD 0000
SSC0 Serial Synchronous Controller 0 16K Bytes
0xFFFD 3FFF
0xFFFD 4000
SSC1 Serial Synchronous Controller 1 16K Bytes
0xFFFD 7FFF
0xFFFD 8000
SSC2 Serial Synchronous Controller 2 16K Bytes
0xFFFD BFFF
0xFFFD C000
Reserved
0xFFFD FFFF
0xFFFE 0000
SPI Serial Peripheral Interface 16K Bytes
0xFFFE 3FFF
0xFFFE 4000
Reserved
0xFFFE FFFF
33
1768BATARM08/03
Peripheral Implementation
USART The USART describes features allowing management of the Modem Signals DTR, DSR, DCD
and RI. For details, see Modem Mode on page 422.
In the AT91RM9200, only the USART1 implements these signals, named DTR1, DSR1, DCD1
and RI1.
The USART0, USART2 and USART3 do not implement all the modem signals. Only RTS and
CTS (RTS0 and CTS0, RTS2 and CTS2, RTS3 and CTS3, respectively) are implemented in
these USARTs for other features.
Thus, programming the USART0, USART2 or the USART3 in Modem Mode may lead to
unpredictable results. In these USARTs, the commands relating to the Modem Mode have no
effect and the status bits relating the status of the modem signals are never activated.
Timer Counter The Timer Counter 0 to 5 are described with five generic clock inputs, TIMER_CLOCK1 to
TIMER_CLOCK5. In the AT91RM9200, these clock inputs are connected to the Master Clock
(MCK), to the Slow Clock (SLCK) and to divisions of the Master Clock. For details, see Clock
Control on page 476.
Table 2 gives the correspondence between the Timer Counter clock inputs and clocks in the
AT91RM9200. Each Timer Counter 0 to 5 displays the same configuration.
34 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The ARM920T cached processor is a member of the ARM9 Thumb family of high-
performance 32-bit system-on-a-chip processors. It provides a complete high perfor-
mance CPU subsystem including:
ARM9TDMI RISC integer CPU
16-Kbyte instruction and 16-Kbyte data caches
Instruction and data memory management units (MMUs)
Write buffer
AMBA (Advanced Microprocessor Bus Architecture) bus interface
Embedded Trace Macrocell (ETM) interface
The ARM9TDMI core within the ARM920T executes both the 32-bit ARM and 16-bit
Thumb instruction sets. The ARM9TDMI processor is a Harvard architecture device,
implementing a five-stage pipeline consisting of Fetch, Decode, Execute, Memory and
Write stages.
The ARM920T processor incorporates two coprocessors:
CP14 - Controls software access to the debug communication channel
CP15 - System Control Processor, providing 16 additional registers that are used to
configure and control the caches, the MMU, protection system, clocking mode and
other system options
The main features of the ARM920T processor are:
ARM9TDMI-based, ARM Architecture v4T
Two Instruction Sets
ARM High-performance 32-bit Instruction Set
Thumb High Code Density 16-bit Instruction Set
5-Stage Pipeline Architecture
Instruction Fetch (F)
Instruction Decode (D)
Execute (E)
Data Memory Access (M)
Register Write (W)
16-Kbyte Data Cache, 16-Kbyte Instruction Cache
Virtually-addressed 64-way Associative Cache
8 Words per Line
Write-though and Write-back Operation
Pseudo-random or Round-robin Replacement
Low-power CAM RAM Implementation
Write Buffer
16-word Data Buffer
4-address Address Buffer
Software Control Drain
Standard ARMv4 Memory Management Unit (MMU)
Access Permission for Sections
35
1768BATARM08/03
Access Permission for Large Pages and Small Pages Can be Specified
Separately for Each Quarter of the Pages
16 Embedded Domains
64-entry Instruction TLB and 64-entry Data TLB
8-, 16-, 32-bit Data Bus for Instructions and Data
Block Diagram
Figure 8. ARM920T Internal Functional Block Diagram
ARM920T
Instruction Instruction
Cache MMU Instruction
Physical
Address
Bus
Instruction
Modified
Virtual
R13 Address
Bus
Instruction Instruction
Virtual Bus
Address
Bus
ICE
Interface
Bus Memory
ICE ARM9TDMI CP15 Interface Controller
Data
Virtual Data
Address Bus
Bus Write
Buffer
Data
R13 Modified
Virtual
Address Data
Bus Physical
Address
Bus
36 AT91RM9200
1768BATARM08/03
AT91RM9200
ARM9TDMI Processor
Instruction Type Instructions are either 32 bits (in ARM state) or 16 bits (in Thumb state).
Data Types ARM9TDMI supports byte (8-bit), half-word (16-bit) and word (32-bit) data types. Words
must be aligned to four-byte boundaries and half-words to two-byte boundaries.
Unaligned data access behavior depends on which instruction is used in a particular
location.
ARM9TDMI Operating The ARM9TDMI, based on ARM architecture v4T, supports seven processor modes:
Modes User: Standard ARM program execution state
FIQ: Designed to support high-speed data transfer or channel processes
IRQ: Used for general-purpose interrupt handling
Supervisor: Protected mode for the operating system
Abort mode: Implements virtual memory and/or memory protection
System: A privileged user mode for the operating system
Undefined: Supports software emulation of hardware coprocessors
Mode changes may be made under software control, or may be brought about by exter-
nal interrupts or exception processing. Most application programs will execute in User
Mode. The non-user modes, known as privileged modes, are entered in order to service
interrupts or exceptions or to access protected resources.
37
1768BATARM08/03
ARM9TDMI Registers The ARM9TDMI processor core consists of a 32-bit datapath and associated control
logic. That datapath contains 31 general-purpose registers, coupled to a full shifter,
Arithmetic Logic Unit and multiplier.
At any one time, 16 registers are visible to the user. The remainder are synonyms used
to speed up exception processing.
Register 15 is the Program Counter (PC) and can be used in all instructions to reference
data relative to the current instruction.
R14 holds the return address after a subroutine call.
R13 is used (by software convention) as a stack pointer.
Mode-specific banked
registers
Registers R0 to R7 are unbanked registers, thus each of them refers to the same 32-bit
physical register in all processor modes. They are general-purpose registers, with no
special uses managed by the architecture, and can be used wherever an instruction
allows a general-purpose register to be specified.
38 AT91RM9200
1768BATARM08/03
AT91RM9200
Registers R8 to R14 are banked registers. This means that each of them depends of the
current processor mode.
For further details, see the ARM Architecture Reference Manual, Rev. DDI0100E.
Modes and Exception All exceptions have banked registers for R14 and R13.
Handling
After an exception, R14 holds the return address for exception processing. This address
is used both to return after the exception is processed and to address the instruction that
caused the exception.
R13 is banked across exception modes to provide each exception handler with a private
stack pointer.
The fast interrupt mode also banks registers 8 to 12 so that interrupt processing can
begin without the need to save these registers.
A seventh processing mode, System Mode, does not have any banked registers. It uses
the User Mode registers. System Mode runs tasks that require a privileged processor
mode and allows them to invoke all classes of exceptions.
Status Registers All other processor states are held in status registers. The current operating processor
status is in the Current Program Status Register (CPSR). The CPSR holds:
four ALU flags (Negative, Zero, Carry, and Overflow),
two interrupt disable bits (one for each type of interrupt),
one bit to indicate ARM or Thumb execution
five bits to encode the current processor mode
All five exception modes also have a Saved Program Status Register (SPSR) which
holds the CPSR of the task immediately before the exception occurred.
Exception Types The ARM9TDMI supports five types of exceptions and a privileged processing mode for
each type. The types of exceptions are:
fast interrupt (FIQ)
normal interrupt (IRQ)
memory aborts (used to implement memory protection or virtual memory)
attempted execution of an undefined instruction
software interrupt (SWIs)
Exceptions are generated by internal and external sources.
More than one exception can occur at the same time.
When an exception occurs, the banked version of R14 and the SPSR for the exception
mode are used to save the state.
To return after handling the exception, the SPSR is moved to the CPSR and R14 is
moved to the PC. This can be done in two ways:
use of a data-processing instruction with the S-bit set, and the PC as the destination
use of the Load Multiple with Restore CPSR instruction (LDM)
39
1768BATARM08/03
ARM Instruction Set The ARM instruction set is divided into:
Overview Branch instructions
Data processing instructions
Status register transfer instructions
Load and Store instructions
Coprocessor instructions
Exception-generating instructions
ARM instructions can be executed conditionally. Every instruction contains a 4-bit condi-
tion code field (bits[31:28]).
For further details, see the ARM920T Technical Reference Manual, Rev. DDI0151C.
Table 10 gives the ARM instruction mnemonic list.
LDRBT Load Register Byte with Translation STRT Store Register with Translation
40 AT91RM9200
1768BATARM08/03
AT91RM9200
Thumb Instruction Set The Thumb instruction set is a re-encoded subset of the ARM instruction set.
Overview The Thumb instruction set is divided into:
Branch instructions
Data processing instructions
Load and Store instructions
Load and Store multiple instructions
Exception-generating instruction
In Thumb mode, eight general-purpose registers are available, R0 to R7, that are the
same physical registers as R0 to R7 when executing ARM instructions. Some Thumb
instructions also access the Program Counter (ARM Register 15), the Link Register
(ARM Register 14) and the Stack Pointer (ARM Register 13). Further instructions allow
limited access to the ARM register 8 to 15.
For further details, see the ARM920T Technical Reference Manual, Rev. DDI0151C.
Table 11 gives the Thumb instruction mnemonic list.
MUL Multiply
41
1768BATARM08/03
CP15 Coprocessor
Coprocessor 15, or System Control Coprocessor CP15, is used when special features
are used with the ARM9TDMI such as:
On-chip Memory Management Unit (MMU)
Instruction and/or Data Cache
Write buffer
To control these features, CP15 provides 16 additional registers. See Table 12.
42 AT91RM9200
1768BATARM08/03
AT91RM9200
31 30 29 28 27 26 25 24
Cond 1 1 1 0
23 22 21 20 19 18 17 16
opcode_1 L CRn
15 14 13 12 11 10 9 8
Rd 1 1 1 1
7 6 5 4 3 2 1 0
opcode_2 1 CRm
43
1768BATARM08/03
Memory Management Unit (MMU)
The ARM920T processor implements an enhanced ARM architecture v4 MMU to pro-
vide translation and access permission checks for the instruction and data address ports
of the ARM9TDMI core. The MMU is controlled from a single set of two-level page
tables stored in the main memory, providing a single address and translation protection
scheme. Independently, instruction and data TLBs in the MMU can be locked and
flushed.
Domain A domain is a collection of sections and pages. The ARM920T supports 16 domains.
Access to the domains is controlled by the Domain Access Control register. For details,
refer to CP15 Register 3, Domain Access Control Register on page 52.
MMU Faults The MMU generates alignment faults, translation faults, domain faults and permission
faults. Alignment fault checking is not affected by whether the MMU is enabled or not.
The access controls of the MMU detect the conditions that produce these faults. If the
fault is a result of memory access, the MMU aborts the access and signals the fault to
the CPU core.The MMU stores the status and address fault in the FSR and FAR regis-
ters (only for faults generated by data access).
The MMU does not store fault information about faults generated by an instruction fetch.
The memory system can abort during line fetches, memory accesses and translation
table access.
44 AT91RM9200
1768BATARM08/03
AT91RM9200
Instruction Cache The ARM920T includes a 16-Kbyte Instruction Cache (ICache). The ICache has 512
(ICache) lines of 32 bytes, arranged as a 64-way set associative cache.
Instruction access is subject to MMU permission and translation checks.
If the ICache is enabled with the MMU disabled, all instructions fetched as threats are
cachable. No protection checks are made and the physical address is flat-mapped to the
modified virtual address.
When the ICache is disabled, the cache contents are ignored and all instruction fetches
appear on the AMBA bus.
On reset, the ICache entries are invalidated and the ICache is disabled. For best perfor-
mance, ICache should be enabled as soon as possible after reset.
The ICache is enabled by writing 1 to I bit of the CP15 Register 1 and disabled by writing
0 to this bit. For more details, see CP15 Register 1, Control on page 49.
The ICache is organized as eight segments, each containing 64 lines with each line
made up of 8 words.The position of the line within the segment is called the index and is
numbered from 0 to 63.
A line in the cache is identified by the index and segment. The index is independent of
the MVA (Modified Virtual Address), and the segment is the bit[7:5] of the MVA.
Data Cache (DCache) The ARM920T includes a 16-Kbyte data cache (DCache). The DCache has 512 lines of
and Write Buffer 32 bytes, arranged as a 64-way set associative cache, and uses MVAs translated by
CP15 Register 13 from the ARM9DTMI core.
DCache The DCache is organized as eight segments, each containing 64 lines with each line
made up of eight words.The position of the line within the segment is called the index
and is a number from 0 to 63.
The Write Buffer can hold up to 16 words of data and four separate addresses.
DCache and Write Buffer operations are closely connected as their configuration is set
in each section by the page descriptor in the MMU translation table.
All data accesses are subject to MMU permission and translation checks. Data
accesses aborted by the MMU cannot cause linefill or data access via the AMBA ASB
interface.
Write-though Operation When a cache hit occurs for a data access, the cache line that contains the data is
updated to contains its value. The new data is also immediately written to the main
memory.
Write-back Operation When a cache hit occurs for a data access, the cache line is marked as dirty, meaning
that its contents are not up-to-date with those in the main memory.
45
1768BATARM08/03
Write Buffer The ARM920T incorporates a 16-entry write buffer to avoid stalling the processor when
writes to external memory are performed. When a store occurs, its data, address and
other details are written to the write buffer at high speed. The write buffer then com-
pletes the store at the main memory speed (typically slower than the ARM speed). In
parallel, the ARM9TDMI processor can execute further instructions at full speed.
Physical Address Tag RAM The ARM920T implements Physical Address Tag RAM (PA TAG RAM) to perform write-
(PA TAG RAM) backs from the data cache. The physical address of all the lines held in the data cache is
stored in the PA TAG memory, removing the need for address translation when evicting
a line from the cache.
When a line is written into the data cache, the physical address TAG is written into the
PA TAG RAM. If this line has to be written back to the main memory, the PA TAG RAM
is read and the physical address is used by the AMBA ASB interface to perform the
write-back.
For a 16-Kbyte DCache, the PA TAG RAM is organized by eight segments with:
64 rows per segments
26 bits per rows
46 AT91RM9200
1768BATARM08/03
AT91RM9200
ID Code
The ID code register is accessed by reading the register 0 with the opcode_2 field set to 0.
The contents of the ID code is shown below:
31 30 29 28 27 26 25 24
imp
23 22 21 20 19 18 17 16
SRev archi
15 14 13 12 11 10 9 8
PNumber
7 6 5 4 3 2 1 0
Layout Rev
LayoutRev[3:0]: Revision
Contains the processor revision number
PNumber[15:4]: Processor Part Number
0x920 value for ARM920T processor.
archi[19:16]: Architecture
Details the implementor architecture code.
0x2 value means ARMv4T architecture.
SRev[23:20]: Specification Revision Number
0x1 value; specification revision number used to distinguished two variants of the same primary part.
imp[31:24]: Implementor Code
0x41 (= A); means ARM Ltd.
47
1768BATARM08/03
Cache Type
The Cache Type register is accessed by reading the register 0 with the opcode_2 field set to 1.
The Cache Type register contains information about the size and architecture of the caches.
31 30 29 28 27 26 25 24
0 0 0 ctype S
23 22 21 20 19 18 17 16
DSize
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
ISize
For details on bits DSize and ISize, refer to the ARM920T Technical Reference Manual, Rev. DDI0151C.
48 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
- - - - - - - -
15 14 13 12 11 10 9 8
- RR V I 0 0 R S
7 6 5 4 3 2 1 0
B 1 1 1 1 C A M
49
1768BATARM08/03
Clocking Mode[31:30] (iA and nF bits)
iA nF Clocking mode
0 0 Fast Bus
0 1 Synchronous
1 0 Reserved
1 1 Asynchronous
50 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
Pointer
15 14 13 12 11 10 9 8
Pointer - - - - - -
7 6 5 4 3 2 1 0
- - - - - - - -
Pointer[31:14]
Points to the first-level translation table base. Read returns the currently active first-level translation table. Write sets the
pointer to the first-level table to the written value.
The non-defined bits should be zero when written and are unpredictable when read.
51
1768BATARM08/03
CP15 Register 3, Domain Access Control Register
Access: Read/write
The CP 15 Register 3, or Domain Access Control Register, defines the domains access permission.
MMU accesses are priory controlled through the use of 16 domains.
Each field of Register 3 is associated with one domain.
31 30 29 28 27 26 25 24
D15 D14 D13 D12
23 22 21 20 19 18 17 16
D11 D10 D9 D8
D
15 14 13 12 11 10 9 8
D7 D6 D5 D4
7 6 5 4 3 2 1 0
D3 D2 D1 D0
52 AT91RM9200
1768BATARM08/03
AT91RM9200
31 30 29 28 27 26 25 24
- - - - - - - -
23 22 21 20 19 18 17 16
- - - - - - - -
D
15 14 13 12 11 10 9 8
- - - - - - - -
7 6 5 4 3 2 1 0
Domain Status
The non-defined bits should be zero when written and are unpredictable when read.
53
1768BATARM08/03
CP15 Register 6, Fault Address Register
Access: Read/write
The CP 15 Register 6, or Fault Address Register (FAR), contains the MVA (Modified Virtual Address) of the access being
attempted when the last fault occurred. The FAR is only updated for data faults, not for prefetch faults.
The ability to write to the FAR is provided to allow a debugger to restore a previous state.
31 30 29 28 27 26 25 24
FAR
23 22 21 20 19 18 17 16
FAR
D
15 14 13 12 11 10 9 8
FAR
7 6 5 4 3 2 1 0
FAR
54 AT91RM9200
1768BATARM08/03
AT91RM9200
Functions Details
Wait for interrupt
Stops execution in low-power state until an interrupt occurs.
Invalidate
The cache line (or lines) is marked as invalid, so no cache hits occur in that line until it is re-allocated to an address.
Clean
Applies to write-back data caches. If the cache line contains stored data that has not yet been written out to the main mem-
ory, it is written to main memory immediately.
Drain write buffer
Stops the execution until all data in the write buffer has been stored in the main memory.
Prefetch
The memory cache line at the specified virtual address is loaded into the cache.
55
1768BATARM08/03
The operation carried out on a single cache line identifies the line using the data transferred in the MCR instruction.
The data is interpreted as using one of the two formats:
MVA format
index format
Below are the details of CP15 Register 7, or Cache Function Register, in MVA format.
31 30 29 28 27 26 25 24
mva
23 22 21 20 19 18 17 16
mva
15 14 13 12 11 10 9 8
mva
7 6 5 4 3 2 1 0
mva - - - - -
Below the details of CP15 Register 7, or Cache Function Register, in Index format:
31 30 29 28 27 26 25 24
index - -
23 22 21 20 19 18 17 16
- - - - - - - -
15 14 13 12 11 10 9 8
- - - - - - - -
7 6 5 4 3 2 1 0
set - - - - -
index[31:26]: Line
Determines the cache line.
set[7:5]: Segment
Determines the cache segment.
The non-defined bits should be zero when written and are unpredictable when read.
56 AT91RM9200
1768BATARM08/03
AT91RM9200
Below are details of the CP15 Register 8 for TLB operation on MVA format and one single entry.
31 30 29 28 27 26 25 24
mva
23 22 21 20 19 18 17 16
mva
15 14 13 12 11 10 9 8
mva - -
7 6 5 4 3 2 1 0
- - - - - - - -
57
1768BATARM08/03
CP15 Register 9, Cache Lockdown Register
Access: Read/write
The CP15 Register 9, or Cache Lockdown Register, is 0x0 on reset. The Cache Lockdown Register allows software to con-
trol which cache line in the ICache or DCache is loaded for a linefill. It prevents lines in the ICache or DCache from being
evicted during a linefill, locking them into the cache.
Reading from the CP15 Register 9 returns the value of the Cache Lockdown Register that is the base pointer for all cache
segments.
Only the bits[31:26] are returned; others are unpredictable.
Writing to the CP15 Register 9 updates the Cache Lockdown Register with both the base and the current victim pointers for
all cache segments.
31 30 29 28 27 26 25 24
index - -
23 22 21 20 19 18 17 16
- - - - - - - -
15 14 13 12 11 10 9 8
- - - - - - - -
7 6 5 4 3 2 1 0
- - - - - - - -
The non-defined bits should be zero when written and are unpredictable when read.
58 AT91RM9200
1768BATARM08/03
AT91RM9200
31 30 29 28 27 26 25 24
Base
23 22 21 20 19 18 17 16
Victim - - - -
D
15 14 13 12 11 10 9 8
- - - - - - - -
7 6 5 4 3 2 1 0
- - - - - - - P
Base[31:26]: Base
The TLB replacement strategy only uses the TLB entries numbered from base to 63. The Victim field provided is in that
range.
Victim[25:20]: Victim Counter
Specifies the TLB entry (line) being overwritten.
P[0]: Preserved
If 0, the TLB entry can be invalidated.
If 1, the TLB entry is protected. It cannot be invalidated during the Invalidate All instruction. Refer to CP15 Register 8, TLB
Operations Register on page 57.
The non-defined bits should be zero when written and are unpredictable when read.
59
1768BATARM08/03
CP15 Register 13, FCSE PID Register
Access: Read/write
The CP15 Register 13, or Fast Context Switch Extension (FCSE) Process Identifier (PID) Register, is set to 0x0 on reset.
Reading from CP15 Register 13 returns the FCSE PID value.
Writing to CP15 Register 13 sets the FCSE PID.
The FCSE PID sets the mapping between the ARM9TDMI and the MMU of the cache memories.
The addresses issued by the ARM9TDMI are in the range of 0 to 32 Mbytes and are translated via the FCSE PID.
31 30 29 28 27 26 25 24
FCSEPID -
23 22 21 20 19 18 17 16
- - - - - - - -
D
15 14 13 12 11 10 9 8
- - - - - - - -
7 6 5 4 3 2 1 0
- - - - - - - -
The non-defined bits should be zero when written and are unpredictable when read.
60 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The AT91RM9200 features a number of complementary debug and test capabilities. A com-
mon JTAG/ICE (In-Circuit Emulator) port is used for standard debugging functions such as
downloading code and single-stepping through programs. An ETM (Embedded Trace Macro-
cell) provides more sophisticated debug features such as address and data comparators, half-
rate clock mode, counters, sequencer and FIFO. The Debug Unit provides a two-pin UART
that can be used to upload an application into internal SRAM. It manages the interrupt han-
dling of the internal COMMTX and COMMRX signals that trace the activity of the Debug
Communication Channel.
A set of dedicated debug and test input/output pins give direct access to these capabilities
from a PC-based test environment.
Features of Debug and Test Features are:
Integrated Embedded In-Circuit-Emulator
Debug Unit
Two-pin UART
Debug Communication Channel
Chip ID Register
Embedded Trace Macrocell: ETM9 Rev2a
Medium Level Implementation
Half-rate Clock Mode
Four Pairs of Address Comparators
Two Data Comparators
Eight Memory Map Decoder Inputs
Two Counters
One Sequencer
One 18-byte FIFO
IEEE1149.1 JTAG Boundary Scan on all Digital Pins
61
1768BATARM08/03
Block Diagram Figure 9. AT91RM9200 Debug and Test Block Diagram
TMS
TCK
TDI
NTRST
ICE/JTAG JTAGSEL
Boundary TAP
Port
TDO
TPK0-TPK15
TPS0-TPS2
PIO
ARM9TDMI ICE ETM
TSYNC
2 TCLK
ARM920T
DTXD
PDC DBGU
DRXD
TST0-TST1
Reset
and
Test NRST
62 AT91RM9200
1768BATARM08/03
AT91RM9200
Application
Examples
Debug Figure 10 on page 63 shows a complete debug environment example. The ICE/JTAG inter-
Environment face is used for standard debugging functions such as downloading code and single-stepping
through the program. The Trace Port interface is used for tracing information. A software
debugger running on a personal computer provides the user interface for configuring a Trace
Port interface utilizing the ICE/JTAG interface.
Host Debugger
ICE/JTAG Trace
Connector Connector
Test Figure 11 below shows a test environment example. Test vectors are sent and interpreted by
the tester. In this example, the board under test is designed using many JTAG compliant
Environment
devices. These devices can be connected together to form a single scan chain.
63
1768BATARM08/03
Figure 11. AT91RM9200-based Application IEEE1149.1 Test Environment Example
JTAG
Interface
ICE/JTAG
Connector Chip n Chip 2
AT91RM920 Chip 1
64 AT91RM9200
1768BATARM08/03
AT91RM9200
Functional
Description
Test Mode Pins Two dedicated pins (TST1, TST0) are used to define the test mode of the device. The user
must make sure that these pins are both tied at low level to ensure normal operating condi-
tions. Other values associated to these pins are manufacturing test reserved.
Embedded In- The ARM9TDMI Embedded In-Circuit Emulator is supported via the ICE/JTAG port. It is con-
Circuit Emulator nected to a host computer via an ICE interface. Debug support is implemented using an
ARM9TDMI core embedded within the ARM920T. The internal state of the ARM920T is exam-
ined through an ICE/JTAG port which allows instructions to be serially inserted into the
pipeline of the core without using the external data bus. Therefore, when in debug state, a
store-multiple (STM) can be inserted into the instruction pipeline. This exports the contents of
the ARM9TDMI registers. This data can be serially shifted out without affecting the rest of the
system.
There are six scan chains inside the ARM920T processor which support testing, debugging,
and programming of the Embedded ICE. The scan chains are controlled by the ICE/JTAG
port.
Embedded ICE mode is selected when JTAGSEL is low. It is not possible to switch directly
between ICE and JTAG operations. A chip reset must be performed (NRST and NTRST) after
JTAGSEL is changed. The test reset input to the embedded ICE (NTRST) is provided sepa-
rately to facilitate debug of the boot program.
For further details on the Embedded In-Circuit-Emulator, see the ARM920T Technical Refer-
ence Manual, ARM Ltd, - DDI 0151C.
Debug Unit The Debug Unit provides a two-pin (DXRD and TXRD) UART that can be used for several
debug and trace purposes and offers an ideal means for in-situ programming solutions and
debug monitor communication. Moreover, the link with two peripheral data controller channels
provides packet handling of these tasks with processor time reduced to a minimum.
The Debug Unit also manages the interrupt handling of the COMMTX and COMMRX signals
that come from the ICE and trace the activity of the Debug Communication Channel. The
Debug Unit allows blockage of access to the system through the ICE interface.
The Debug Unit can be used to upload an application into internal SRAM. It is activated by the
boot program when no valid application is detected. The protocol used to load the application
is XMODEM.
A specific register, the Debug Unit Chip ID Register, informs about the product version and its
internal configuration.
AT91RM9200 Debug Unit Chip ID value is: 0x09290781, on 32-bit width.
For further details on the Debug Unit, see the Debug Unit datasheet; Atmel literature number,
2641.
For further details on the Debug Unit and Boot program, see the Boot Program Specifications.
Embedded Trace The AT91RM9200 features an Embedded Trace Macrocell (ETM), which is closely connected
Macrocell to the ARM9TDMI Processor. The Embedded Trace is a standard mid-level implementation
and contains the following resources:
Four pairs of address comparators
Two data comparators
65
1768BATARM08/03
Eight memory map decoder inputs
Two counters
One sequencer
Four external inputs
One external output
One 18-byte FIFO
The Embedded Trace Macrocell of the AT91RM9200 works in half-rate clock mode and thus
integrates a clock divider. This assures that the maximum frequency of all the trace port sig-
nals do not exceed one half of the ARM920T clock speed.
The Embedded Trace Macrocell input and output resources are not used in the AT91RM9200.
The Embedded Trace is a real-time trace module with the capability of tracing the ARM9TDMI
instruction and data.
The Embedded Trace debug features are only accessible in the AT91RM9200 BGA package.
For further details on Embedded Trace Macrocell, see the ETM9 (Rev2a) Technical Refer-
ence Manual, ARM Ltd. -DDI 0157E.
TPS-TPS0
ARM920T Trace FIFO
Control TPK15-TPK0
Bus Tracker
TSYNC
Scan Chain 6
ETM9
TMS
TDO
TCK
TDI
66 AT91RM9200
1768BATARM08/03
AT91RM9200
Implementation This section gives an overview of the Embedded Trace resources. For further details, see the
Details Embedded Trace Macrocell Specification, ARM Ltd. -IHI 0014H.
Three-state Sequencer The sequencer has three possible next states (one dedicated to itself and two others) and can
change on every clock cycle. The sate transition is controlled with internal events. If the user
needs multiple-stage trigger schemes, the trigger event is based on a sequencer state.
Address Comparator In single mode, address comparators compare either the instruction address or the data
address against the user-programmed address.
In range mode, the address comparators are arranged in pairs to form a virtual address range
resource.
Details of the address comparator programming are:
The first comparator is programmed with the range start address.
The second comparator is programmed with the range end address.
The resource matches if the address is within the following range:
(address > = range start address) AND (address < range end address)
Unpredictable behavior occurs if the two address comparators are not configurated in the
same way.
Data Comparator Each full address comparator is associated with a specific data comparator. A data compara-
tor is used to observe the data bus only when load and store operations occur.
A data comparator has both a value register and a mask register, therefore it is possible to
compare only certain bits of a preprogrammed value against the data bus.
Memory Decoder Inputs The eight memory map decoder inputs are connected to custom address decoders. The
address decoders divide the memory into regions of on-chip SRAM, on-chip ROM, and periph-
erals. The address decoders also optimize the ETM9 trace trigger.
FIFO An 18-byte FIFO is used to store data tracing. The FIFO is used to separate the pipeline status
from the trace packet. So, the FIFO can be used to buffer trace packets.
A FIFO overflow is detected by the embedded trace macrocell when the FIFO is full or when
the FIFO has less bytes than the user-programmed number.
For further details, see the ETM9 (Rev2a) Technical Reference Manual, ARM Ltd. DDI 0157E.
67
1768BATARM08/03
Half-rate Clocking Mode The ETM9 is implemented in half-rate mode that allows both rising and falling edge data trac-
ing of the trace clock.
The half-rate mode is implemented to maintain the signal clock integrity of high speed systems
(up to 100 Mhz).
ARM920T Clock
Trace Clock
TraceData
Care must be taken on the choice of the trace capture system as it needs to support half-rate
clock functionality.
Application Board The TCLK signal needs to be set with care, some timing parameters are required.
Restriction
Refer to AT91RM9200 JTAG/ICE Timings on page 621 and ETM Timings on page 624.
The specified target system connector is the AMP Mictor connector.
The connector must be oriented on the application board as described below in Figure 14. The
view of the PCB is shown from above with the trace connector mounted near the edge of the
board. This allows the Trace Port Analyzer to minimize the physical intrusiveness of the inter-
connected target.
ARM920T Clock
Trace Clock
TraceData
68 AT91RM9200
1768BATARM08/03
AT91RM9200
IEEE 1149.1 JTAG IEEE 1149.1 JTAG Boundary Scan allows pin-level access independent of the device packag-
Boundary Scan ing technology.
IEEE 1149.1 JTAG Boundary Scan is enabled when JTAGSEL is high. The SAMPLE,
EXTEST and BYPASS functions are implemented. In ICE debug mode, the ARM processor
responds with a non-JTAG chip ID that identifies the processor to the ICE system. This is not
IEEE 1149.1 JTAG-compliant.
It is not possible to switch directly between JTAG and ICE operations. A chip reset must be
performed (NRST and NTRST) after JTAGSEL is changed.
Two Boundary Scan Descriptor Language (BSDL) files are provided to set up testing. Each
BSDL file is dedicated to a specific packaging.
JTAG Boundary Scan The Boundary Scan Register (BSR) contains 449 bits which correspond to active pins and
Register associated control signals.
Each AT91RM9200 input pin has a corresponding bit in the Boundary Scan Register for
observability.
Each AT91RM9200 output pin has a corresponding 2-bit register in the BSR. The OUTPUT bit
contains data which can be forced on the pad. The CTRL bit can put the pad into high
impedance.
Each AT91RM9200 input/output pin corresponds to a 3-bit register in the BSR. The OUTPUT
bit contains data that can be forced on the pad. The INPUT bit facilitates the observability of
data applied to the pad. The CTRL bit selects the direction of the pad.
69
1768BATARM08/03
Table 20. JTAG Boundary Scan Register (Continued)
Bit Associated BSR
Number Pin Name Pin Type Cells
432 NCS1/SDCS Output OUTPUT
431 NCS2 Output OUTPUT
430 NCS[2:3]/NBS3 Output CTRL
429 NCS3 Output OUTPUT
428 NOE/NRD Output OUTPUT
427 INPUT
NWE/NWR0 Output
426 OUTPUT
425 INPUT
NUB/NWR1/NBS1 Output
424 OUTPUT
423 NBS3 Output OUTPUT
422 SDCKE Output OUTPUT
421 SDCKE/RAS/CAS/WE/SDA10 Output CTRL
420 RAS Output OUTPUT
419 CAS Output OUTPUT
418 WE Output OUTPUT
417 INPUT
D0 I/O
416 OUTPUT
415 D[3:0] I/O CTRL
414 INPUT
D1 I/O
413 OUTPUT
412 INPUT
D2 I/O
411 OUTPUT
410 INPUT
D3 I/O
409 OUTPUT
408 INPUT
D4 I/O
407 OUTPUT
406 D[7:4] I/O CTRL
405 INPUT
D5 I/O
404 OUTPUT
403 INPUT
D6 I/O
402 OUTPUT
401 INPUT
D7 I/O
400 OUTPUT
399 INPUT
D8 I/O
398 OUTPUT
70 AT91RM9200
1768BATARM08/03
AT91RM9200
71
1768BATARM08/03
Table 20. JTAG Boundary Scan Register (Continued)
Bit Associated BSR
Number Pin Name Pin Type Cells
363 INPUT
362 PC22/D22 I/O OUTPUT
361 CTRL
360 INPUT
359 PC23/D23 I/O OUTPUT
358 CTRL
357 INPUT
356 PC24/D24 I/O OUTPUT
355 CTRL
354 INPUT
353 PC25/D25 I/O OUTPUT
352 CTRL
351 INPUT
350 PC26/D26 I/O OUTPUT
349 CTRL
348 INPUT
347 PC27/D27 I/O OUTPUT
346 CTRL
345 INPUT
344 PC28/D28 I/O OUTPUT
343 CTRL
342 INPUT
341 PC29/D29 I/O OUTPUT
340 CTRL
339 INPUT
338 PC30/D30 I/O OUTPUT
337 CTRL
336 INPUT
335 PC31/D31 I/O OUTPUT
334 CTRL
333 INPUT
332 PC10/NCS4/CFCS I/O OUTPUT
331 CTRL
72 AT91RM9200
1768BATARM08/03
AT91RM9200
73
1768BATARM08/03
Table 20. JTAG Boundary Scan Register (Continued)
Bit Associated BSR
Number Pin Name Pin Type Cells
297 INPUT
296 PC6/NWAIT I/O OUTPUT
295 CTRL
294 INPUT
293 PA0/MISO/PCK3 I/O OUTPUT
292 CTRL
291 INPUT
290 PA1/MOSI/PCK0 I/O OUTPUT
289 CTRL
288 INPUT
287 PA2/SPCK/IRQ4 I/O OUTPUT
286 CTRL
285 INPUT
284 PA3/NPCS0/IRQ5 I/O OUTPUT
283 CTRL
282 INPUT
281 PA4/NPCS1/PCK1 I/O OUTPUT
280 CTRL
279 INPUT
278 PA5/NPCS2/TXD3 I/O OUTPUT
277 CTRL
276 INPUT
275 PD0/ETX0 I/O OUTPUT
274 CTRL
273 INPUT
272 PD1/ETX1 I/O OUTPUT
271 CTRL
270 INPUT
269 PD2/ETX2 I/O OUTPUT
268 CTRL
267 INPUT
266 PD3/ETX3 I/O OUTPUT
265 CTRL
74 AT91RM9200
1768BATARM08/03
AT91RM9200
75
1768BATARM08/03
Table 20. JTAG Boundary Scan Register (Continued)
Bit Associated BSR
Number Pin Name Pin Type Cells
231 INPUT
230 PA14/ERXER/TCLK1 I/O OUTPUT
229 CTRL
228 INPUT
227 PA15/EMDC/TCLK2 I/O OUTPUT
226 CTRL
225 INPUT
224 PA16/EMDIO/IRQ6 I/O OUTPUT
223 CTRL
222 INPUT
221 PA17/TXD0/TIOA0 I/O OUTPUT
220 CTRL
219 INPUT
218 PA18/RXD0/TIOB0 I/O OUTPUT
217 CTRL
216 INPUT
215 PA19/SCK0/TIOA1 I/O OUTPUT
214 CTRL
213 INPUT
212 PA20/CTS0/TIOB1 I/O OUTPUT
211 CTRL
210 INPUT
209 PA21/RTS0/TIOA2 I/O OUTPUT
208 CTRL
207 INPUT
206 PA22/RXD2/TIOB2 I/O OUTPUT
205 CTRL
204 INPUT
203 PA23/TXD2/IRQ3 I/O OUTPUT
202 CTRL
201 INPUT
200 PA24/SCK2/PCK1 I/O OUTPUT
199 CTRL
76 AT91RM9200
1768BATARM08/03
AT91RM9200
77
1768BATARM08/03
Table 20. JTAG Boundary Scan Register (Continued)
Bit Associated BSR
Number Pin Name Pin Type Cells
165 INPUT
164 PB4/RK0/MCDA2 I/O OUTPUT
163 CTRL
162 INPUT
161 PB5/RF0/MCDA3 I/O OUTPUT
160 CTRL
159 INPUT
158 PB6/TF1/TIOA3 I/O OUTPUT
157 CTRL
156 INPUT
155 PB7/TK1/TIOB3 I/O OUTPUT
154 CTRL
153 INPUT
152 PB8/TD1/TIOA4 I/O OUTPUT
151 CTRL
150 INPUT
149 PB9/RD1/TIOB4 I/O OUTPUT
148 CTRL
147 INPUT
146 PB10/RK1/TIOA5 I/O OUTPUT
145 CTRL
144 INPUT
143 PB11/RF1/TIOB5 I/O OUTPUT
142 CTRL
141 INPUT
140 PB12/TF2/ETX2 I/O OUTPUT
139 CTRL
138 INPUT
137 PB13/TK2/ETX3 I/O OUTPUT
136 CTRL
135 INPUT
134 PB14/TD2/ETXER I/O OUTPUT
133 CTRL
78 AT91RM9200
1768BATARM08/03
AT91RM9200
79
1768BATARM08/03
Table 20. JTAG Boundary Scan Register (Continued)
Bit Associated BSR
Number Pin Name Pin Type Cells
99 INPUT
98 PB20/TXD1 I/O OUTPUT
97 CTRL
96 INPUT
95 PB21/RXD1 I/O OUTPUT
94 CTRL
93 INPUT
92 PB22/SCK1 I/O OUTPUT
91 CTRL
90 INPUT
89 PD13/TPK1 I/O OUTPUT
88 CTRL
87 INPUT
86 PD14/TPK2 I/O OUTPUT
85 CTRL
84 INPUT
83 PD15/TD0/TPK3 I/O OUTPUT
82 CTRL
81 INPUT
80 PB23/DCD1 I/O OUTPUT
79 CTRL
78 INPUT
77 PB24/CTS1 I/O OUTPUT
76 CTRL
75 INPUT
74 PB25/DSR1/EF100 I/O OUTPUT
73 CTRL
72 INPUT
71 PB26/RTS1 I/O OUTPUT
70 CTRL
69 INPUT
68 PB27/PCK0 I/O OUTPUT
67 CTRL
80 AT91RM9200
1768BATARM08/03
AT91RM9200
81
1768BATARM08/03
Table 20. JTAG Boundary Scan Register (Continued)
Bit Associated BSR
Number Pin Name Pin Type Cells
33 INPUT
32 PD27/TPK15 I/O OUTPUT
31 CTRL
30 INPUT
29 PB28/FIQ I/O OUTPUT
28 CTRL
27 INPUT
26 PB29/IRQ0 I/O OUTPUT
25 CTRL
24 A0/NLB/NBS0 Output OUPUT
A[3:0]/NLB/NWR2/NBS0
23 Output CTRL
/NBS2
22 A1/NWR2/NBS2 Output OUTPUT
21 A2 Output OUTPUT
20 A3 Output OUTPUT
19 A4 Output OUTPUT
18 A[7:4] Output CTRL
17 A5 Output OUTPUT
16 A6 Output OUTPUT
15 A7 Output OUTPUT
14 A8 Output OUTPUT
13 A[11:8] Output CTRL
12 A9 Output OUTPUT
11 A10 Output OUTPUT
10 SDA10 Output OUTPUT
9 A11 Output OUTPUT
8 A12 Output OUTPUT
7 A[15:12] Output CTRL
6 A13 Output OUTPUT
5 A14 Output OUTPUT
4 A15 Output OUTPUT
3 A16/BA0 Output OUTPUT
2 A17/BA1 Output OUTPUT
1 A18 Output OUTPUT
82 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
PART NUMBER
15 14 13 12 11 10 9 8
PART NUMBER MANUFACTURER IDENTITY
7 6 5 4 3 2 1 0
MANUFACTURER IDENTITY 1
83
1768BATARM08/03
84 AT91RM9200
1768BATARM08/03
AT91RM9200
Boot Program
Overview The Boot Program downloads an application in any of the AT91 products integrating a ROM. It
integrates a Bootloader and a boot Uploader to assure correct information download.
The Bootloader is activated first. It looks for a sequence of eight valid ARM exception vectors
in a DataFlash connected to the SPI, an EEPROM connected to the Two-wire Interface (TWI)
or an 8-bit memory device connected to the external bus interface (EBI) (if device integrates
EBI). All these vectors must be B-branch or LDR load register instructions except for the sixth
instruction. This vector is used to store information, such as the size of the image to download
and the type of DataFlash device.
If a valid sequence is found, code is downloaded into the internal SRAM. This is followed by a
remap and a jump to the first address of the SRAM.
If no valid ARM vector sequence is found, the boot Uploader is started. It initializes the Debug
Unit serial port (DBGU) and the USB Device Port. It then waits for any transaction and down-
loads a piece of code into the internal SRAM via a Device Firmware Upgrade (DFU) protocol
for USB and XMODEM protocol for the DBGU. After the end of the download, it branches to
the application entry point at the first address of the SRAM.
The main features of the Boot Program are:
Default Boot Program stored in ROM-based products
Downloads and runs an application from external storage media into internal SRAM
Downloaded code size depends on embedded SRAM size
Automatic detection of valid application
Bootloader supporting a wide range of non-volatile memories
SPI DataFlash connected on SPI NPCS0
Two-wire EEPROM
8-bit parallel memories on NCS0 (if device integrates EBI)
Boot Uploader in case no valid program is detected in external NVM and supporting
several communication media
Serial communication on a DBGU (XModem protocol)
USB Device Port (DFU Protocol)
85
1768BATARM08/03
Flow Diagram
The Boot Program implements the algorithm presented in Figure 15.
Device
Setup
Timeout 10 ms Bootloader
Timeout 40 ms
DBGU Serial
Run
Download
OR Boot Uploader
USB Download
Run
DFU* protocol
86 AT91RM9200
1768BATARM08/03
AT91RM9200
Bootloader The Boot Program is started from address 0x0000_0000 (ARM reset vector) when the on-chip
boot mode is selected (BMS high during the reset, only on devices with EBI integrated). The
first operation is the search for a valid program in the off-chip non-volatile memories. If a valid
application is found, this application is loaded into internal SRAM and executed by branching
at address 0x0000_0000 after remap. This application may be the application code or a sec-
ond-level Bootloader.
To optimize the downloaded application code size, the Boot Program embeds several func-
tions that can be reused by the application. The Boot Program is linked at address
0x0010_0000 but the internal ROM is mapped at both 0x0000_0000 and 0x0010_0000 after
reset. All the call to functions is PC relative and does not use absolute addresses. The ARM
vectors are present at both addresses, 0x0000_0000 and 0x0010_0000.
To access the functions in ROM, a structure containing chip descriptor and function entry
points is defined at a fixed address in ROM.
If no valid application is detected, the debug serial port or the USB device port must be con-
nected to allow the upload. A specific application provided by Atmel (DFU uploader) loads the
application into internal SRAM through the USB. To load the application through the debug
serial port, a terminal application (HyperTerminal) running the Xmodem protocol is required.
Internal Internal
SRAM ROM
REMAP
0x0020_0000 0x0010_0000
Internal Internal
ROM SRAM
0x0000_0000 0x0000_0000
After reset, the code in internal ROM is mapped at both addresses 0x0000_0000 and
0x0010_0000:
87
1768BATARM08/03
Valid Image The Bootloader software looks for a valid application by analyzing the first 32 bytes corre-
Detection sponding to the ARM exception vectors. These bytes must implement ARM instructions for
either branch or load PC with PC relative addressing. The sixth vector, at offset 0x18, contains
the size of the image to download and the DataFlash parameters.
The user must replace this vector with his own vector.
31 28 27 24 23 20 19 16 15 12 11 0
1 1 1 0 1 1 I P U 1 W 0 Rn Rd
31 28 27 24 23 0
1 1 1 0 1 0 1 0 Offset (24 bits)
00 ea00000b B 0x2c
004 e59ff014 LDR PC, [PC,20]
08 e59ff014 LDR PC, [PC,20]
0c e59ff014 LDR PC, [PC,20]
10 e59ff014 LDR PC, [PC,20]
14 00001234 LDR PC, [PC,20] <- Code size = 4660 bytes
18 e51fff20 LDR PC, [PC,-0xf20]
1c e51fff20 LDR PC, [PC,-0xf20]
In download mode (DataFlash, EEPROM or 8-bit memory in device with EBI integrated), the
size of the image to load into SRAM is contained in the location of the sixth ARM vector. Thus
the user must replace this vector by the correct vector for his application.
88 AT91RM9200
1768BATARM08/03
AT91RM9200
Structure of ARM The ARM exception vector 6 is used to store information needed by the Boot ROM down-
Vector 6 loader. This information is described below.
31 17 16 13 12 8 7 0
The first eight bits contain the number of blocks to download. The size of a block is 512 bytes,
allowing download of up to 128K bytes.
The bits 13 to 16 determine the DataFlash page number.
DataFlash page number = 2(Nb of pages)
The last 15 bits contain the DataFlash page size.
Example The following vector contains the information to describe a AT45DB642 DataFlash which con-
tains 11776 bytes to download.
Vector 6 is 0x0841A017 (00001000010000011010000000010111b):
Size to download: 0x17 * 512 bytes = 11776 bytes
Number pages (1101b): 13 ==> Number of DataFlash pages = 213 = 8192
DataFlash page size(000010000100000b) = 1056
For download in the EEPROM or 8-bit external memory (if device integrates EBI), only the size
to be downloaded is decoded.
89
1768BATARM08/03
Bootloader The Boot Program performs device initialization followed by the download procedure. If unsuc-
Sequence cessful, the upload is done via the USB or debug serial port.
Download Procedure The download procedure checks for a valid boot on several devices. The first device checked
is a serial DataFlash connected to the NPCS0 of the SPI, followed by the serial EEPROM con-
nected to the TWI and by an 8-bit parallel memory connected on NCS0 of the External Bus
Interface (if EBI is implemented in the product).
90 AT91RM9200
1768BATARM08/03
AT91RM9200
Serial DataFlash The Boot Program supports all Atmel DataFlash devices. Table 21 summarizes the parame-
Download ters to include in the ARM vector 6 for all devices.
The DataFlash has a Status Register that determines all the parameters required to access
the device.
Thus, to be compatible with the future design of the DataFlash, parameters are coded in the
ARM vector 6.
Start
Memory Uploader
Only for Device without
EBI integrated
Yes
8 vectors
(except vector 6) are LDR No
or Branch instruction ?
Yes
Read the Two-Wire EEPROM into the
internal SRAM
(code size to read in vector 6)
End
91
1768BATARM08/03
Serial Two-wire Generally, serial EEPROMs have no identification code. The bootloader checks for an
EEPROM Download acknowledgment on the first read. The device address on the two-wire bus must be 0x0.
The bootloader supports the devices listed in Table 23.
Start
Memory Uploader
Only for Device without
EBI integrated
Yes
8 vectors
(except vector 6) are LDR No
or Branch instruction ?
Yes
Read the Two-Wire EEPROM into the
internal SRAM
(code size to read in vector 6)
End
92 AT91RM9200
1768BATARM08/03
AT91RM9200
8-bit Parallel Flash Eight-bit parallel Flash download is supported if the product integrates an External Bus Inter-
Download (Applicable face (EBI).
to Devices with EBI)
All 8-bit memory devices supported by the EBI when NCS0 is configured in 8-bit data bus
width are supported by the bootloader.
Start
8 vectors
(except vector 6) are LDR No Memory uploader
or Branch instruction ?
Yes
Read the external memory into the
internal SRAM
(code size to read in vector 6)
End
93
1768BATARM08/03
Boot Uploader If no valid boot device has been found during the Bootloader sequence, initialization of serial
communication devices (DBGU and USB device ports) is performed.
Initialization of the DBGU serial port (115200 bauds, 8, N, 1) and Xmodem protocol
start
Initialization of the USB Device Port and DFU protocol start
Download of the application
The boot Uploader performs the DFU and Xmodem protocols to upload the application into
internal SRAM at address 0x0020_0000.
The Boot Program uses a piece of internal SRAM for variables and stacks. To prevent any
upload error, the size of the application to upload must be less than the SRAM size minus 3K
bytes.
After the download, the peripheral registers are reset, the interrupts are disabled and the
remap is performed. After the remap, the internal SRAM is at address 0x0000_0000 and the
internal ROM at address 0x0010_0000. The instruction setting the PC to 0 is the one just after
the remap command. This instruction is fetched in the pipe before doing the remap and exe-
cuted just after. This fetch cycle executes the downloaded image.
External
Communication
Channels
DBGU Serial Port The upload is performed through the DBGU serial port initialized to 115200 Baud, 8, n, 1.
The DBGU sends the character C (0x43) to start an Xmodem protocol. Any terminal perform-
ing this protocol can be used to send the application file to the target. The size of the binary file
to send depends on the SRAM size embedded in the product (Refer to the microcontroller
datasheet to determine SRAM size embedded in the microcontroller). In all cases, the size of
the binary file must be lower than SRAM size because the Xmodem protocol requires some
SRAM memory to work.
Xmodem Protocol The Xmodem protocol supported is the 128-byte length block. This protocol uses a two char-
acter CRC-16 to guarantee detection of a maximum bit error.
Xmodem protocol with CRC is accurate provided both sender and receiver report successful
transmission. Each block of the transfer looks like:
<SOH><blk #><255-blk #><--128 data bytes--><checksum> in which:
<SOH> = 01 hex
<blk #> = binary number, starts at 01, increments by 1, and wraps 0FFH to 00H
(not to 01)
<255-blk #> = 1s complement of the blk#.
<checksum> = 2 bytes CRC16
Figure 23 shows a transmission using this protocol.
94 AT91RM9200
1768BATARM08/03
AT91RM9200
Host Device
ACK
ACK
ACK
EOT
ACK
USB Device Port A 48 MHz USB clock is necessary to use USB Device port. It has been programmed earlier in
the device initialization with PLLB configuration.
DFU Protocol The DFU allows upgrade of the firmware of USB devices. The DFU algorithm is a part of the
USB specification. For more details, refer to USB Device Firmware Upgrade Specification,
Rev. 1.0.
There are four distinct steps when carrying out a firmware upgrade:
1. Enumeration: The device informs the host of its capabilities.
2. Reconfiguration: The host and the device agree to initiate a firmware upgrade.
3. Transfer: The host transfers the firmware image to the device. Status requests are
employed to maintain synchronization between the host and the device.
4. Manifestation: Once the device reports to the host that it has completed the reprogram-
ming operations, the host issues a reset and the device executes the upgraded
firmware.
Host Device
USB reset
USB reset
95
1768BATARM08/03
Hardware and The software limitations of the Boot Program are:
Software The downloaded code size is less than the SRAM size embedded in the product.
Constraints The device address of the EEPROM must be 0 on the TWI bus.
The code is always downloaded from the device address 0x0000_0000 (DataFlash,
EEPROM) to the address 0x0000_0000 of the internal SRAM (after remap).
The downloaded code must be position-independent or linked at address 0x0000_0000.
The hardware limitations of the Boot Program are:
The DataFlash must be connected to NPCS0 of the SPI.
The 8-bit parallel Flash must be connected to NCS0 of the EBI (applicable for devices with
integrated EBI).
The SPI and TWI drivers use several PIOs in alternate functions to communicate with devices.
Care must be taken when these PIOs are used by the application. The devices connected
could be unintentionally driven at boot time, and electrical conflicts between SPI or TWI output
pins and the connected devices may appear.
To assure correct functionality, it is recommended to plug in critical devices to other pins or to
boot on an external 16-bit parallel memory (if product integrates an EBI) by setting bit BMS.
Table 24 contains a list of pins that are driven during the Boot Program execution. These pins
are driven during the boot sequence for a period of about 6 ms if no correct boot program is
found. The download through the TWI takes about 5 sec for 64K bytes due to the TWI bit rate
(100 Kbits/s).
For the DataFlash driven by SPCK signal at 12 MHz, the time to download 64K bytes is
reduced to 66 ms.
Before performing the jump to the application in internal SRAM, all the PIOs and peripherals
used in the Boot Program are set to their reset state.
96 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview An embedded software service is an independent software object that drives device resources
for frequently implemented tasks. The object-oriented approach of the software provides an
easy way to access services to build applications.
An AT91 service has several purposes:
It gives software examples dedicated to the AT91 devices.
It can be used on several AT91 device families.
It offers an interface to the software stored in the ROM.
The main features of the software services are:
Compliant with ATPCS
Compliant with ANSI/ISO Standard C
Compiled in ARM/Thumb Interworking
ROM Entry Service
Tempo, Xmodem and DataFlash services
CRC and Sine tables
Service Definition
Service Structure
Methods In the service structure, pointers to functions are supposed to be initialized by default to the
standard functions. Only the default standard functions reside in ROM. Default methods can
be overloaded by custom application methods.
Methods do not declare any static variables nor invoke global variables. All methods are
invoked with a pointer to the service structure. A method can access and update service data
without restrictions.
Similarly, there is no polling in the methods. In fact, there is a method to start the functionality
(a read to give an example), a method to get the status (is the read achieved?), and a call-
back, initialized by the start method. Thus, using service, the client application carries out a
synchronous read by starting the read and polling the status, or an asynchronous read speci-
fying a callback when starting the read operation.
Service Entry Point Each AT91 service, except for the ROM Entry Service (see page 101), defines a function
named AT91F_Open_<Service>. It is the only entry point defined for a service. Even if the
functions AT91F_Open_<Service> may be compared with object constructors, they do not
act as constructors in that they initiate the service structure but they do not allocate it. Thus the
customer application must allocate it.
97
1768BATARM08/03
Example
// Allocation of the service structure
AT91S_Pipe pipe;
// Opening of the service
AT91PS_Pipe pPipe = AT91F_OpenPipe(&pipe, );
Method pointers in the service structure are initialized to the default methods defined in the
AT91 service. Other fields in the service structure are initialized to default values or with the
arguments of the function AT91F_Open_<Service>.
In summary, an application must know what the service structure is and where the function
AT91F_Open_<Service> is.
The default function AT91F_Open_<Service> may be redefined by the application or com-
prised in an application-defined function.
Using a Service
Opening a Service The entry point to a service is established by initializing the service structure. An open function
is associated with each service structure, except for the ROM Entry Service (see page 101).
Thus, only the functions AT91F_Open_<service> are visible from the user side. Access to
the service methods is made via function pointers in the service structure.
The function AT91F_Open_<service> has at least one argument: a pointer to the service
structure that must be allocated elsewhere. It returns a pointer to the base service structure or
a pointer to this service structure.
The function AT91F_Open_<service> initializes all data members and method pointers. All
function pointers in the service structure are set to the services functions.
The advantage of this method is to offer a single entry point for a service. The methods of a
service are initialized by the open function and each member can be overloaded.
Overloading a Method Default methods are defined for all services provided in ROM. These methods may not be
adapted to a project requirement. It is possible to overload default methods by methods
defined in the project.
A me th o d i s a p o in te r t o a f u nc t io n . Th i s po i nt er i s i ni t i al i z ed b y t h e fu n c ti o n
AT91F_Open_<Service>. To overload one or several methods in a service, the function
pointer must be updated to the new method.
It is possible to overload just one method of a service or all the methods of a service. In this
latter case, the functionality of the service is user-defined, but still works on the same data
structure.
Note: Calling the default function AT91F_Open_<Service> ensures that all methods and data are
initialized.
98 AT91RM9200
1768BATARM08/03
AT91RM9200
This can be done by writing a new function My_OpenService(). This new Open function
must call the library-defined function AT91F_Open_<Service>, and then update one or sev-
eral function pointers:
Table 25. Overloading a Method with the Overloading of the Open Service Function
Default service behavior in ROM Overloading AT91F_ChildMethod by My_ChildMethod
// Init the service with default methods // Allocation of the service structure
AT91PS_Service AT91F_OpenService( AT91S_Service service;
AT91PS_Service pService)
{ // Opening of the service
pService->data = 0; AT91PS_Service pService =
pService->MainMethod =AT91F_MainMethod; My_OpenService(&service);
pService->ChildMethod=AT91F_ChildMethod;
return pService;
}
99
1768BATARM08/03
This also can be done directly by overloading the method after the use of AT91F_Open_<Ser-
vice> method:
Table 26. Overloading a Method without the Overloading of the Open Service Function.
Default service behavior in ROM Overloading AT91F_ChildMethod by My_ChildMethod
}
// Init the service with default methods
AT91PS_Service AT91F_OpenService(
AT91PS_Service pService)
{
pService->data = 0;
pService->MainMethod =AT91F_MainMethod;
pService->ChildMethod=AT91F_ChildMethod;
return pService;
}
100 AT91RM9200
1768BATARM08/03
AT91RM9200
Definition Several AT91 products embed ROM. In most cases, the ROM integrates a bootloader and
several services that may speed up the application and reduce the application code size.
When software is fixed in the ROM, the address of each object (function, constant, table, etc.)
must be related to a customer application. This is done by providing an address table to the
linker. For each version of ROM, a new address table must be provided and all client applica-
tions must be recompiled.
The Embedded Software Services offer another solution to access objects stored in ROM. For
each embedded service, the customer application requires only the address of the Service
Entry Point (see page 97).
Even if these services have only one entry point (AT91F_Open_<Service> function), they must
be specified to the linker. The Embedded Software Services solve this problem by providing a
dedicated service: the ROM Entry Service.
The goal of this product-dedicated service is to provide just one address to access all ROM
functionalities.
ROM Entry Service The ROM Entry Service of a product is a structure named AT91S_RomBoot. Some members of
this structure point to the open functions of all services stored in ROM (function
AT91F_Open_<Service>) but also the CRC and Sine Arrays. Thus, only the address of the
AT91S_RomBoot has to be published.
Table 27. Initialization of the ROM Entry Service and Use with an Open Service Method
Application Memory Space ROM Memory Space
// Init the ROM Entry Service AT91S_TempoStatus AT91F_OpenCtlTempo(
AT91S_RomBoot const *pAT91; AT91PS_CtlTempo pCtlTempo,
pAT91 = AT91C_ROM_BOOT_ADDRESS; void const *pTempoTimer )
{
// Allocation of the service structure
AT91S_CtlTempo tempo; ...
}
// Call the Service Open method
pAT91->OpenCtlTempo(&tempo, ...); AT91S_TempoStatus AT91F_CtlTempoCreate (
AT91PS_CtlTempo pCtrl,
// Use of tempo methods AT91PS_SvcTempo pTempo)
tempo.CtlTempoCreate(&tempo, ...); {
...
The application obtains the address of the ROM Entry Service and initializes an instance of
the AT91S_RomBoot structure. To obtain the Open Service Method of another service stored in
ROM, the application uses the appropriate member of the AT91S_RomBoot structure.
The address of the AT91S_RomBoot can be found at the beginning of the ROM, after the
exception vectors.
101
1768BATARM08/03
Tempo Service
Presentation The Tempo Service allows a single hardware system timer to support several software timers
running concurrently. This works as an object notifier.
There are two objects defined to control the Tempo Service: AT91S_CtlTempo and
AT91S_SvcTempo.
The application declares one instance of AT91S_CtlTempo associated with the hardware
system timer. Additionally, it controls a list of instances of AT91S_SvcTempo.
Each time the application requires another timer, it asks the AT91S_CtlTempo to create a
new instance of AT91S_SvcTempo, then the application initializes all the settings of
AT91S_SvcTempo.
102 AT91RM9200
1768BATARM08/03
AT91RM9200
103
1768BATARM08/03
Using the Service The first step is to find the address of the open service method AT91F_OpenCtlTempo using
the ROM Entry Service.
Allocate one instance of AT91S_CtlTempo and AT91S_SvcTempo in the application mem-
ory space:
// Allocate the service and the control tempo
AT91S_CtlTempo ctlTempo;
AT91S_SvcTempo svcTempo1;
At this stage, the application can use the AT91S_CtlTempo service members.
If the application wants to overload an object member, it can be done now. For example, if
AT91F_CtlTempoCreate(&ctlTempo, &svcTempo1) method is to be replaced by the applica-
tion defined as my_CtlTempoCreate(...), the procedure is as follows:
// Overload AT91F_CtlTempoCreate
ctlTempo.CtlTempoCreate = my_CtlTempoCreate;
Start the timeout by calling Start method of the svcTempo1 object. Depending on the function
parameters, either a callback is started at the end of the countdown or the status of the time-
out is checked by reading the TickTempo member of the svcTempo1 object.
// Start the timeout
svcTempo1.Start(&svcTempo1,100,0,NULL,NULL);
// Wait for the timeout of 100 (unity depends on the timer programmation)
// No repetition and no callback.
while (svcTempo1.TickTempo);
When the application needs another software timer to control a timeout, it:
Allocates one instance of AT91S_SvcTempo in the application memory space
// Allocate the service
AT91S_SvcTempo svcTempo2;
104 AT91RM9200
1768BATARM08/03
AT91RM9200
Xmodem Service
Presentation The Xmodem service is an application of the communication pipe abstract layer. This layer is
media-independent (USART, USB, etc.) and gives entry points to carry out reads and writes
on an abstract media, the pipe.
Communication Pipe The pipe communication structure is a virtual structure that contains all the functions required
Service to read and write a buffer, regardless of the communication media and the memory
management.
The pipe structure defines:
a pointer to a communication service structure AT91PS_SvcComm
a pointer to a buffer manager structure AT91PS_Buffer
pointers on read and write functions
pointers to callback functions associated to the read and write functions
The following structure defines the pipe object:
// Pipe methods
AT91S_PipeStatus (*Write) (
struct _AT91S_Pipe *pPipe,
char const * pData,
unsigned int size,
void (*callback) (AT91S_PipeStatus, void *),
void *privateData);
AT91S_PipeStatus (*Read) (
struct _AT91S_Pipe *pPipe,
char *pData,
unsigned int size,
void (*callback) (AT91S_PipeStatus, void *),
void *privateData);
AT91S_PipeStatus (*AbortWrite) (struct _AT91S_Pipe *pPipe);
AT91S_PipeStatus (*AbortRead) (struct _AT91S_Pipe *pPipe);
AT91S_PipeStatus (*Reset) (struct _AT91S_Pipe *pPipe);
char (*IsWritten) (struct _AT91S_Pipe *pPipe,char const *pVoid);
char (*IsReceived) (struct _AT91S_Pipe *pPipe,char const *pVoid);
} AT91S_Pipe, *AT91PS_Pipe;
The Xmodem protocol implementation demonstrates how to use the communication pipe.
105
1768BATARM08/03
Description of the Buffer The AT91PS_Buffer is a pointer to the AT91S_Buffer structure manages the buffers. This
Structure structure embeds the following functions:
pointers to functions that manage the read buffer
pointers to functions that manage the write buffer
All the functions can be overloaded by the application to adapt buffer management.
A simple implementation of buffer management for the Xmodem Service is provided in the
boot ROM source code.
106 AT91RM9200
1768BATARM08/03
AT91RM9200
Description of the The SvcComm structure provides the interface between low-level functions and the pipe
SvcComm Structure object.
It contains pointers of functions initialized to the lower level functions (e.g. SvcXmodem).
The Xmodem Service implementation gives an example of SvcComm use.
typedef struct _AT91S_Service
{
// Methods:
AT91S_SvcCommStatus (*Reset) (struct _AT91S_Service *pService);
AT91S_SvcCommStatus (*StartTx)(struct _AT91S_Service *pService);
AT91S_SvcCommStatus (*StartRx)(struct _AT91S_Service *pService);
AT91S_SvcCommStatus (*StopTx) (struct _AT91S_Service *pService);
AT91S_SvcCommStatus (*StopRx) (struct _AT91S_Service *pService);
char (*TxReady)(struct _AT91S_Service *pService);
char (*RxReady)(struct _AT91S_Service *pService);
// Data:
struct _AT91S_Buffer *pBuffer; // Link to a buffer object
void *pChild;
} AT91S_SvcComm, *AT91PS_SvcComm;
107
1768BATARM08/03
Description of the The SvcXmodem service is a reusable implementation of the Xmodem protocol. It supports
SvcXmodem Structure only the 128-byte packet format and provides read and write functions. The SvcXmodem
structure defines:
a pointer to a handler initialized to readHandler or writeHandler
a pointer to a function that processes the xmodem packet crc
a pointer to a function that checks the packet header
a pointer to a function that checks data
With this structure, the Xmodem protocol can be used with all media (USART, USB, etc.).
Only private methods may be overloaded to adapt the Xmodem protocol to a new media.
The default implementation of the Xmodem uses a USART to send and receive packets. Read
and write functions implement peripheral data controller facilities to reduce interrupt overhead.
It assumes the USART is initialized, the memory buffer allocated and the interrupts
programmed.
A periodic timer is required by the service to manage timeouts and the periodic transmission of
the character C (Refer to Xmodem protocol). This feature is provided by the Tempo Service.
The following structure defines the Xmodem Service:
// Public Methods:
AT91S_SvcCommStatus (*Handler) (struct _AT91PS_SvcXmodem *, unsigned int);
AT91S_SvcCommStatus (*StartTx) (struct _AT91PS_SvcXmodem *, unsigned int);
AT91S_SvcCommStatus (*StopTx) (struct _AT91PS_SvcXmodem *, unsigned int);
// Private Methods:
AT91S_SvcCommStatus (*ReadHandler) (struct _AT91PS_SvcXmodem *, unsigned int
csr);
AT91S_SvcCommStatus (*WriteHandler) (struct _AT91PS_SvcXmodem *, unsigned int
csr);
unsigned short (*GetCrc) (char *ptr, unsigned int count);
char (*CheckHeader) (unsigned char currentPacket, char *packet);
char (*CheckData) (struct _AT91PS_SvcXmodem *);
char *pData;
unsigned int dataSize; // = XMODEM_DATA_STX or XMODEM_DATA_SOH
char packetDesc[AT91C_XMODEM_PACKET_SIZE];
unsigned char packetId; // Current packet
char packetStatus;
char isPacketDesc;
char eot; // end of transmition
} AT91S_SvcXmodem, *AT91PS_SvcXmodem
108 AT91RM9200
1768BATARM08/03
AT91RM9200
109
1768BATARM08/03
Using the Service The following steps show how to initialize and use the Xmodem Service in an application:
Variables definitions:
AT91S_RomBoot const *pAT91; // struct containing Openservice functions
AT91S_SBuffer sXmBuffer; // Xmodem Buffer allocation
AT91S_SvcXmodem svcXmodem; // Xmodem service structure allocation
AT91S_Pipe xmodemPipe;// xmodem pipe communication struct
AT91S_CtlTempo ctlTempo; // Tempo struct
AT91PS_Buffer pXmBuffer; // Pointer on a buffer structure
AT91PS_SvcComm pSvcXmodem; // Pointer on a Media Structure
Initialisations
// Call Open methods:
pAT91 = AT91C_ROM_BOOT_ADDRESS;
// OpenCtlTempo on the system timer
pAT91->OpenCtlTempo(&ctlTempo, (void *) &(pAT91->SYSTIMER_DESC));
ctlTempo.CtlTempoStart((void *) &(pAT91->SYSTIMER_DESC));
// Xmodem buffer initialisation
pXmBuffer = pAT91->OpenSBuffer(&sXmBuffer);
pSvcXmodem = pAT91->OpenSvcXmodem(&svcXmodem, AT91C_BASE_DBGU, &ctlTempo);
// Open communication pipe on the xmodem service
pAT91->OpenPipe(&xmodemPipe, pSvcXmodem, pXmBuffer);
// Init the DBGU peripheral
// Open PIO for DBGU
AT91F_DBGU_CfgPIO();
// Configure DBGU
AT91F_US_Configure (
(AT91PS_USART) AT91C_BASE_DBGU, // DBGU base address
MCK, // Master Clock
AT91C_US_ASYNC_MODE, // mode Register to be programmed
BAUDRATE , // baudrate to be programmed
0); // timeguard to be programmed
// Enable Transmitter
AT91F_US_EnableTx((AT91PS_USART) AT91C_BASE_DBGU);
// Enable Receiver
AT91F_US_EnableRx((AT91PS_USART) AT91C_BASE_DBGU);
// Initialize the Interrupt for System Timer and DBGU (shared interrupt)
// Initialize the Interrupt Source 1 for SysTimer and DBGU
AT91F_AIC_ConfigureIt(AT91C_BASE_AIC,
AT91C_ID_SYS,
AT91C_AIC_PRIOR_HIGHEST,
AT91C_AIC_SRCTYPE_INT_LEVEL_SENSITIVE,
AT91F_ASM_ST_DBGU_Handler);
110 AT91RM9200
1768BATARM08/03
AT91RM9200
DataFlash Service
Presentation The DataFlash Service allows the Serial Peripheral Interface (SPI) to support several Serial
DataFlash and DataFlash Cards for reading, programming and erasing operations.
This service is based on SPI interrupts that are managed by a specific handler. It also uses the
corresponding PDC registers.
For more information on the commands available in the DataFlash Service, refer to the rele-
vant DataFlash documentation.
111
1768BATARM08/03
Table 30. DataFlash Service Methods (Continued)
Associated Function Pointers & Methods Used by Default Description
// Typical Use: Member of AT91S_SvcDataFlash structure
AT91S_SvcDataFlash svcDataFlash; Read a Page in DataFlash.
svcDataFlash.PageRead(...); Input Parameters:
Pointer on DataFlash Service Structure.
// Default Method: DataFlash address.
AT91S_SvcDataFlashStatus AT91F_DataFlashPageRead ( Data buffer destination pointer.
AT91PS_SvcDataFlash pSvcDataFlash, Number of bytes to read.
unsigned int src, Output Parameters:
unsigned char *dataBuffer, Returns 0 if DataFlash is Busy.
int sizeToRead ) Returns 1 if DataFlash Ready.
// Typical Use: Member of AT91S_SvcDataFlash structure.
AT91S_SvcDataFlash svcDataFlash; Continuous Stream Read.
svcDataFlash.ContinuousRead(...); Input Parameters:
Pointer on DataFlash Service Structure.
// Default Method: DataFlash address.
AT91S_SvcDataFlashStatus Data buffer destination pointer.
AT91F_DataFlashContinuousRead (
Number of bytes to read.
AT91PS_SvcDataFlash pSvcDataFlash,
Output Parameters:
int src,
Returns 0 if DataFlash is Busy.
unsigned char *dataBuffer,
Returns 1 if DataFlash is Ready.
int sizeToRead )
// Typical Use: Member of AT91S_SvcDataFlash structure.
AT91S_SvcDataFlash svcDataFlash; Read the Internal DataFlash SRAM Buffer 1 or 2.
svcDataFlash.ReadBuffer(...); Input Parameters:
Pointer on DataFlash Service Structure.
// Default Method: Choose Internal DataFlash Buffer 1 or 2 command.
AT91S_SvcDataFlashStatus AT91F_DataFlashReadBuffer DataFlash address.
(
Data buffer destination pointer.
AT91PS_SvcDataFlash pSvcDataFlash,
Number of bytes to read.
unsigned char BufferCommand,
Output Parameters:
unsigned int bufferAddress,
Returns 0 if DataFlash is Busy.
unsigned char *dataBuffer,
Returns 1 if DataFlash is Ready.
int sizeToRead )
Returns 4 if DataFlash Bad Command.
Returns 5 if DataFlash Bad Address.
// Typical Use: Member of AT91S_SvcDataFlash structure
AT91S_SvcDataFlash svcDataFlash; Read a Page in the Internal SRAM Buffer 1 or 2.
svcDataFlash.MainMemoryToBufferTransfert(...); Input Parameters:
Pointer on DataFlash Service Structure.
// Default Method: Choose Internal DataFlash Buffer 1 or 2 command.
AT91S_SvcDataFlashStatus Page to read.
AT91F_MainMemoryToBufferTransfert(
Output Parameters:
AT91PS_SvcDataFlash pSvcDataFlash,
Returns 0 if DataFlash is Busy.
unsigned char BufferCommand,
Returns 1 if DataFlash is Ready.
unsigned int page)
Returns 4 if DataFlash Bad Command.
112 AT91RM9200
1768BATARM08/03
AT91RM9200
113
1768BATARM08/03
Table 30. DataFlash Service Methods (Continued)
Associated Function Pointers & Methods Used by Default Description
// Typical Use: Member of AT91S_SvcDataFlash structure
AT91S_SvcDataFlash svcDataFlash; Page Program through Internal SRAM Buffer 1 or 2.
svcDataFlash.PagePgmBuf(...); Input Parameters:
Pointer on DataFlash Service Structure.
// Default Method: Choose Internal DataFlash Buffer 1 or 2 command.
AT91S_SvcDataFlashStatus Source buffer.
AT91F_DataFlashPagePgmBuf(
DataFlash destination address.
AT91PS_SvcDataFlash pSvcDataFlash,
Number of bytes to write.
unsigned char BufferCommand,
Output Parameters:
unsigned char *src,
Returns 0 if DataFlash is Busy.
unsigned int dest,
Returns 1 if DataFlash is Ready.
unsigned int SizeToWrite)
Returns 4 if DataFlash Bad Command.
// Typical Use: Member of AT91S_SvcDataFlash structure.
AT91S_SvcDataFlash svcDataFlash; Write data to the Internal SRAM buffer 1 or 2.
svcDataFlash.WriteBuffer(...); Input Parameters:
Pointer on DataFlash Service Structure.
// Default Method: Choose Internal DataFlash Buffer 1 or 2 command.
AT91S_SvcDataFlashStatus Pointer on data buffer to write.
AT91F_DataFlashWriteBuffer (
Address in the internal buffer.
AT91PS_SvcDataFlash pSvcDataFlash,
Number of bytes to write.
unsigned char BufferCommand,
Output Parameters:
unsigned char *dataBuffer,
Returns 0 if DataFlash is Busy.
unsigned int bufferAddress,
Returns 1 if DataFlash is Ready.
int SizeToWrite )
Returns 4 if DataFlash Bad Command.
Returns 5 if DataFlash Bad Address.
// Typical Use: Member of AT91S_SvcDataFlash structure.
AT91S_SvcDataFlash svcDataFlash; Write Internal Buffer to the DataFlash Main Memory.
svcDataFlash.WriteBufferToMain(...); Input Parameters:
Pointer on DataFlash Service Structure.
// Default Method: Choose Internal DataFlash Buffer 1 or 2 command.
AT91S_SvcDataFlashStatus AT91F_WriteBufferToMain ( Main memory address on DataFlash.
AT91PS_SvcDataFlash pSvcDataFlash, Output Parameters:
unsigned char BufferCommand, Returns 0 if DataFlash is Busy.
unsigned int dest ) Returns 1 if DataFlash is Ready.
114 AT91RM9200
1768BATARM08/03
AT91RM9200
115
1768BATARM08/03
Using the Service The first step is to find the address of the open service method AT91F_OpenSvcDataFlash using
the ROM Entry Service.
1. Allocate one instance of AT91S_SvcDataFlash and AT91S_Dataflash in the application
memory space:
// Allocate the service and a device structure.
AT91S_SvcDataFlash svcDataFlash;
AT91S_Dataflash Device; // member of AT91S_SvcDataFlash service
3. Configure the DataFlash structure with its correct features and link it to the device
structure in the AT91S_SvcDataFlash service structure:
// Example with an ATMEL AT45DB321B DataFlash
Device.pages_number = 8192;
Device.pages_size = 528;
Device.page_offset = 10;
Device.byte_mask = 0x300;
// Link to the service structure
svcDataFlash.pDevice = &Device;
4. Now the different methods can be used. Following is an example of a Page Read of
528 bytes on page 50:
// Result of the read operation in RxBufferDataFlash
unsigned char RxBufferDataFlash[528];
svcDataFlash.PageRead(&svcDataFlash,
(50*528),RxBufferDataFlash,528);
116 AT91RM9200
1768BATARM08/03
AT91RM9200
CRC Service
Presentation This service differs from the preceding ones in that it is structured differently: it is composed
of an array and some methods directly accessible via the AT91S_RomBoot structure.
117
1768BATARM08/03
Using the Service Compute the CRC16 CCITT of a 256-byte buffer and save it in the crc16 variable:
// Compute CRC16 CCITT
unsigned char BufferToCompute[256];
short crc16;
... (BufferToCompute Treatment)
pAT91->CRCCCITT(&BufferToCompute,256,&crc16);
Sine Service
Presentation This service differs from the preceding one in that it is structured differently: it is composed of
an array and a method directly accessible through the AT91S_RomBoot structure.
// Array Embedded:
const short AT91C_SINUS180_TAB[256]
118 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview This chapter describes the AT91RM9200 reset signals and how to use them in order to assure
correct operation of the device.
The AT91RM9200 has two reset input lines called NRST and NTRST. Each line provides,
respectively:
Initialization of the User Interface registers (defined in the user interface of each
peripheral) and:
Sample the signals needed at bootup
Compel the processor to fetch the next instruction at address zero.
Initialization of the embedded ICE TAP controller.
The NRST signal must be considered as the System Reset signal and the reader must take
care when designing the logic to drive this reset signal. NTRST is typically used by the hard-
ware debug interface which uses the In-Circuit Emulator unit and Initializes it without affecting
the normal operation of the ARM processor. This line shall also be driven by an on board
logic.
Both NRST and NTRST are active low signals that asynchronously reset the logic in the
AT91RM92000.
Reset Conditions
NRST Conditions NRST is the active low reset input. When power is first applied to the system, a power-on reset
(also denominated as cold reset) must be applied to the AT91RM9200. During this transient
state, it is mandatory to hold the reset signal low long enough for the power supply to reach a
working nominal level and for the oscillator to reach a stable operating frequency. Typically,
these features are provided by every power supply supervisor which, under a threshold volt-
age limit, the electrical environment is considered as not nominal. Power-up is not the only
event to be be considered as power-down or a brownout are also occurrences that assert the
NRST signal. The threshold voltage must be selected according to the minimum operating
voltage of the AT91RM9200 power supply lines marked as VDD in Figure 25. (See DC Char-
acteristics on page 596.)
The choice of the reset holding delay depends on the start-up time of the low frequency oscil-
lator as shown below in Figure 25. (See 32 kHz Oscillator Characteristics on page 599.)
NRST
119
1768BATARM08/03
NRST can also be asserted in circumstances other than the power-up sequence, such as a
manual command. This assertion can be performed asynchronously, but exit from reset is
synchronized internally to the default active clock. During normal operation, NRST must be
active for a minimum delay time to ensure correct behavior. See Figure 26 and Table 33.
RST1
NRST
NTRST Assertion As with the NRST signal, at power-up, the NTRST signal must be valid while the power supply
has not obtained the the minimum recommended working level. (See DC Characteristics on
page 596.). A clock on TCK is not required to validate this reset request.
As with the NRST signal, NTRST can also be asserted in circumstances other than the power-
up sequence, such as a manual command or an ICE Interface action. This assertion and de-
assertion can be performed asynchronously but must be active for a minimum delay time.
(See JTAG/ICE Timings on page 621.)
Reset
Management
System Reset The system reset functionality is provided through the NRST signal.
This Reset signal is used to compel the microcontroller unit to assume a set of initial
conditions:
Sample the Boot Mode Select (BMS) logical state.
Restore the default states (default values) of the user interface.
Require the processor to perform the next instruction fetch from address zero.
With the exception of the program counter and the Current Program Status Register, the pro-
cessors registers do not have defined reset states. When the microcontrollers NRST input is
asserted, the processor immediately stops execution of the current instruction independently
of the clock.
The system reset circuitry must take two types of reset requests into account:
The cold reset needed for the power-up sequence.
The user reset request.
Both have the same effect but can have different assertion time requirements regarding the
NRST pin. In fact, the cold reset assertion has to overlap the start-up time of the system. The
user reset request requires a shorter assertion delay time than does cold reset.
Test Reset Test reset functionality is provided through the NTRST signal.
120 AT91RM9200
1768BATARM08/03
AT91RM9200
The NTRST control pin initializes the selected TAP controller. The TAP controller involved in
this reset is determined according to the initial logical state applied on the JTAGSEL pin after
the last valid NRST.
In Boundary Scan Mode, after a NTRST assertion, the IDCODE instruction is set onto the out-
put of the instruction register in the Test-Logic-Reset controller state.
Otherwise, in ICE Mode, the reset action is as follows:
The core exits from Debug Mode.
The IDCORE instruction is requested.
In either Boundary Scan or ICE Mode a reset can be performed from the same or different cir-
cuitry, as shown in Figure 27 below, upon system reset at power-up or upon user request.
Reset
Controller NTRST NTRST
Reset
Controller NRST
Reset
Controller NRST
AT91RM9200 AT91RM9200
(1) (2)
Required Features The following table presents the features required of a reset controller in order to obtain an
for the Reset optimal system with the AT91RM9200 processor.
Controller
Table 34. Reset Controller Functions Synthesis
Feature Description
Power Supply Monitoring Overlaps the transient state of the system during power-up/down and brownout.
Reset Active Timeout Overlaps the start-up time of the boot-up oscillator by holding the reset signal during this delay.
Period
Manual Reset Command Asserts the reset signal from a logic command and holds the reset signal with a shorter delay than that
of the Reset Active Timeout Period.
121
1768BATARM08/03
122 AT91RM9200
1768BATARM08/03
AT91RM9200
Memory Controller(MC)
Overview The Memory Controller (MC) manages the ASB bus and controls access by up to four mas-
ters. It features a bus arbiter and an address decoder that splits the 4G bytes of address
space into areas to access the embedded SRAM and ROM, the embedded peripherals and
the external memories through the External Bus Interface (EBI). It also features an abort sta-
tus and a misalignment detector to assist in application debug.
The Memory Controller allows booting from the embedded ROM or from an external non-vola-
tile memory connected to the Chip Select 0 of the EBI. The Remap command switches
addressing of the ARM vectors (0x0 - 0x20) on the embedded SRAM.
Key Features of the RM9200 Memory Controller are:
Programmable Bus Arbiter Handling Four Masters
Internal Bus is Shared by ARM920T, PDC, USB Host Port and Ethernet MAC
Masters
Each Master Can Be Assigned a Priority Between 0 and 7
Address Decoder Provides Selection For
Eight External 256-Mbyte Memory Areas
Four Internal 1-Mbyte Memory Areas
One 256-Mbyte Embedded Peripheral Area
Boot Mode Select Option
Non-volatile Boot Memory Can Be Internal or External
Selection is Made By BMS Pin Sampled at Reset
Abort Status Registers
Source, Type and All Parameters of the Access Leading to an Abort are Saved
Misalignment Detector
Alignment Checking of All Data Accesses
Abort Generation in Case of Misalignment
Remap Command
Provides Remapping of an Internal SRAM in Place of the Boot NVM
123
1768BATARM08/03
Block Diagram
Figure 28. Memory Controller Block Diagram
Memory Controller
ASB
ARM920T
Processor Abort Internal
Abort
Status Memories
Address
Decoder
BMS
EMAC
DMA
Bus Misalignment
Arbiter Detector
External
UHP Bus
DMA Interface
User
Interface
Peripheral
Data
Controller Memory
APB Controller
Bridge Interrupt
Peripheral 0 AIC
Peripheral 1 APB
From Master
to Slave
Peripheral N
124 AT91RM9200
1768BATARM08/03
AT91RM9200
Functional The Memory Controller (MC) handles the internal ASB bus and arbitrates the accesses of up
Description to four masters.
It is made up of:
A bus arbiter
An address decoder
An abort status
A misalignment detector
The Memory Controller handles only little-endian mode accesses. All masters must work in lit-
tle-endian mode only.
Bus Arbiter The Memory Controller has a user-programmable bus arbiter. Each master can be assigned a
priority between 0 and 7, where 7 is the highest level. The bus arbiter is programmed in the
register MC_MPR (Master Priority Register).
The same priority level can be assigned to more than one master. If requests occur from two
masters having the same priority level, the following default priority is used by the bus arbiter
to determine the first to serve: Master 0, Master 1, Master 2, Master 3.
The masters are:
the ARM920T as the Master 0
the Peripheral Data Controller as the Master 1
the USB Host Port as the Master 2
the Ethernet MAC as the Master 3
Address Decoder The Memory Controller features an Address Decoder that first decodes the four highest bits of
the 32-bit address bus and defines 11 separate areas:
One 256-Mbyte address space for the internal memories
Eight 256-Mbyte address spaces, each assigned to one of the eight chip select lines of the
External Bus Interface
One 256-Mbyte address space reserved for the embedded peripherals
An undefined address space of 1536M bytes that returns an Abort if accessed
125
1768BATARM08/03
External Memory Figure 29 shows the assignment of the 256-Mbyte memory areas.
Areas
Figure 29. External Memory Areas
0x0000 0000
256M Bytes Internal Memories
0x0FFF FFFF
0x1000 0000
256M Bytes Chip Select 0
0x1FFF FFFF
0x2000 0000
256M Bytes Chip Select 1
0x2FFF FFFF
0x3000 0000
256M Bytes Chip Select 2
0x3FFF FFFF
0x4000 0000 EBI
256M Bytes Chip Select 3 External
0x4FFF FFFF
Bus
0x5000 0000 Interface
256M Bytes Chip Select 4
0x5FFF FFFF
0x6000 0000
256M Bytes Chip Select 5
0x6FFF FFFF
0x7000 0000
256M Bytes Chip Select 6
0x7FFF FFFF
0x8000 0000
256M Bytes Chip Select 7
0x8FFF FFFF
0x9000 0000
0xEFFF FFFF
0xF000 0000
256M Bytes Peripherals
0xFFFF FFFF
Internal Memory Within the Internal Memory address space, the Address Decoder of the Memory Controller
Mapping decodes eight more address bits to allocate 1-Mbyte address spaces for the embedded
memories.
The allocated memories are accessed all along the 1-Mbyte address space and so are
repeated n times within this address space, n equaling 1M byte divided by the size of the
memory.
When the address of the access is undefined within the internal memory area, i.e. over the
address 0x0040 0000, the Address Decoder returns an Abort to the master.
126 AT91RM9200
1768BATARM08/03
AT91RM9200
0x0FFF FFFF
Internal Memory Depending on the BMS pin state at reset and as a function of the remap command, the mem-
Area 0 ory mapped at address 0x0 is different. Before execution of the remap command the on-chip
ROM (BMS = 1) or the non-volatile memory connected to external chip select zero (BMS = 0)
is mapped into Internal Memory Area 0. After the remap command, the internal SRAM at
address 0x0020 0000 is mapped into Internal Memory Area 0. The memory mapped into Inter-
nal Memory Area 0 is accessible in both its original location and at address 0x0.
The first 32 bytes of Internal Memory Area 0 contain the ARM processor exception vectors.
Table 35. Internal Memory Area Depending on BMS and the Remap Command
Before Remap After Remap
BMS State 1 0 X
Internal Memory Area 0 Internal ROM External Memory Area 0 Internal SRAM
Boot Mode Select The BMS pin state allows the device to boot out of an internal or external boot memory, actu-
ally the input level on the BMS pin during the last 2 clock cycles, before the reset selects the
type of boot memory according to the following conditions:
If high, the Internal ROM, which is generally mapped within the Internal Memory Area 1,
is also accessible through the Internal Memory Area 0
If low, the External Memory Area 0, which is generally accessible from address 0x1000
0000, is also accessible through the Internal Memory Area 0.
The BMS pin is multiplexed with an I/O line. After reset, this pin can be used as any standard
PIO line.
Remap Command After execution, the Remap Command causes the Internal SRAM to be accessed through the
Internal Memory Area 0.
As the ARM vectors (Reset, Abort, Data Abort, Prefetch Abort, Undefined Instruction, Inter-
rupt, and Fast Interrupt) are mapped from address 0x0 to address 0x20, the Remap
Command allows the user to redefine dynamically these vectors under software control.
The Remap Command is accessible through the Memory Controller User Interface by writing
the MC_RCR (Remap Control Register) RCB field to one.
127
1768BATARM08/03
The Remap Command can be cancelled by writing the MC_RCR RCB field to one, which acts
as a toggling command. This allows easy debug of the user-defined boot sequence by offering
a simple way to put the chip in the same configuration as just after a reset.
Table 35 on page 127 is provided to summarize the effect of these two key features on the
nature of the memory mapped to the address 0x0.
Misalignment The Memory Controller features a Misalignment Detector that checks the consistency of the
Detector accesses.
For each access, regardless of the master, the size of access and the 0 and 1 bits of the
address bus are checked. If the type of access is a word (32-bit) and the 0 and 1 bits are not 0,
or if the type of the access is a half-word (16-bit) and the 0 bit is not 0, an abort is returned to
the master and the access is cancelled. Note that the accesses of the ARM processor when it
is fetching instructions are not checked.
128 AT91RM9200
1768BATARM08/03
AT91RM9200
The misalignments are generally due to software errors leading to wrong pointer handling.
These errors are particularly difficult to detect in the debug phase.
As the requested address is saved in the Abort Status and the address of the instruction gen-
erating the misalignment is saved in the Abort Link Register of the processor, detection and
correction of this kind of software error is simplified.
Memory Controller The Memory Controller itself does not generate any interrupt. However, as indicated in Figure
Interrupt 28, the Memory Controller receives an interrupt signal from the External Bus Interface, which
might be activated in case of Refresh Error detected by the SDRAM Controller. This interrupt
signal just transits through the Memory Controller, which can neither enable/disable it nor
return its activity.
This Memory Controller interrupt signal is ORed with the other System Peripheral interrupt
lines (RTC, ST, DBGU, PMC) to provide the System Interrupt on Source 1 of the Advanced
Interrupt Controller.
User Interface
Base Address: 0xFFFFFF00
129
1768BATARM08/03
MC Remap Control Register
Register Name: MC_RCR
Access Type: Write-only
Absolute Address: 0xFFFF FF00
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
RCB
130 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
MST3 MST2 MST1 MST0
15 14 13 12 11 10 9 8
ABTTYP ABTSZ
7 6 5 4 3 2 1 0
MISADD UNDADD
0 0 Byte
0 1 Half-word
1 0 Word
1 1 Reserved
0 0 Data Read
0 1 Data Write
1 0 Code Fetch
1 1 Reserved
131
1768BATARM08/03
MST2: UHP Abort Source
0: The last aborted access was not due to the UHP.
1: The last aborted access was due to the UHP.
MST3: EMAC Abort Source
0: The last aborted access was not due to the EMAC.
1: The last aborted access was due to the EMAC.
SVMST0: Saved ARM920T Abort Source
0: No abort due to the ARM920T occurred since the last read of MC_ASR or it is notified in the bit MST0.
1: At least one abort due to the ARM920T occurred since the last read of MC_ASR.
SVMST1: Saved PDC Abort Source
0: No abort due to the PDC occurred since the last read of MC_ASR or it is notified in the bit MST1.
1: At least one abort due to the PDC occurred since the last read of MC_ASR.
SVMST2: Saved UHP Abort Source
0: No abort due to the UHP occurred since the last read of MC_ASR or it is notified in the bit MST2.
1: At least one abort due to the UHP occurred since the last read of MC_ASR.
SVMST3: Saved EMAC Abort Source
0: No abort due to the EMAC occurred since the last read of MC_ASR or it is notified in the bit MST3.
1: At least one abort due to the EMAC occurred since the last read of MC_ASR.
132 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
ABTADD
15 14 13 12 11 10 9 8
ABTADD
7 6 5 4 3 2 1 0
ABTADD
133
1768BATARM08/03
MC Master Priority Register
Register Name: MC_MPR
Access Type: Read/Write
Reset Value: 0x3210
Absolute Address: 0xFFFF FF0C
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
MSTP3 MSTP2
7 6 5 4 3 2 1 0
MSTP1 MSTP0
134 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The External Bus Interface (EBI) is designed to ensure the successful data transfer between
several external devices and the embedded Memory Controller of an ARM-based device.
The Static Memory, SDRAM and Burst Flash Controllers are all featured external Memory
Controllers on the EBI. These external Memory Controllers are capable of handling several
types of external memory and peripheral devices, such as SRAM, PROM, EPROM, EEPROM,
Flash, SDRAM and Burst Flash.
The EBI also supports the CompactFlash and the SmartMedia protocols via integrated cir-
cuitry that greatly reduces the requirements for external components. Furthermore, the EBI
handles data transfers with up to eight external devices, each assigned to eight address
spaces defined by the embedded Memory Controller. Data transfers are performed through a
16-bit or 32-bit data bus, an address bus of up to 26 bits, up to eight chip select lines
(NCS[7:0]) and several control pins that are generally multiplexed between the different exter-
nal Memory Controllers.
Features of the EBI are:
Integrates Three External Memory Controllers:
Static Memory Controller
SDRAM Controller
Burst Flash Controller
Additional Logic for SmartMediaTM and CompactFlashTM Support
Optimized External Bus:
16- or 32-bit Data Bus
Up to 26-bit Address Bus, Up to 64-Mbytes Addressable
Up to 8 Chip Selects, Each Reserved to One of the Eight Memory Areas
Optimized Pin Multiplexing to Reduce Latencies on External Memories
Configurable Chip Select Assignment:
Burst Flash Controller or Static Memory Controller on NCS0
SDRAM Controller or Static Memory Controller on NCS1
Static Memory Controller on NCS3, Optional SmartMedia Support
Static Memory Controller on NCS4 - NCS6, Optional CompactFlash Support
Static Memory Controller on NCS7
135
1768BATARM08/03
Block Diagram Figure 31 below shows the organization of the External Bus Interface.
A[15:2], A[22:18]
A16/BA0
MUX
Burst Flash Logic A17/BA1
Controller
NCS0/BFCS
NCS1/SDCS
NCS2
NCS3/SMCS
Static NRD/NOE/CFOE
Memory
Controller NWR0/NWE/CFWE
NWR1/NBS1/CFIOR
NWR3/NBS3/CFIOW
SDCK
SDCKE
RAS
CAS
SDWE
SmartMedia
Logic SDA10
D[31:16]
A[24:23]
PIO
A25/CFRNW
CompactFlash
Logic NCS4/CFCS
NCS5/CFCE1
NCS6/CFCE2
NCS7
BFRDY/SMOE
User Interface
BFWE
NWAIT
APB
136 AT91RM9200
1768BATARM08/03
AT91RM9200
137
1768BATARM08/03
Table 37. I/O Lines Description (Continued)
Name Function Type Active Level
Burst Flash Controller
BFCK Burst Flash Clock Output
BFCS Burst Flash Chip Select Line Output Low
BFAVD Burst Flash Address Valid Signal Output Low
BFBAA Burst Flash Address Advance Signal Output Low
BFOE Burst Flash Output Enable Output Low
BFRDY Burst Flash Ready Signal Input High
BFWE Burst Flash Write Enable Output Low
The connection of some signals through the Mux logic is not direct and depends on the Mem-
ory Controller in use at the moment.
Table 38 below details the connections between the three Memory Controllers and the EBI
pins.
Table 38. EBI Pins and Memory Controllers I/O Line Connections
EBI Pins SDRAMC I/O Lines BFC I/O Lines SMC I/O Lines
NWR1/NBS1/CFIOR NBS1 Not Supported NWR1/NUB
A0/NBS0 Not Supported Not Supported A0/NLB
A1 Not Supported A0 A1
A[11:2] A[9:0] A[10:1] A[11:2]
SDA10 A10 Not Supported Not Supported
A12 Not Supported A11 A12
A[14:13] A[12:11] A[13:12] A[14:13]
A[25:15] Not Supported A[24:14] A[25:15]
D[31:16] D[31:16] Not Supported Not Supported
D[15:0] D[15:0] D[15:0] D[15:0]
138 AT91RM9200
1768BATARM08/03
AT91RM9200
Application
Example
Hardware Interface Table 39 below details the connections to be applied between the EBI pins and the external
devices for each Memory Controller.
139
1768BATARM08/03
Table 39. EBI Pins and External Device Connections (Continued)
Pins of the Interfaced Device
2 x 8-bit SmartMedia
8-bit Static Static 16-bit Static Burst Flash or NAND
Pin Device Devices Device Device SDRAM CompactFlash Flash
Controller SMC BFC SDRAMC SMC
NWR3/NBS3/CFIOW DQM3 IOW
BFCK CK
BFAVD AVD
BFBAA/SMWE BAA WE
BFOE OE
BFRDY/SMOE RDY OE
BFWE WE
SDCK CLK
SDCKE CKE
RAS RAS
CAS CAS
SDWE WE
NWAIT WAIT
(2)
Pxx CD1 or CD2
Pxx(2) CE
(2)
Pxx RDY
Notes: 1. Not directly connected to the CompactFlash slot. Permits the control of the bidirectional buffer between the EBI data bus and
the CompactFlash slot.
2. Any PIO line.
3. The REG signal of the CompactFlash can be driven by any of the following address bits: A24, A22 to A11. For details, see
CompactFlash Support on page 143.
4. The CLE and ALE signals of the SmartMedia device may be driven by any address bit. For details, see SmartMedia and
NAND Flash Support on page 146.
5. NWR1 enables upper byte writes. NWR0 enables lower byte writes.
140 AT91RM9200
1768BATARM08/03
AT91RM9200
Connection Figure 32 below shows an example of connections between the EBI and external devices.
Examples
EBI
D0-D31
RAS 2M x 8 2M x 8
CAS
SDCK
D0-D7
SDRAM D8-D15
SDRAM
SDCKE D0-D7 D0-D7
SDWE
CS
A0/NBS0 CS
NWR1/NBS1 CLK CLK
A1/NWR2/NBS2 CKE A0-A9, A11 A2-A11, A13 A0-A9, A11 A2-A11, A13
CKE
NWR3/NBS3 SDWE WE A10 SDA10 SDWE A10 SDA10
WE
NRD/NOE RAS BA0 A16/BA0 BA0 A16/BA0
RAS
NWR0/NWE CAS BA1 A17/BA1 BA1 A17/BA1
CAS
DQM
DQM
NBS0 NBS1
SDA10
A2-A15
A16/BA0
A17/BA1
A18-A25
2M x 8 2M x 8
D16-D23 D0-D7
SDRAM D24-D31
SDRAM
D0-D7
NCS0/BFCS
NCS1/SDCS CS CS
NCS2 CLK CLK
CKE A0-A9, A11 A2-A11, A13
NCS3 CKE A0-A9, A11
SDWE WE A10 SDA10 SDWE
NCS4 WE A10 A2-A11, A13
RAS BA0 A16/BA0
NCS5 RAS BA0 SDA10
CAS BA1 A17/BA1
NCS6 CAS BA1 A16/BA0
NCS7 DQM DQM A17/BA1
NBS3
NBS2
BFCLK 2M x 16
BFOE
BFWE Burst Flash
BFAVD
BFRDY D0-D15
D0-D15 A0-A20 A1-A21
CE
CLK
OE
WE
AVD
RDY
128K x 8 128K x 8
SRAM SRAM
A1-A17 A1-A17
D0-D7 D0-D7 A0-A16 D8-D15 D0-D7 A0-A16
CS CS
OE OE
NRD/NOE NRD/NOE
WE WE
A0/NWR0/NBS0 NWR1/NBS1
141
1768BATARM08/03
Product
Dependencies
I/O Lines The pins used for interfacing the External Bus Interface may be multiplexed with the PIO lines.
The programmer must first program the PIO controller to assign the External Bus Interface
pins to their peripheral function. If I/O lines of the External Bus Interface are not used by the
application, they can be used for other purposes by the PIO Controller.
Functional The EBI transfers data between the internal ASB Bus (handled by the Memory Controller) and
the external memories or peripheral devices. It controls the waveforms and the parameters of
Description
the external address, data and control busses and is composed of the following elements:
The Static Memory Controller (SMC)
The SDRAM Controller (SDRAMC)
The Burst Flash Controller (BFC)
A chip select assignment feature that assigns an ASB address space to the external
devices.
A multiplex controller circuit that shares the pins between the different Memory
Controllers.
Programmable CompactFlash support logic
Programmable SmartMedia and NAND Flash support logic
Bus Multiplexing The EBI offers a complete set of control signals that share the 32-bit data lines, the address
lines of up to 26 bits and the control signals through a multiplex logic operating in function of
the memory area requests.
Multiplexing is specifically organized in order to guarantee the maintenance of the address
and output control lines at a stable state while no external access is being performed. Multi-
plexing is also designed to respect the data float times defined in the Memory Controllers.
Furthermore, refresh cycles of the SDRAM are executed independently by the SDRAM Con-
troller without delaying the other external Memory Controller accesses. Lastly, it prevents
burst accesses on the same page of a burst Flash from being interrupted which avoids the
need to restart a high-latency first access.
Pull-up Control The EBI permits enabling of on-chip pull-up resistors on the data bus lines not multiplexed with
the PIO Controller lines. The pull-up resistors are enabled after reset. Setting the DBPUC bit
disables the pull-up resistors on the D0 to D15 lines. Enabling the pull-up resistor on the D16 -
D31 lines can be performed by programming the appropriate PIO controller.
Static Memory For information on the Static Memory Controller, refer to the SMC Overview on page 151.
Controller
SDRAM Controller For information on the SDRAM Controller, refer to the SDRAMC description on Overview on
page 135.
Burst Flash For information on the Burst Flash Controller, refer to the BFC Overview on page 209.
Controller
142 AT91RM9200
1768BATARM08/03
AT91RM9200
CompactFlash The External Bus Interface integrates circuitry that interfaces to CompactFlash devices.
Support The CompactFlash logic is driven by the Static Memory Controller (SMC) on the NCS4
address space. Programming the CS4A field of the Chip Select Assignment Register (See
EBI Chip Select Assignment Register on page 149.) to the appropriate value enables this
logic. Access to an external CompactFlash device is then made by accessing the address
space reserved to NCS4 (i.e., between 0x5000 0000 and 0x5FFF FFFF).
When multiplexed with CFCE1 and CFCE2 signals, the NCS5 and NCS6 signals become
unavailable. Performing an access within the address space reserved to NCS5 and NCS6
(i.e., between 0x6000 0000 and 0x7FFF FFFF) may lead to an unpredictable outcome.
The True IDE Mode is not supported and in I/O Mode, the signal _IOIS16 is not managed.
I/O Mode, Common Within the NCS4 address space, the current transfer address is used to distinguish I/O mode,
Memory Mode and common memory mode and attribute memory mode. More precisely, the A23 bit of the trans-
Attribute Memory fer address is used to select I/O Mode. Any EBI address bit not required by the CompactFlash
Mode device (i.e., bit A24 or bits A22 to A11) can be used to separate common memory mode and
attribute memory mode. Using the A22 bit, for example, leads to the address map in Figure 33
below. In this figure, i stands for any hexadecimal digit.
Note: In the above example, the A22 pin of the EBI can be used to drive the REG signal of the Com-
pactFlash Device.
Read/Write Signals In I/O mode, the CompactFlash logic drives the read and write command signals of the SMC
on CFIOR and CFIOW signals, while the CFOE and CFWE signals are deactivated. Likewise,
in common memory mode and attribute memory mode, the SMC signals are driven on the
CFOE and CFWE signals, while the CFIOR and CFIOW are deactivated. Figure 34 on page
144 demonstrates a schematic representation of this logic.
Attribute memory mode, common memory mode and I/O mode are supported by setting the
address setup and hold time on the NCS4 chip select to the appropriate values. For details on
these signal waveforms, please refer to the section: Setup and Hold Cycles on page 164 of
the Static Memory Controller documentation.
143
1768BATARM08/03
Figure 34. CompactFlash Read/Write Control Signals
A23
1
1 CFIOR
CFIOW
NRD_NOE
NWR0_NWE CFOE
1 CFWE
1
Access Type The CFCE1 and CFCE2 signals enable upper- and lower-byte access on the data bus of the
CompactFlash device in accordance with Table 40 below. The odd byte access on the D[7:0]
bus is only possible when the SMC is configured to drive 8-bit memory devices on the NCS4
pin. The Chip Select Register (DBW field in SMC Chip Select Registers on page 186) of the
NCS4 address space must be set as shown in Table 40 to enable the required access type.
The CFCE1 and CFCE2 waveforms are identical to the NCS4 waveform. For details on these
waveforms and timings, refer to the Static Memory Controller Overview on page 151.
Multiplexing of Table 41 below and Table 42 on page 145 illustrate the multiplexing of the CompactFlash logic
CompactFlash Signals signals with other EBI signals on the EBI pins. The EBI pins in Table 41 are strictly dedicated
on EBI Pins to the CompactFlash interface as soon as the CS4A field of the Chip Select Assignment Reg-
ister is set (See EBI Chip Select Assignment Register on page 149.). These pins must not be
used to drive any other memory devices.
The EBI pins in Table 42 on page 145 remain shared between all memory areas when the
CompactFlash interface is enabled (CS4A = 1).
144 AT91RM9200
1768BATARM08/03
AT91RM9200
CompactFlash Figure 35 below illustrates an example of a CompactFlash application. CFCS and CFRNW
Application Example signals are not directly connected to the CompactFlash slot, but do control the direction and
the output enable of the buffers between the EBI and the CompactFlash Device. The timing of
the CFCS signal is identical to the NCS4 signal. Moreover, the CFRNW signal remains valid
throughout the transfer, as does the address bus. The CompactFlash _WAIT signal is con-
nected to the NWAIT input of the Static Memory Controller. For details on these waveforms
and timings, refer to the Static Memory Controller Overview on page 135.
D[15:0] D[15:0]
DIR /OE
A25/CFRNW
NCS4/CFCS
_CD1
CD (PIO)
_CD2
/OE
A[10:0] A[10:0]
A22/REG _REG
NOE/CFOE _OE
NWE/CFWE _WE
NWR1/CFIOR _IORD
NWR3/CFIOW _IOWR
NCS5/CFE1 _CE1
NCS6/CFE2 _CE2
NWAIT _WAIT
145
1768BATARM08/03
SmartMedia and The EBI integrates circuitry that interfaces to SmartMedia and NAND Flash devices.
NAND Flash The SmartMedia logic is driven by the Static Memory Controller on the NCS3 address space.
Support Programming the CS3A field in the Chip Select Assignment Register to the appropriate value
enables the SmartMedia logic (See EBI Chip Select Assignment Register on page 149.).
Access to an external SmartMedia device is then made by accessing the address space
reserved to NCS3 (i.e., between 0x4000 0000 and 0x4FFF FFFF).
The SmartMedia Logic drives the read and write command signals of the SMC on the SMOE
and SMWE signals when the NCS3 signal is active. SMOE and SMWE are invalidated as
soon as the transfer address fails to lie in the NCS3 address space. For details on these wave-
forms, refer to the Static Memory Controller Overview on page 151.
The SMWE and SMOE signals are multiplexed with BFRDY and BFBAA signals of the Burst
Flash Controller. This multiplexing is controlled in the MUX logic part of the EBI by the CS3A
field of the Chip Select Assignment Register (See EBI Chip Select Assignment Register on
page 149.). This logic also controls the direction of the BFRDY/SMOE pad.
BFRDY BFRDY_SMOE
BFBAA
BFBAA_SMWE
NCS3 SMOE
NRD_NOE
SMWE
NWR0_NWE
The address latch enable and command latch enable signals on the SmartMedia device are
driven by address bits A22 and A21 of the EBI address bus. The user should note that any bit
on the EBI address bus can also be used for this purpose. The command, address or data
words on the data bus of the SmartMedia device are distinguished by using their address
within the NCS3 address space. The chip enable (CE) signal of the device and the ready/busy
(R/B) signals are connected to PIO lines. The CE signal then remains asserted even when
NCS3 is not selected, preventing the device from returning to standby mode. Some functional
limitation with the supported burst Flash device will occur when the SmartMedia device is acti-
vated due to the fact that the SMOE and SMWE signals are multiplexed with BFRDY and
BFBAA signals respectively.
146 AT91RM9200
1768BATARM08/03
AT91RM9200
D[7:0]
AD[7:0]
A[22:21]
ALE
CLE
NCS3/SMCS
Not Connected
EBI
SmartMedia
BFBAA/SMWE
NWE
BFRDY/SMOE
NOE
PIO CE
PIO R/B
147
1768BATARM08/03
External Bus Interface (EBI) User Interface
AT91RM9200 EBI User Interface Base Address: 0xFFFF FF60
148 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
CS4A CS3A CS1A CS0A
149
1768BATARM08/03
EBI Configuration Register
Register Name: EBI_CFGR
Access Type: Read/write
Reset Value: 0x0
Offset: 0x04
Absolute Address: 0xFFFF FF64
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
DBPUC
150 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Static Memory Controller (SMC) generates the signals that control the access to external
static memory or peripheral devices. The SMC is fully programmable and can address up to
512M bytes. It has eight chip selects and a 26-bit address bus. The 16-bit data bus can be
configured to interface with 8- or 16-bit external devices. Separate read and write control sig-
nals allow for direct memory and peripheral interfacing. The SMC supports different access
protocols allowing single clock cycle memory accesses. It also provides an external wait
request capability.
The main features of the SMC are:
External memory mapping, 512-Mbyte address space
Up to 8 Chip Select Lines
8- or 16-bit Data Bus
Remap of Boot Memory
Multiple Access Modes Supported
Byte Write or Byte Select Lines
Two different Read Protocols for each Memory Bank
Multiple Device Adaptability
Compliant with LCD Module
Programmable Setup Time Read/Write
Programmable Hold Time Read/Write
Multiple Wait State Management
Programmable Wait State Generation
External Wait Request
Programmable Data Float Time
151
1768BATARM08/03
Block Diagram Figure 38. Static Memory Controller Block Diagram
PIO
SMC Controller
Memory SMC
Chip Select NCS[7:0]
Controller
NRD/NOE
NWR0/NWE
NWR1/NUB
A0/NLB
MCK A[25:1]
PMC
D[15:0]
NWAIT
User Interface
APB
152 AT91RM9200
1768BATARM08/03
AT91RM9200
Application Example
Hardware Interface Figure 39 shows an example of static memory device connection to the SMC.
D0-D15
A0/NLB
NWR1/NUB
NWR0/NWE
NRD/NOE
128K x 8 128K x 8
SRAM SRAM
A1-A17 A1-A17
D0-D7 D0-D7 A0-A16 D8-D15 D0-D7 A0-A16
NCS0
NCS1
NCS2 CS CS
NCS3 NRD/NOE NRD/NOE
NCS4 OE OE
NWR0/NWE WE NWR1/NUB WE
NCS5
NCS6
NCS7
A1-A25
Static
Memory
Controller
Product Dependencies
I/O Lines The pins used for interfacing the Static Memory Controller may be multiplexed with the PIO
lines. The programmer must first program the PIO controller to assign the Static Memory Con-
troller pins to their peripheral function. If I/O lines of the Static Memory Controller are not used
by the application, they can be used for other purposes by the PIO Controller.
153
1768BATARM08/03
Functional Description
External Memory
Interface
External Memory The memory map is defined by hardware and associates the internal 32-bit address space
Mapping with the external 26-bit address bus. Note that A[25:0] is only significant for 8-bit memory.
A[25:1] is used for 16-bit memory. If the physical memory device is smaller than the page size,
it wraps around and appears to be repeated within the page. The SMC correctly handles any
valid access to the memory device within the page. See Figure 40.
Chip Select Lines The Static Memory Controller provides up to eight chip select lines: NCS0 to NCS7.
NCS[7:0]
NCS7
Memory Enable
NRD NCS6
Memory Enable
SMC NWR[1:0] NCS5
Memory Enable
A[25:0] NCS4
Memory Enable
D[15:0] NCS3
Memory Enable
NCS2
Memory Enable
NCS1
Memory Enable
NCS0
Memory Enable
Output Enable
Write Enable
A[25:0]
8 or 16
D[15:0] or D[7:0]
154 AT91RM9200
1768BATARM08/03
AT91RM9200
Data Bus Width A data bus width of 8 or 16 bits can be selected for each chip select. This option is controlled
by the DBW field in the SMC_CSR for the corresponding chip select. See SMC Chip Select
Registers on page 186.
Figure 42 shows how to connect a 512K x 8-bit memory on NCS2 (DBW = 10).
D[7:0] D[7:0]
D[15:8]
A[25:1] A[25:1]
SMC A0 A0
NWR1
NWR0 Write Enable
NRD Output Enable
NCS2 Memory Enable
Figure 43 shows how to connect a 512K x 16-bit memory on NCS2 (DBW = 01).
D[7:0] D[7:0]
D[15:8] D[15:8]
A[25:1] A[24:0]
155
1768BATARM08/03
Write Access
Write Access Type Each chip select with a 16-bit data bus can operate with one of two different types of write
access:
Byte Write Access supports two byte write and a single read signal.
Byte Select Access selects upper and/or lower byte with two byte select lines, and
separate read and write signals.
This option is controlled by the BAT field in the SMC_CSR for the corresponding chip select.
See SMC Chip Select Registers on page 186.
Byte Write Access Byte Write Access is used to connect 2 x 8-bit devices as a 16-bit memory page.
The signal A0/NLB is not used.
The signal NWR1/NUB is used as NWR1 and enables upper byte writes.
The signal NWR0/NWE is used as NWR0 and enables lower byte writes.
The signal NRD/NOE is used as NRD and enables half-word and byte reads.
Figure 44 shows how to connect two 512K x 8-bit devices in parallel on NCS2 (BAT = 0)
D[7:0] D[7:0]
D[15:8]
A[24:1] A[18:0]
SMC A0
NWR1
NWR0 Write Enable
NRD Read Enable
NCS2 Memory Enable
D[15:8]
A[18:0]
Write Enable
Read Enable
Memory Enable
Byte Select Access Byte Select Access is used to connect 16-bit devices in a memory page.
The signal A0/NLB is used as NLB and enables the lower byte for both read and write
operations.
The signal NWR1/NUB is used as NUB and enables the upper byte for both read and write
operations.
The signal NWR0/NWE is used as NWE and enables writing for byte or half-word.
The signal NRD/NOE is used as NOE and enables reading for byte or half-word.
156 AT91RM9200
1768BATARM08/03
AT91RM9200
Figure 45 shows how to connect a 16-bit device with byte and half-word access (e.g., SRAM
device type) on NCS2 (BAT = 1).
Figure 45. Connection to a 16-bit Data Path Device with Byte and Half-word Access
D[7:0] D[7:0]
D[15:8] D[15:8]
A[19:1] A[18:0]
Figure 46 shows how to connect a 16-bit device without byte access (e.g., Flash device type)
on NCS2 (BAT = 1).
Figure 46. Connection to a 16-bit Data Path Device without Byte Write Capability
D[7:0] D[7:0]
D[15:8] D[15:8]
A[19:1] A[18:0]
SMC NLB
NUB
NWE Write Enable
NOE Output Enable
157
1768BATARM08/03
Write Data Hold Time During write cycles, data output becomes valid after the rising edge of MCK and remains valid
after the rising edge of NWE. During a write access, the data remain on the bus 1/2 period of
MCK after the rising edge of NWE. See Figure 47 and Figure 48.
MCK
A[25:0]
NCS2
NWE
D[15:0]
MCK
A[25:0]
NCS2
NWE
D[15:0]
158 AT91RM9200
1768BATARM08/03
AT91RM9200
Read Access
Read Protocols The SMC provides two alternative protocols for external memory read accesses: standard and
early read. The difference between the two protocols lies in the behavior of the NRD signal.
For write accesses, in both protocols, NWE has the same behavior. In the second half of the
master clock cycle, NWE always goes low (see Figure 56 on page 164).
The protocol is selected by the DRP field in SMC_CSR (See SMC Chip Select Registers on
page 186.). Standard read protocol is the default protocol after reset.
Note: In the following waveforms and descriptions, NRD represents NRD as well as NOE since the
two signals have the same waveform. Likewise, NWE represents NWE, NWR0 and NWR1
unless NWR0 and NWR1 are otherwise represented. In addition, NCS represents NCS[7:0]
(see I/O Lines on page 153, Table 44 and Table 45).
Standard Read Protocol Standard read protocol implements a read cycle during which NRD and NWE are similar. Both
are active during the second half of the clock cycle. The first half of the clock cycle allows time
to ensure completion of the previous access as well as the output of address lines and NCS
before the read cycle begins.
During a standard read protocol, NCS is set low and address lines are valid at the beginning of
the external memory access, while NRD goes low only in the second half of the master clock
cycle to avoid bus conflict. See Figure 49.
MCK
A[25:0]
NCS
NRD
D[15:0]
159
1768BATARM08/03
Early Read Protocol Early read protocol provides more time for a read access from the memory by asserting NRD
at the beginning of the clock cycle. In the case of successive read cycles in the same memory,
NRD remains active continuously. Since a read cycle normally limits the speed of operation of
the external memory system, early read protocol can allow a faster clock frequency to be
used. However, an extra wait state is required in some cases to avoid contentions on the
external bus.
MCK
A[25:0]
NCS
NRD
D[15:0]
160 AT91RM9200
1768BATARM08/03
AT91RM9200
Wait State The SMC can automatically insert wait states. The different types of wait states managed are
Management listed below:
Standard wait states
External wait states
Data float wait states
Chip select change wait states
Early Read wait states
Standard Wait States Each chip select can be programmed to insert one or more wait states during an access on
the corresponding memory area. This is done by setting the WSEN field in the corresponding
SMC_CSR (See SMC Chip Select Registers on page 186.). The number of cycles to insert is
programmed in the NWS field in the same register.
Below is the correspondence between the number of standard wait states programmed and
the number of clock cycles during which the NWE pulse is held low:
0 wait states 1/2 clock cycle
1 wait state 1 clock cycle
For each additional wait state programmed, an additional cycle is added.
MCK
A[25:0]
NCS
NWE
External Wait States The NWAIT input pin is used to insert wait states beyond the maximum standard wait states
programmable or in addition to. If NWAIT is asserted low, then the SMC adds a wait state and
no changes are made to the output signals, the internal counters or the state. When NWAIT is
de-asserted, the SMC completes the access sequence.
The input of the NWAIT signal is an asynchronous input. To avoid any metastability problems,
NWAIT is synchronized before using it. This operation results in a two-cycle delay.
NWS must be programmed as a function of synchronization time and delay between NWAIT
falling and control signals falling (NRD/NWE), otherwise SMC will not function correctly.
NWS > Wait Delay from nrd/nwe + external_nwait Synchronization Delay + 1
If NWAIT is asserted during a setup or hold timing, the SMC does not function correctly.
161
1768BATARM08/03
Figure 52. NWAIT behaviour in Read Access
MCK
A[25:0]
NWAIt
NWAIT
internally synchronized
NRD
NCS
(2)
(1) NWAIT
Synchronization Delay
Wait Delay from NRD
MCK
A[25:0]
NWAIT
NWAIT
internally synchronized
NWE
D[15:0]
Data Float Wait States Some memory devices are slow to release the external bus. For such devices, it is necessary
to add wait states (data float wait states) after a read access before starting a write access or
a read access to a different external memory.
The Data Float Output Time (tDF) for each external memory device is programmed in the TDF
field of the SMC_CSR register for the corresponding chip select (See SMC Chip Select Reg-
isters on page 186.). The value of TDF indicates the number of data float wait cycles
(between 0 and 15) to be inserted and represents the time allowed for the data output to go to
high impedance after the memory is disabled.
Data float wait states do not delay internal memory accesses. Hence, a single access to an
external memory with long tDF will not slow down the execution of a program from internal
memory.
162 AT91RM9200
1768BATARM08/03
AT91RM9200
To ensure that the external memory system is not accessed while it is still busy, the SMC
keeps track of the programmed external data float time during internal accesses.
Internal memory accesses and consecutive read accesses to the same external memory do
not add data float wait states.
MCK
A[25:0]
NCS
D[15:0]
Chip Select Change A chip select wait state is automatically inserted when consecutive accesses are made to two
Wait State different external memories (if no other type of wait state has already been inserted). If a wait
state has already been inserted (e.g., data float wait state), then no more wait states are
added.
MCK
NCS1
NCS2
NWE
163
1768BATARM08/03
Early Read Wait State In early read protocol, an early read wait state is automatically inserted when an external write
cycle is followed by a read cycle to allow time for the write cycle to end before the subsequent
read cycle begins (see Figure 56). This wait state is generated in addition to any other pro-
grammed wait states (i.e., data float wait state).
No wait state is added when a read cycle is followed by a write cycle, between consecutive
accesses of the same type, or between external and internal memory accesses.
MCK
A[25:0]
NCS
NRD
NWE
D[15:0]
Setup and Hold The SMC allows some memory devices to be interfaced with different setup, hold and pulse
Cycles delays. These parameters are programmable and define the timing of each portion of the read
and write cycles. However, it is not possible to use this feature in early read protocol.
If an attempt is made to program the setup parameter as not equal to zero and the hold
parameter as equal to zero with WSEN = 0 (0 standard wait state), the SMC does not operate
correctly.
If consecutive accesses are made to two different external memories and the second memory
is programmed with setup cycles, then no chip select change wait state is inserted (see Figure
61 on page 166).
When a data float wait state (tDF) is programmed on the first memory bank and when the sec-
ond memory bank is programmed with setup cycles, the SMC behaves as follows:
If the number of tDF is higher or equal to the number of setup cycles, the number of setup
cycles inserted is equal to 0 (see Figure 62 on page 167).
If the number of the setup cycle is higher than the number of tDF, the number of tDF inserted
is 0 (see Figure 63 on page 167).
Read Access The read cycle can be divided into a setup, a pulse length and a hold. The setup parameter
can have a value between 1.5 and 7.5 clock cycles, the hold parameter between 0 and 7 clock
cycles and the pulse length between 1.5 and 128.5 clock cycles, by increments of one.
164 AT91RM9200
1768BATARM08/03
AT91RM9200
MCK
A[25:0]
NRD
MCK
A[25:0]
NRD
Write Access The write cycle can be divided into a setup, a pulse length and a hold. The setup parameter
can have a value between 1.5 and 7.5 clock cycles, the hold parameter between 0.5 and 7
clock cycles and the pulse length between 1 and 128 clock cycles by increments of one.
MCK
A[25:0]
NWE
D[15:0]
165
1768BATARM08/03
Figure 60. Write Access with Setup
MCK
A[25:0]
NWE
D[15:0]
Figure 61. Consecutive Accesses with Setup Programmed on the Second Access
Setup
MCK
A[25:0]
NCS1
NCS2
NWE
NRD
166 AT91RM9200
1768BATARM08/03
AT91RM9200
Figure 62. First Access with Data Float Wait States (TDF = 2) and Second Access with Setup (NRDSETUP = 1)
Setup
MCK
A[25:0]
NCS1
NCS2
NRD
D[15:0]
Figure 63. First Access with Data Float Wait States (TDF = 2) and Second Access with Setup (NRDSETUP = 3)
Setup
MCK
A[25:0]
NCS1
NCS2
NRD
D[15:0]
167
1768BATARM08/03
LCD Interface The SMC can be configured to work with an external liquid crystal display (LCD) controller by
Mode setting the ACSS (Address to Chip Select Setup) bit in the SMC_CSR registers (See SMC
Chip Select Registers on page 186.).
In LCD mode, NCS is shortened by one/two/three clock cycles at the leading and trailing
edges, providing positive address setup and hold. For read accesses, the data is latched in
the SMC when NCS is raised at the end of the access.
Additionally, WSEN must be set and NWS programmed with a value of two or more superior to
ACSS. In LCD mode, it is not recommended to use RWHOLD or RWSETUP. If the above con-
ditions are not satisfied, SMC does not operate correctly.
MCK
A[25:0]
NRD
NCS
ACSS ACSS
MCK
A[25:0]
NWE
NCS
ACCS ACCS
D[15:0]
168 AT91RM9200
1768BATARM08/03
AT91RM9200
Read Accesses in Figure 66 on page 169 through Figure 69 on page 172 show examples of the alternatives for
Standard and Early external memory read protocol.
Protocols
MCK
A[25:0]
NRD
NWE
NCS1
Chip Select
Change Wait
NCS2
D[15:0] (Mem 1)
tWHDX tWHDX
D[15:0] (Mem 2)
169
1768BATARM08/03
Figure 67. Early Read Protocol without tDF
Read Write Early Read Read Read Write Early Read Read
Mem 1 Mem 1 Wait Cycle Mem 1 Mem 2 Mem 2 Wait Cycle Mem 2
MCK
A[25:0]
NRD
NWE
NCS1
Chip Select
Change Wait
NCS2
D[15:0] (Mem 1)
D[15:0] (Mem 2)
170 AT91RM9200
1768BATARM08/03
AT91RM9200
MCK
A[25:0]
NRD
NWE
NCS1
NCS2
tDF tDF
(tDF = 1) (tDF = 1)
D[15:0]
(Mem 1)
D[15:0]
(to write)
tWHDX tDF
(tDF = 2)
D[15:0]
(Mem 2)
171
1768BATARM08/03
Figure 69. Early Read Protocol with tDF
Write Early Read Write Write
Read Mem 1 Mem 1 Read Wait Read Mem 1 Mem 2 Read Mem 2 Mem 2 Mem 2
Data Data Data
Float Wait Float Wait Float Wait
MCK
A[25:0]
NRD
NWE
NCS1
NCS2
tDF tDF
D[15:0]
(Mem 1)
D[15:0]
(to write)
tDF (tDF = 2)
D[15:0]
(Mem 2)
172 AT91RM9200
1768BATARM08/03
AT91RM9200
Accesses with Setup Figure 70 and Figure 71 show an example of read and write accesses with Setup and Hold
and Hold Cycles.
Figure 70. Read Accesses in Standard Read Protocol with Setup and Hold(1)
MCK
A0/NLB
NRD/NOE
NWR0/NWE
NWR1/NUB
Setup Hold Setup Hold
NCS
Note: 1. Read access memory data bus width = 8, RWSETUP = 1, RWHOLD = 1,WSEN= 1, NWS = 0
MCK
A0/NLB
NRD/NOE
NWR0/NWE
NWR1/NUB
NCS
Note: 1. Write access, memory data bus width = 8, RWSETUP = 1, RWHOLD = 1, WSEN = 1, NWS = 0
173
1768BATARM08/03
Accesses Using Figure 72 on page 174 through Figure 75 on page 177 show examples of accesses using
NWAIT Input Signal NWAIT.
Figure 72. Write Access using NWAIT in Byte Select Type Access(1)
Chip Select
Wait
MCK
NWAIT
NWAIT
internally
synchronized
A[25:1] 000008A
NRD/NOE
NWR0/NWE
A0/NLB
NWR1/NUB
NCS
D[15:0] 1312
Note: 1. Write access memory, data bus width = 16 bits, WSEN = 1, NWS = 6
174 AT91RM9200
1768BATARM08/03
AT91RM9200
Figure 73. Write Access using NWAIT in Byte Write Type Access(1)
Chip Select
Wait
MCK
NWAIT
NWAIT
internally
synchronized
A[25:1] 000008C
A0/NLB
NRD/NOE
NWR0/NWE
NWR1/NUB
NCS
D[15:0] 1716
Note: 1. Write access memory, data bus width = 16 bits, WSEN = 1, NWS = 5
175
1768BATARM08/03
Figure 74. Write Access using NWAIT(1)
Chip Select
Wait
MCK
NWAIT
NWAIT
internally
synchronized
A[25:1] 0000033
A0/NLB
NRD/NOE
NWR0/NWE
NWR1/NUB
NCS
D[15:0] 0403
Note: 1. Write access memory, data bus width = 8 bits, WSEN = 1, NWS = 4
176 AT91RM9200
1768BATARM08/03
AT91RM9200
MCK
NWAIT
NWAIT
internally
synchronized
A[25:1] 0002C44
A0/NLB
NRD/NOE
NWR0/NWE
NWR1/NUB
NCS
0003
D[15:0]
Note: 1. Read access, memory data bus width = 16, NWS = 5, WSEN = 1
Memory Access Figure 76 on page 178 through Figure 82 on page 184 show the waveforms for read and write
Example Waveforms accesses to the various associated external memory devices. The configurations described
are shown in Table 46.
177
1768BATARM08/03
Figure 76. 0 Wait State, 16-bit Bus Width, Word Transfer
MCK
NCS
NLB
NUB
Read Access
D[15:0] B 2 B1 B 4 B3
D[15:0] B 2 B1 B4 B3
Write Access
Byte Write/
Byte Select Option NWE
D[15:0] B2 B1 B 4 B3
178 AT91RM9200
1768BATARM08/03
AT91RM9200
MCK
NCS
NLB
NUB
Read Access
NRD
D[15:0] B2 B1 B4 B3
Write Access
Byte Write/
Byte Select Option
NWE
179
1768BATARM08/03
Figure 78. 1 Wait State, 16-bit Bus Width, Half-Word Transfer
1 Wait State
MCK
A[25:1]
NCS
NLB
NUB
Read Access
D[15:0]
B2 B 1
D[15:0] B 2 B1
Write Access
Byte Write/
Byte Select Option
NWE
D[15:0] B 2 B1
180 AT91RM9200
1768BATARM08/03
AT91RM9200
MCK
NCS
Read Access
D[15:0] X B1 X B2 X B3 X B4
D[15:0] X B1 X B2 X B3 X B4
Write Access
NWR0
NWR1
D[15:0] X B1 X B2 X B3 X B4
181
1768BATARM08/03
Figure 80. 1 Wait State, 8-bit Bus Width, Half-Word Transfer
MCK
NCS
Read Access
D[15:0] X B1 X B2
D[15:0] X B1 X B2
Write Access
NWR0
NWR1
D[15:0] X B1 X B2
182 AT91RM9200
1768BATARM08/03
AT91RM9200
MCK
A[25:0]
NCS
Read Access
D[15:0] XB1
D[15:0] X B1
Write Access
NWR0
NWR1
D[15:0] X B1
183
1768BATARM08/03
Figure 82. 0 Wait State, 16-bit Bus Width, Byte Transfer
MCK
NCS
NLB
NUB
Read Access
D[15:0] X B1 B2X
Write Access
NWR0
NWR1
184 AT91RM9200
1768BATARM08/03
AT91RM9200
185
1768BATARM08/03
SMC Chip Select Registers
Register Name: SMC_CSR0..SMC_CSR7
Access Type: Read/write
Reset Value: See Table 47 on page 185
31 30 29 28 27 26 25 24
RWHOLD RWSETUP
23 22 21 20 19 18 17 16
ACSS
15 14 13 12 11 10 9 8
DRP DBW BAT TDF
7 6 5 4 3 2 1 0
WSEN NWS
186 AT91RM9200
1768BATARM08/03
AT91RM9200
RWSETUP NRD Setup NWR Setup RWHOLD NRD Hold NWR Hold
(1)
0 0 0 cycle or cycle 0 0 0 0 cycle
0 cycles(2)
0 0 1 1 + cycles 1 + cycles 0 0 1 1 cycles 1 cycle
0 1 0 2 + cycles 2 + cycles 0 1 0 2 cycles 2 cycles
0 1 1 3 + cycles 3 + cycles 0 1 1 3 cycles 3 cycles
1 0 0 4 + cycles 4 + cycles 1 0 0 4 cycles 4 cycles
1 0 1 5 + cycles 5 + cycles 1 0 1 5 cycles 5 cycles
1 1 0 6 + cycles 6 + cycles 1 1 0 6 cycles 6 cycles
1 1 1 7 + cycles 7 + cycles 1 1 1 7 cycles 7 cycles
Notes: 1. In Standard Read Protocol.
2. In Early Read Protocol. (It is not possible to use the parameters RWSETUP or RWHOLD in this mode)
187
1768BATARM08/03
Figure 83. Read/write Setup
MCK
A[25:0]
NRD/NOE
NWE
RWSETUP
MCK
A[25:0]
NRD/NOE
RWHOLD
MCK
A[25:0]
NWE
D[15:0]
RWHOLD
188 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The SDRAM Controller (SDRAMC) extends the memory capabilities of a chip by providing the
interface to an external 16-bit or 32-bit SDRAM device. The page size supports ranges from
2048 to 8192 and the number of columns from 256 to 2048. It supports byte (8-bit), half-word
(16-bit) and word (32-bit) accesses.
The SDRAM Controller supports a read or write burst length of one location. It does not sup-
port byte read/write bursts or half-word write bursts. It keeps track of the active row in each
bank, thus maximizing SDRAM performance, e.g., the application may be placed in one bank
and data in the other banks. So as to optimize performance, it is advisable to avoid accessing
different rows in the same bank.
Features of the SDRAMC are:
Numerous Configurations Supported
2K, 4K, 8K Row Address Memory Parts
SDRAM with Two or Four Internal Banks
SDRAM with 16- or 32-bit Data Path
Programming Facilities
Word, Half-word, Byte Access
Automatic Page Break When Memory Boundary Has Been Reached
Multibank Ping-pong Access
Timing Parameters Specified by Software
Automatic Refresh Operation, Refresh Rate is Programmable
Energy-saving Capabilities
Self-refresh and Low-power Modes Supported
Error Detection
Refresh Error Interrupt
SDRAM Power-up Initialization by Software
Latency is Set to Two Clocks (CAS Latency of 1, 3 Not Supported)
Auto Precharge Command Not Used
189
1768BATARM08/03
Block Diagram Figure 86. SDRAM Controller Block Diagram
SDRAMC PIO
Controller
SDRAMC
Chip Select SDCK
Memory SDCKE
Controller
SDRAMC SDCS
Interrupt
BA[1:0]
RAS
MCK CAS
PMC
SDWE
NBS[3:0]
A[12:0]
D[31:0]
User Interface
APB
I/O Lines
Description
Table 48. I/O Line Description
Name Description Type Active Level
SDCK SDRAM Clock Output
SDCKE SDRAM Clock Enable Output High
SDCS SDRAM Controller Chip Select Output Low
BA[1:0] Bank Select Signals Output
RAS Row Signal Output Low
CAS Column Signal Output Low
SDWE SDRAM Write Enable Output Low
NBS[3:0] Data Mask Enable Signals Output Low
A[12:0] Address Bus Output
D[31:0] Data Bus I/O
190 AT91RM9200
1768BATARM08/03
AT91RM9200
Application
Example
Hardware Interface Figure 87 shows an example of SDRAM device connection to the SDRAM Controller by using
a 32-bit data bus width. Figure 88 shows an example of SDRAM device connection by using a
16-bit data bus width. Care should be taken, as these examples are given for a direct connec-
tion of the devices to the SDRAM Controller, without External Bus Interface, nor PIO Controller
multiplexing.
Figure 87. SDRAM Controller Connections to SDRAM Devices: 32-bit Data Bus Width
D0-D31
RAS
CAS 2M x 8 2M x 8
SDCK
SDCKE D0-D7 SDRAM D8-D15 SDRAM
D0-D7 D0-D7
SDWE
NBS0 CS
CS
NBS1 CLK CLK
NBS2 CKE A0-A9, A11 A0-A9, A11 A0-A9, A11 A0-A9, A11
CKE
NBS3 SDWE WE A10 A10 SDWE A10 A10
WE
RAS BA0 BA0 BA0 BA0
RAS
CAS BA1 BA1 BA1 BA1
CAS
DQM
DQM
NBS0 NBS1
A0-A12
BA0
BA1
2M x 8 2M x 8
D16-D23 SDRAM D24-D31 SDRAM
D0-D7 D0-D7
SDCS
CS CS
CLK CLK
CKE A0-A9, A11 A0-A9, A11
CKE A0-A9, A11
SDWE WE A10 A10 SDWE
WE A10 A0-A9, A11
RAS BA0 BA0
RAS BA0 A10
CAS BA1 BA1
CAS BA1 BA0
DQM DQM BA1
SDRAM NBS2 NBS3
Controller
Figure 88. SDRAM Controller Connections to SDRAM Devices: 16-bit Data Bus Width
D0-D31
RAS
CAS 2M x 8 2M x 8
SDCK
SDCKE D0-D7
SDRAM D8-D15 SDRAM
D0-D7 D0-D7
SDWE
NBS0 CS
CS
NBS1 CLK CLK
CKE A0-A9, A11 A0-A9, A11 A0-A9, A11 A0-A9, A11
CKE
SDWE WE A10 A10 SDWE A10 A10
WE
RAS BA0 BA0 BA0 BA0
RAS
CAS BA1 BA1 BA1 BA1
CAS
DQM
DQM
NBS0 NBS1
A0-A12
BA0
BA1
SDRAM SDCS
Controller
191
1768BATARM08/03
Software Interface The SDRAM Controllers function is to make the SDRAM device access protocol transparent
to the user. Table 49 to Table 54 illustrate the SDRAM device memory mapping therefore
seen by the user in correlation with the device structure. Various configurations are illustrated.
27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
192 AT91RM9200
1768BATARM08/03
AT91RM9200
27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
27 26 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
193
1768BATARM08/03
Product
Dependencies
SDRAM Devices The initialization sequence is generated by software. The SDRAM devices are initialized by
Initialization the following sequence:
1. A minimum pause of 200 s is provided to precede any signal toggle.
2. An All Banks Precharge command is issued to the SDRAM devices.
3. Eight auto-refresh (CBR) cycles are provided.
4. A mode register set (MRS) cycle is issued to program the parameters of the SDRAM
devices, in particular CAS latency and burst length.
5. A Normal Mode command is provided, 3 clocks after tMRD is met.
6. Write refresh rate into the count field in the SDRAMC Refresh Timer register. (Refresh
rate = delay between refresh cycles).
After these six steps, the SDRAM devices are fully functional.
The commands (NOP, MRS, CBR, normal mode) are generated by programming the com-
mand field in the SDRAMC Mode register
CK
:0]
10
11]
CS
AS
AS
WE
BS
Inputs Stable for Precharge All Banks 1st Auto-refresh 8th Auto-refresh MRS Command Valid Command
200 sec
194 AT91RM9200
1768BATARM08/03
AT91RM9200
I/O Lines The pins used for interfacing the SDRAM Controller may be multiplexed with the PIO lines.
The programmer must first program the PIO controller to assign the SDRAM Controller pins to
their peripheral function. If I/O lines of the SDRAM Controller are not used by the application,
they can be used for other purposes by the PIO Controller.
Interrupt The SDRAM Controller interrupt (Refresh Error notification) is connected to the Memory Con-
troller. This interrupt may be ORed with other System Peripheral interrupt lines and is finally
provided as the System Interrupt Source (Source 1) to the AIC (Advanced Interrupt
Controller).
Using the SDRAM Controller interrupt requires the AIC to be programmed first.
Functional
Description
SDRAM Controller The SDRAM Controller allows burst access or single access. To initiate a burst access, the
Write Cycle SDRAM Controller uses the transfer type signal provided by the master requesting the access.
If the next access is a sequential write access, writing to the SDRAM device is carried out. If
the next access is a write-sequential access, but the current access is to a boundary page, or
if the next access is in another row, then the SDRAM Controller generates a precharge com-
mand, activates the new row and initiates a write command. To comply with SDRAM timing
parameters, additional clock cycles are inserted between precharge/active (tRP) commands
and active/write (tRCD) commands. For definition of these timing parameters, refer to the
SDRAMC Configuration Register on page 204. This is described in Figure 90 below.
SDCS
SDCK
A[12:0] Row n col a col b col c col d col e col f col g col h col i col j col k col l
RAS
CAS
SDWE
D[31:0] Dna Dnb Dnc Dnd Dne Dnf Dng Dnh Dni Dnj Dnk Dnl
195
1768BATARM08/03
SDRAM Controller The SDRAM Controller allows burst access or single access. To initiate a burst access, the
Read Cycle SDRAM Controller uses the transfer type signal provided by the master requesting the access.
If the next access is a sequential read access, reading to the SDRAM device is carried out. If
the next access is a sequential read access, but the current access is to a boundary page, or if
the next access is in another row, then the SDRAM Controller generates a precharge com-
mand, activates the new row and initiates a read command. To comply with SDRAM timing
parameters, an additional clock cycle is inserted between the precharge/active (tRP) command
and the active/read (tRCD) command, After a read command, additional wait states are gener-
ated to comply with cas latency. The SDRAM Controller supports a cas latency of two. For
definition of these timing parameters, refer to SDRAMC Configuration Register on page 204.
This is described in Figure 91 below.
tRCD = 3 CAS = 2
SDCS
SDCK
RAS
CAS
SDWE
196 AT91RM9200
1768BATARM08/03
AT91RM9200
Border When the memory row boundary has been reached, an automatic page break is inserted. In
Management this case, the SDRAM controller generates a precharge command, activates the new row and
initiates a read or write command. To comply with SDRAM timing parameters, an additional
clock cycle is inserted between the precharge/active (tRP) command and the active/read (tRCD)
command. This is described in Figure 92 below.
SDCS
SDCK
Row n
A[12:0] col a col b col c col d Row m col a col b col c col d col e
RAS
CAS
SDWE
D[31:0] Dna Dnb Dnc Dnd Dma Dmb Dmc Dmd Dme
197
1768BATARM08/03
SDRAM Controller An auto-refresh command is used to refresh the SDRAM device. Refresh addresses are gen-
Refresh Cycles erated internally by the SDRAM device and incremented after each auto-refresh automatically.
The SDRAM Controller generates these auto-refresh commands periodically. A timer is
loaded with the value in the register SDRAMC_TR that indicates the number of clock cycles
between refresh cycles.
A refresh error interrupt is generated when the previous auto-refresh command did not per-
form. It will be acknowledged by reading the Interrupt Status Register (SDRAMC_ISR).
When the SDRAM Controller initiates a refresh of the SDRAM device, internal memory
accesses are not delayed. However, if the CPU tries to access the SDRAM, the slave will indi-
cate that the device is busy and the ARM BWAIT signal will be asserted. See Figure 93 below.
SDCS
SDCK
Row n
A[12:0] col c col d Row m col a
RAS
CAS
SDWE
D[31:0] Dma
Dnb Dnc Dnd
(input)
198 AT91RM9200
1768BATARM08/03
AT91RM9200
Power
Management
Self-refresh Mode Self-refresh mode is used in power-down mode, i.e., when no access to the SDRAM device is
possible. In this case, power consumption is very low. The mode is activated by programming
the self-refresh command bit (SRCB) in SDRAMC_SRR. In self-refresh mode, the SDRAM
device retains data without external clocking and provides its own internal clocking, thus per-
forming its own auto-refresh cycles. All the inputs to the SDRAM device become dont care
except SDCKE, which remains low. As soon as the SDRAM device is selected, the SDRAM
Controller provides a sequence of commands and exits self-refresh mode, so the self-refresh
command bit is disabled.
To re-activate this mode, the self-refresh command bit must be re-programmed.
The SDRAM device must remain in self-refresh mode for a minimum period of tRAS and may
remain in self-refresh mode for an indefinite period. This is described in Figure 94 below.
A[12:0] Row
SDCK
SDCKE
SDCS
RAS
CAS
SDWE
Access Request
to the SDRAM Controller
199
1768BATARM08/03
Low-power Mode Low-power mode is used in power-down mode, i.e., when no access to the SDRAM device is
possible. In this mode, power consumption is greater than in self-refresh mode. This state is
similar to normal mode (No low-power mode/No self-refresh mode), but the SDCKE pin is low
and the input and output buffers are deactivated as soon as the SDRAM device is no longer
accessible. In contrast to self-refresh mode, the SDRAM device cannot remain in low-power
mode longer than the refresh period (64 ms for a whole device refresh operation). As no auto-
refresh operations are performed in this mode, the SDRAM Controller carries out the refresh
operation. In order to exit low-power mode, a NOP command is required. The exit procedure is
faster than in self-refresh mode.
When self-refresh mode is enabled, it is recommended to avoid enabling low-power mode.
When low-power mode is enabled, it is recommended to avoid enabling self-refresh mode.
This is described in Figure 95 below.
SDCS
SDCK
RAS
CAS
SDCKE
D[31:0]
Dna Dnb Dnc Dnd Dne Dnf
(input)
200 AT91RM9200
1768BATARM08/03
AT91RM9200
201
1768BATARM08/03
SDRAMC Mode Register
Register Name: SDRAMC_MR
Access Type: Read/Write
Reset Value: 0x00000010
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
DBW MODE
MODE Description
0 0 0 0 Normal mode. Any access to the SDRAM is decoded normally.
0 0 0 1 The SDRAM Controller issues a NOP command when the SDRAM device is accessed regardless of the
cycle.
0 0 1 0 The SDRAM Controller issues an All Banks Precharge command when the SDRAM device is accessed
regardless of the cycle.
0 0 1 1 The SDRAM Controller issues a Load Mode Register command when the SDRAM device is accessed
regardless of the cycle. The address offset with respect to the SDRAM device base address is used to
program the Mode Register. For instance, when this mode is activated, an access to the SDRAM_Base +
offset address generates a Load Mode Register command with the value offset written to the SDRAM
device Mode Register.
0 1 0 0 The SDRAM Controller issues a Refresh Command when the SDRAM device is accessed regardless of
the cycle. Previously, an All Banks Precharge command must be issued.
202 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
COUNT
7 6 5 4 3 2 1 0
COUNT
203
1768BATARM08/03
SDRAMC Configuration Register
Register Name: SDRAMC_CR
Access Type: Read/Write
Reset Value: 0x2A99C140
31 30 29 28 27 26 25 24
TXSR TRAS
23 22 21 20 19 18 17 16
TRAS TRCD TRP
15 14 13 12 11 10 9 8
TRP TRC TWR
7 6 5 4 3 2 1 0
TWR CAS NB NR NC
NC Column Bits
0 0 8
0 1 9
1 0 10
1 1 11
NR Row Bits
0 0 11
0 1 12
1 0 13
1 1 Reserved
NB Number of Banks
0 2
1 4
204 AT91RM9200
1768BATARM08/03
AT91RM9200
205
1768BATARM08/03
SDRAMC Self-refresh Register
Register Name: SDRAMC_SRR
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
SRCB
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
LPCB
206 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
RES
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
RES
207
1768BATARM08/03
SDRAMC Interrupt Mask Register
Register Name: SDRAMC_IMR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
RES
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
RES
208 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Burst Flash Controller (BFC) provides an interface for external 16-bit Burst Flash devices
and handles an address space of 256M bytes. It supports byte, half-word and word aligned
accesses and can access up to 32M bytes of Burst Flash devices. The BFC also supports
data bus and address bus multiplexing. The Burst Flash interface supports only continuous
burst reads. Programmable burst lengths of four or eight words are not possible. The BFC
never generates an abort signal, regardless of the requested address within the 256M bytes of
address space.
The BFC can operate with two burst read protocols depending on whether or not the address
increment of the Burst Flash device is signal controlled. The Burst Flash Controller Mode Reg-
ister (BFC_MR) located in the BFC user interface is used in programming Asynchronous or
Burst Operating Modes. In Burst Mode, the read protocol, Clock Controlled Address Advance,
automatically increments the address at each clock cycle. Whereas in Signal Controlled
Address Advance protocol the address is incremented only when the Burst Address Advance
signal is active. When Address and Data Bus Multiplexing Mode is chosen, the sixteen lowest
address bits are multiplexed with the data bus.
The BFC clock speed is programmable to be either master clock or master clock divided by 2
or 4. Page size handling (16 bytes to 1024 bytes) is required by some Burst Flash devices
unable to handle continuous burst read. The number of latency cycles after address valid goes
up to sixteen cycles. The number of latency cycles after output enable runs between one and
three cycles. The Burst Flash Controller can also be programmed to suspend and maintain the
current burst. This attribute gives other devices the possibility to share the BFC busses without
any loss of efficiency. In Burst Mode, the BFC can restart a sequential access without any
additional latency.
Features of the Burst Flash Controller are:
Multiple Access Modes Supported
Asynchronous or Burst Mode Byte, Half-word or Word Read Accesses
Asynchronous Mode Half-word Write Accesses
Adaptability to Different Device Speed Grades
Programmable Burst Flash Clock Rate
Programmable Data Access Time
Programmable Latency after Output Enable
Adaptability to Different Device Access Protocols and Bus Interfaces
Two Burst Read Protocols: Clock Control Address Advance or Signal Controlled
Address Advance
Multiplexed or Separate Address and Data Busses
Continuous Burst and Page Mode Accesses Supported
209
1768BATARM08/03
Block Diagram
Figure 96. Burst Flash Controller Block Diagram
PIO
BFC Controller
Memory BFC
Chip Select BFCK
Controller
BFCS
BFAVD
BFBAA
BFOE
MCK BFWE
PMC
BFRDY
A[24:0]
D[15:0]
User Interface
APB
I/O Lines
Description
Table 56. I/O Lines Description
Name Description Type Active Level
BFCK Burst Flash Clock Output
BFCS Burst Flash Chip Select Output Low
BFAVD Burst Flash Address Valid Output Low
BFBAA Burst Flash Address Advance Output Low
BFOE Burst Flash Output Enable Output Low
BFWE Burst Flash Write Enable Output Low
BFRDY Burst Flash Ready Input High
A[24:0] Address Bus Output
D[15:0] Data Bus I/O
210 AT91RM9200
1768BATARM08/03
AT91RM9200
Application Example
Burst Flash The Burst Flash Interface provides control, address and data signals to the Burst Flash Mem-
Interface ory. These signals are detailed in the Functional Description on page 212 which describes
the BFC functionality and operating modes. Figure 97 below presents an illustration of the
possible connections of the BFC to some popular Burst Flash Memories.
AT49SN6416A Am29DBS643D
AT49SN3208A Am29BDS323D
BFC BFC
AT49BN1604 AT49BP1604
BFC BFC
[D0:D15] [D0:D15] [D0:D15] [AD0:AD15]
[A0:A19] [A0:A19] [A16:A19] [A16:A19]
BFCK clk BFCK clk
BFCS ce BFCS ce
BFAVD avd/lba BFAVD avd
BFWE we BFWE we
211
1768BATARM08/03
Product Dependencies
Supported Burst The Burst Flash Controller is designed to preferentially support the following ATMEL Burst
Flash Devices Flash devices:
AT49SN6416A and AT49SN6416AT (64 Mbits x 16)
AT49SN3208A and AT49SN3208AT (32 Mbits x 16)
AT49BN3208 and AT49BN3208T (32 Mbits x 16)
I/O Lines The pins used for interfacing the Burst Flash Controller may be multiplexed with the PIO lines.
The programmer must first program the PIO controller to assign the Burst Flash Controller
pins to their peripheral function. If I/O lines of the Burst Flash Controller are not used by the
application, they can be used for other purposes by the PIO Controller.
Burst Flash After reset, the BFC is disabled and, therefore, must be enabled by programming the field
Controller Reset BFCOM. See Burst Flash Controller Mode Register on page 221. At this time, the Burst
State Flash Controller operates in Asynchronous Mode. The Burst Flash memory can be pro-
grammed by writing and reading in Asynchronous Mode.
Burst Flash The BFC clock rate is programmable to be either Master Clock, Master Clock divided by 2 or
Controller Clock Master Clock divided by 4. The clock selection is necessary in Burst Mode as well as in Asyn-
Selection chronous Mode. The latency fields in the mode register and all burst Flash control signal
waveforms are related to the Burst Flash Clock (BFCK) period.
The BFC clock rate is selected by the BFCC field. See Burst Flash Controller Mode Register
on page 221.
212 AT91RM9200
1768BATARM08/03
AT91RM9200
Burst Flash In Asynchronous Mode, the Burst Flash Controller clock is off. The BFCK signal is driven low.
Controller The BFC performs read access to bytes (8-bits), half-words (16-bits), and words (32-bits). In
Asynchronous the last case, the BFC autonomously transforms the word read request into two separate half-
Mode word reads. This is fully transparent to the user.
The BFC performs only half-word write requests. Write requests for bytes or words are ignored
by the BFC.
For any access in the address space, the address is driven on the address bus while a pulse is
driven on the BFAVD signal (see Figure 99 on page 214, and Figure 100 on page 215). The
Burst Flash address is also driven on the data bus if the multiplexed data and address bus
options are enabled. (Figure 99 on page 214).
For write access, the signal BFWE is asserted in the following BFCK clock cycle.
For read access, the signal BFOE is asserted one cycle later. This additional cycle in read
accesses has been inserted to switch the I/O pad direction so as to avoid conflict on the
Burst Flash data bus when address and data busses are multiplexed.
The Address Valid Latency (AVL) determines the length of the pulses as a number of Master
Clock cycles. The AVL field (See Burst Flash Controller Mode Register on page 221.) is
coded as the Address Valid Latency minus 1. Waveforms in Figure 99 on page 214 and Figure
100 on page 215 show the AVL field definition in read and write accesses.
In read access, the access finishes with the rising edge of BFOE.
In write access, data and address lines are released one half cycle after the rising edge of
BFWE.
After a read access to the Burst Flash, it takes Output Enable Latency (OEL) cycles for the
Burst Flash device to release the data bus. The OEL field (See Burst Flash Controller Mode
Register on page 221.) gives the OEL expressed in BFCK Clock cycles. This prevents other
memory controllers from using the Data Bus until it is released by the Burst Flash device.
In Figure 99 on page 214 (multiplexed address and data busses), one idle cycle (OEL = 1) is
inserted between the read and write accesses. The Burst Flash device must release the data
bus before the BFC can drive the address. As shown in Figure 100 on page 215, where bus-
ses are not multiplexed, the write access can start as soon as the read access ends. In the
same way, the OEL has no impact when a read follows a write access.
Waveforms in Figure 99 on page 214 below and Figure 100 on page 215 are related to the
Burst Flash Controller Clock even though the BFCK pin is driven low in Asynchronous Mode.
The BFCC field (See Burst Flash Controller Mode Register on page 221.) is used as a mea-
sure of the burst Flash speed and must also be programmed in Asynchronous Mode.
213
1768BATARM08/03
Figure 99. Asynchronous Read and Write Accesses with Multiplexed Address and Data Buses
BFCS
BFCK
BFAVD
AVL
BFOE
AVL
BFWE
D[15:0]
Data
Input
OEL = 1
Asynchronous Asynchronous
Read Access Write Access
Address Valid Latency = 4 BFCK cycles (AVL field = 3)
Output Enable Latency (OEL) = 1 BFCK cycle
214 AT91RM9200
1768BATARM08/03
AT91RM9200
Figure 100. Asynchronous Read and Write Accesses with Non-multiplexed Address and Data
BFCS
BFCK
BFAVD
AVL
BFOE
AVL
BFWE
D[15:0] Data
Output
D[15:0]
Data
Input
OEL = 1
Asynchronous Asynchronous
Read Access Write Access
Burst Flash Writing the Burst Flash Controller Operating Mode field (BFCOM) to 2 (see Burst Flash Con-
Controller troller Mode Register on page 221) puts the BFC in Burst Mode. The BFC Clock is driven on
Synchronous the BFCK pin. Only read accesses are treated and write accesses are ignored. The BFC sup-
ports read access of bytes, half-words or words.
Mode
Burst Read Protocols The BFC supports two burst read protocols:
Clock Controlled Address Advance, the internal address of the burst Flash is automatically
incremented at each BFCK cycle.
Signal Controlled Address Advance, the internal address of the burst Flash is incremented
only when the BFBAA signal is active.
Read Access in Burst When a read access is requested in Burst Mode, the requested address is registered in the
Mode BFC. For subsequent read accesses, the address is compared to the previous one. Then the
two following cases are considered:
1. In case of a non-sequential access, the current burst is broken and the BFC launches a
new burst by performing an address latch cycle. The address is presented on the
address bus in any case and on the data bus if the multiplexed bus option is enabled.
215
1768BATARM08/03
This new address is registered in the BFC and is then used as reference for further
accesses.
2. In case of sequential access, and provided that the BFOEH mode is selected in the
mode register (See Burst Flash Controller Mode Register on page 221.), the internal
burst address is incremented:
Through the BFBAA pin, if the Signal Controlled Address Advance is enabled.
By enabling the clock during one clock cycle in Clock Controlled Address Advance
Mode.
These protocols are illustrated in Figure 101 below and Figure 102 on page 217. The Address
Valid Latency (AVL+1, see Burst Flash Controller Mode Register on page 221) gives the
number of cycles from the first rising clock edge when BFAVD is asserted to the rising edge
that causes the read of data D1.
Note: This rising edge is also used to latch D0 in the BFC.
Figure 101. Burst Suspend and Resume with Signal Control Address Advance
BFCS
Internal BFC
Selection Signal
BFCK
BFAVD
AVL OEL = 2
BFOE
BFWE
Burst Suspend Burst Resume
BFBAA
D[15:0]
Address (1)
Output
D[15:0]
D0 D1 D2 D3 D4 D4 D5 D6
Input
D0 D2 D5
Sampling Sampling Sampling
D1 D3 D4
Sampling Sampling Sampling
Burst Suspend and Resume (BFOEH = 1) Address Valid Latency = 4 BFCK cycles (AVL field = 3)
Signal Control Address Advance (BAAEN = 1) Output Enable Latency (OEL) = 2 BFCK cycles
(1) Only if Multiplexed Address & Data Buses
216 AT91RM9200
1768BATARM08/03
AT91RM9200
Figure 102. Burst Suspend and Resume with Clock Control Address Advance
Burst Suspend Burst Resume
BFCS
Internal BFC
Selection Signal
BFCK
BFAVD
AVL OEL = 2
BFOE
BFWE
D[15:0] D0 D1 D2 D3 D4 D4 D5 D6
Input
D0 D2 D5
Sampling Sampling Sampling
D1 D3 D4
Sampling Sampling Sampling
Burst Suspend and Resume (BFOEH = 1) Address Valid Latency = 4 BFCK cycles (AVL = 3)
Clock Control Address Advance (BAAEN = 0) Output Enable Latency (OEL) = 2 BFCK cycles
(1) Only if Multiplexed Address & Data Buses
Burst Suspension for The BFC can suspend a burst to enable other internal transfers, or other memory controllers
Transfer Enabling to use the memory address and data busses if they are shared. Two modes are provided on
the BFOEH bit (Burst Flash Output Enable Handling, see Burst Flash Controller Mode Regis-
ter on page 221):
BFOEH = 1: the BFC suspends the burst when it is no longer selected and the BFOE pin
is deasserted. When a new sequential access on the Burst Flash device is requested, the
burst is resumed and the BFOE pin is asserted again. The data is available on the data
bus after OEL cycles. This mode provides a minimal access latency. (Refer to Figure 101
on page 216 and Figure 102 above).
BFOEH = 0: the BFC suspends the burst when it is no longer selected and the BFOE pin
is deasserted. When a new access to the Burst Flash device is requested, either
sequential or not, a new burst is initialized and the next data is available as defined by the
AVL latency field in the Mode Register. This mode is provided for Burst Flash devices for
which the deassertion of the BFOE signal causes an irreversible break of the burst. Figure
103 on page 218 shows the access request to the BFC and the deassertion of the BFOE
signal due to a deselection of the BFC (Suspend). When the BFC is requested again, a
new burst is started even though the requested address is sequential to the previously
requested address.
217
1768BATARM08/03
Figure 103. Burst Flash Controller with No Burst Enable Handling
Internal
Clock (2)
Internal A0 A1 A2 A3
Address Bus
Burst Suspend Begin New Burst
Internal BFC
Selection Signal
BFCS
BFCK
BFAVD
AVL OEL = 1 AVL
BFOE
BFWE
D[15:0]
Input D0 D1 D2 D2 D3 D4
BFBAA
D0 D3
Sampling Sampling
Address Valid Latency = 4 BFCK cycles
Output Enable Latency (OEL) = 1 BFCK cycle
D1 D2
Sampling Sampling
(1) Only if Multiplexed Address & Data Busses No Burst Output Enable Handling (BFOEH = 0)
(2) Master Clock Mode (BFCC =1) Signal Control Advance Address (BAAEN = 1)
Continuous Burst The BFC performs continuous burst reads. It is also possible to program page sizes from 16
Reads bytes up to 1024 bytes. This is done by setting the appropriate value in the PAGES field of the
Burst Flash Controller Mode Register on page 221.
Page Mode In Page Mode, the BFC stops the current burst and starts a new burst each time the requested
address matches a page boundary. Figure 104 on page 219 illustrates a 16-byte page size.
Data D0 to D10 belong to two separate pages and are accessed through two burst accesses.
This mode is provided for Burst Flash devices that cannot handle continuous burst read (in
which case, a continuous burst access to address D0 would cause the Burst Flash internal
218 AT91RM9200
1768BATARM08/03
AT91RM9200
address to wrap around address D0). Page Mode can be disabled by programming a null
value in the PAGES field of the Burst Flash Controller Mode Register on page 221.
BFCK ..
BFAVD
AVL AVL
(1)
BFOE
BFWE
D[15:0]
Output
D[15:0] ..
D0 D1 D6 D7 D0 D8 D9 D10
Input
BFBAA
D0 D7 D8 (1)
Sampling Sampling Sampling
Burst Read in Page Mode (16 Bytes) Address Valid Latency = 3 BFCK cycles (AVL field = 2)
Signal Control Advance Address (BAAEN = 1) Output Enable Latency (OEL) = 1 BFCK cycle
(1) A New Page Begins at D8 Page Size = 16 Bytes
Ready Enable Mode In Ready Enable Mode (bit RDYEN in the Burst Flash Controller Mode Register on page
221), the BFC uses the Ready Signal (BFRDY) from the burst Flash device as an indicator of
the next data availability. The BFRDY signal must be asserted one BFCK cycle before data is
valid. In Figure 105 on page 220 below, the BFRDY signal indicates on edge (A) that the
expected D4 data will not be available on the next rising BFCK edge. The BFRDY signal
remains low until rising at edge (B). D4 is then sampled on edge (C).
When the RDYEN mode is disabled (RDYEN = 0), the BFRDY signal at the BFC input inter-
face is ignored. This mode is provided for Burst Flash devices that do not handle the BFRDY
signal.
219
1768BATARM08/03
Figure 105. Burst Read Using BFRDY Signal
BFCK
BFAVD
AVL
BFOE
D[15:0]
Input D0 D1 D2 D3 D4 D5 D6 D7
BFBAA
BFRDY
Sampling D0 D1 D2 D3 D4 D5 D6 D7
220 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
RDYEN MUXEN BFOEH BAAEN
15 14 13 12 11 10 9 8
OEL PAGES
7 6 5 4 3 2 1 0
AVL BFCC BFCOM
221
1768BATARM08/03
PAGES: Page Size
This field defines the page size handling and the page size.
222 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Peripheral Data Controller (PDC) transfers data between on-chip serial peripherals such
as the UART, USART, SSC, SPI, MCI and the on- and off-chip memories. Using the Peripheral
Data Contoller avoids processor intervention and removes the processor interrupt-handling
overhead.This significantly reduces the number of clock cycles required for a data transfer
and, as a result, improves the performance of the microcontroller and makes it more power
efficient.
The PDC channels are implemented in pairs, each pair being dedicated to a particular periph-
eral. One channel in the pair is dedicated to the receiving channel and one to the transmitting
channel of each UART, USART, SSC and SPI.
The user interface of a PDC channel is integrated in the memory space of each peripheral. It
contains:
A 32-bit memory pointer register
A 16-bit transfer count register
A 32-bit register for next memory pointer
A 16-bit register for next transfer count
The peripheral triggers PDC transfers using transmit and receive signals. When the pro-
grammed data is transferred, an end of transfer interrupt is generated by the corresponding
peripheral.
Important features of the PDC are:
Generates Transfers to/from Peripherals Such as DBGU, USART, SSC, SPI and MCI
Supports Up to Twenty Channels (Product Dependent)
One Master Clock Cycle Needed for a Transfer from Memory to Peripheral
Two Master Clock Cycles Needed for a Transfer from Peripheral to Memory
Memory
RHR PDC Channel 1 Control
Controller
223
1768BATARM08/03
Functional
Description
Configuration The PDC channels user interface enables the user to configure and control the data transfers
for each channel. The user interface of a PDC channel is integrated into the user interface of
the peripheral (offset 0x100), which it is related to.
Per peripheral, it contains four 32-bit Pointer Registers (RPR, RNPR, TPR, and TNPR) and
four 16-bit Counter Registers (RCR, RNCR, TCR, and TNCR).
The size of the buffer (number of transfers) is configured in an internal 16-bit transfer counter
register, and it is possible, at any moment, to read the number of transfers left for each
channel.
The memory base address is configured in a 32-bit memory pointer by defining the location of
the first address to access in the memory. It is possible, at any moment, to read the location in
memory of the next transfer and the number of remaining transfers. The PDC has dedicated
status registers which indicate if the transfer is enabled or disabled for each channel. The sta-
tus for each channel is located in the peripheral status register. Transfers can be enabled
and/or disabled by setting TXTEN/TXTDIS and RXTEN/RXTDIS in PDC Transfer Control
Register. These control bits enable reading the pointer and counter registers safely without
any risk of their changing between both reads.
The PDC sends status flags to the peripheral visible in its status-register (ENDRX, ENDTX,
RXBUFF, and TXBUFE).
ENDRX flag is set when the PERIPH_RCR register reaches zero.
RXBUFF flag is set when both PERIPH_RCR and PERIPH_RNCR reach zero.
ENDTX flag is set when the PERIPH_TCR register reaches zero.
TXBUFE flag is set when both PERIPH_TCR and PERIPH_TNCR reach zero.
These status flags are described in the peripheral status register.
Memory Pointers Each peripheral is connected to the PDC by a receiver data channel and a transmitter data
channel. Each channel has an internal 32-bit memory pointer. Each memory pointer points to
a location anywhere in the memory space (on-chip memory or external bus interface memory).
Depending on the type of transfer (byte, half-word or word), the memory pointer is incre-
mented by 1, 2 or 4, respectively for peripheral transfers.
If a memory pointer is reprogrammed while the PDC is in operation, the transfer address is
changed, and the PDC performs transfers using the new address.
Transfer Counters There is one internal 16-bit transfer counter for each channel used to count the size of the
block already transferred by its associated channel. These counters are decremented after
each data transfer. When the counter reaches zero, the transfer is complete and the PDC
stops transferring data.
If the Next Counter Register is equal to zero, the PDC disables the trigger while activating the
related peripheral end flag.
If the counter is reprogrammed while the PDC is operating, the number of transfers is updated
and the PDC counts transfers from the new value.
Programming the Next Counter/Pointer registers chains the buffers. The counters are decre-
mented after each data transfer as stated above, but when the transfer counter reaches zero,
224 AT91RM9200
1768BATARM08/03
AT91RM9200
the values of the Next Counter/Pointer are loaded into the Counter/Pointer registers in order to
re-enable the triggers.
For each channel, two status bits indicate the end of the current buffer (ENDRX, ENTX) and
the end of both current and next buffer (RXBUFF, TXBUFE). These bits are directly mapped to
the peripheral status register and can trigger an interrupt request to the AIC.
The peripheral end flag is automatically cleared when one of the counter-registers (Counter or
Next Counter Register) is written.
Note: When the Next Counter Register is loaded into the Counter Register, it is set to zero.
Data Transfers The peripheral triggers PDC transfers using transmit (TXRDY) and receive (RXRDY) signals.
When the peripheral receives an external character, it sends a Receive Ready signal to the
PDC which then requests access to the system bus. When access is granted, the PDC starts
a read of the peripheral Receive Holding Register (RHR) and then triggers a write in the
memory.
After each transfer, the relevant PDC memory pointer is incremented and the number of trans-
fers left is decremented. When the memory block size is reached, a signal is sent to the
peripheral and the transfer stops.
The same procedure is followed, in reverse, for transmit transfers.
Priority of PDC The Peripheral Data Controller handles transfer requests from the channel according to priori-
Transfer Requests ties fixed for each product.These priorities are defined in the product datasheet.
If simultaneous requests of the same type (receiver or transmitter) occur on identical peripher-
als, the priority is determined by the numbering of the peripherals.
If transfer requests are not simultaneous, they are treated in the order they occurred.
Requests from the receivers are handled first and then followed by transmitters requests.
225
1768BATARM08/03
Peripheral Data Controller (PDC) User Interface
Table 57. Register Mapping
Offset Register Register Name Read/Write Reset
0x100 PDC Receive Pointer Register PERIPH(1)_RPR Read/Write 0x0
0x104 PDC Receive Counter Register PERIPH_RCR Read/Write 0x0
0x108 PDC Transmit Pointer Register PERIPH_TPR Read/Write 0x0
0x10C PDC Transmit Counter Register PERIPH_TCR Read/Write 0x0
0x110 PDC Receive Next Pointer Register PERIPH_RNPR Read/Write 0x0
0x114 PDC Receive Next Counter Register PERIPH_RNCR Read/Write 0x0
0x118 PDC Transmit Next Pointer Register PERIPH_TNPR Read/Write 0x0
0x11C PDC Transmit Next Counter Register PERIPH_TNCR Read/Write 0x0
0x120 PDC Transfer Control Register PERIPH_PTCR Write-only -
0x114 PDC Transfer Status Register PERIPH_PTSR Read-only 0x0
Note: 1. PERIPH: Ten registers are mapped in the peripheral memory space at the same offset. These can be defined by the user
according to the function and the peripheral desired (DBGU, USART, SSC, SPI, MCI etc).
23 22 21 20 19 18 17 16
RXPTR
15 14 13 12 11 10 9 8
RXPTR
7 6 5 4 3 2 1 0
RXPTR
226 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
--
15 14 13 12 11 10 9 8
RXCTR
7 6 5 4 3 2 1 0
RXCTR
23 22 21 20 19 18 17 16
TXPTR
15 14 13 12 11 10 9 8
TXPTR
7 6 5 4 3 2 1 0
TXPTR
23 22 21 20 19 18 17 16
--
15 14 13 12 11 10 9 8
TXCTR
7 6 5 4 3 2 1 0
TXCTR
227
1768BATARM08/03
PDC Receive Next Pointer Register
Register Name: PERIPH_RNPR
Access Type: Read/Write
31 30 29 28 27 26 25 24
RXNPTR
23 22 21 20 19 18 17 16
RXNPTR
15 14 13 12 11 10 9 8
RXNPTR
7 6 5 4 3 2 1 0
RXNPTR
23 22 21 20 19 18 17 16
--
15 14 13 12 11 10 9 8
RXNCR
7 6 5 4 3 2 1 0
RXNCR
23 22 21 20 19 18 17 16
TXNPTR
15 14 13 12 11 10 9 8
TXNPTR
7 6 5 4 3 2 1 0
TXNPTR
228 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
--
15 14 13 12 11 10 9 8
TXNCR
7 6 5 4 3 2 1 0
TXNCR
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
TXTDIS TXTEN
7 6 5 4 3 2 1 0
RXTDIS RXTEN
229
1768BATARM08/03
PDC Transfer Status Register
Register Name: PERIPH_PTSR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
TXTEN
7 6 5 4 3 2 1 0
RXTEN
230 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Advanced Interrupt Controller (AIC) is an 8-level priority, individually maskable, vectored
interrupt controller, providing handling of up to thirty-two interrupt sources. It is designed to
substantially reduce the software and real-time overhead in handling internal and external
interrupts.
The AIC drives the nFIQ (fast interrupt request) and the nIRQ (standard interrupt request)
inputs of an ARM processor. Inputs of the AIC are either internal peripheral interrupts or exter-
nal interrupts coming from the product's pins.
The 8-level Priority Controller allows the user to define the priority for each interrupt source,
thus permitting higher priority interrupts to be serviced even if a lower priority interrupt is being
treated.
Internal interrupt sources can be programmed to be level sensitive or edge triggered. External
interrupt sources can be programmed to be positive-edge or negative-edge triggered or high-
level or low-level sensitive.
The fast forcing feature redirects any internal or external interrupt source to provide a fast
interrupt rather than a normal interrupt.
Important Features of the AIC are:
Controls the Interrupt Lines (nIRQ and nFIQ) of an ARM Processor
Thirty-two Individually Maskable and Vectored Interrupt Sources
Source 0 is Reserved for the Fast Interrupt Input (FIQ)
Source 1 is Reserved for System Peripherals (ST, RTC, PMC, DBGU)
Source 2 to Source 31 Control up to Thirty Embedded Peripheral Interrupts or
External Interrupts
Programmable Edge-triggered or Level-sensitive Internal Sources
Programmable Positive/Negative Edge-triggered or High/Low Level-sensitive
External Sources
8-level Priority Controller
Drives the Normal Interrupt of the Processor
Handles Priority of the Interrupt Sources 1 to 31
Higher Priority Interrupts Can Be Served During Service of Lower Priority Interrupt
Vectoring
Optimizes Interrupt Service Routine Branch and Execution
One 32-bit Vector Register per Interrupt Source
Interrupt Vector Register Reads the Corresponding Current Interrupt Vector
Protect Mode
Easy Debugging by Preventing Automatic Operations when Protect ModeIs Are
Enabled
Fast Forcing
Permits Redirecting any Normal Interrupt Source on the Fast Interrupt of the
Processor
General Interrupt Mask
Provides Processor Synchronization on Events Without Triggering an Interrupt
231
1768BATARM08/03
Block Diagram Figure 107. Block Diagram
FIQ AIC
ARM
IRQ0-IRQn Processor
Up to
Thirty-two nFIQ
Embedded Sources
PeripheralEE
Embedded nIRQ
Peripheral
Embedded
Peripheral
APB
Standalone
Applications OS Drivers RTOS Drivers
Hard Real Time Tasks
General OS Interrupt Handler
External Peripherals
Embedded Peripherals
(External Interrupts)
APB
232 AT91RM9200
1768BATARM08/03
AT91RM9200
I/O Line
Table 58. I/O Line Description
Description
Pin Name Pin Description Type
FIQ Fast Interrupt Input
IRQ0 - IRQn Interrupt 0 - Interrupt n Input
Product Dependencies
I/O Lines The interrupt signals FIQ and IRQ0 to IRQn are normally multiplexed through the PIO control-
lers. Depending on the features of the PIO controller used in the product, the pins must be
programmed in accordance with their assigned interrupt function. This is not applicable when
the PIO controller used in the product is transparent on the input path.
Power The Advanced Interrupt Controller is continuously clocked. The Power Management Controller
Management has no effect on the Advanced Interrupt Controller behavior.
The assertion of the Advanced Interrupt Controller outputs, either nIRQ or nFIQ, wakes up the
ARM processor while it is in Idle Mode. The General Interrupt Mask feature enables the AIC to
wake up the processor without asserting the interrupt line of the processor, thus providing syn-
chronization of the processor on an event.
Interrupt Sources The Interrupt Source 0 is always located at FIQ. If the product does not feature an FIQ pin, the
Interrupt Source 0 cannot be used.
The Interrupt Source 1 is always located at System Interrupt. This is the result of the OR-wir-
ing of the system peripheral interrupt lines, such as the System Timer, the Real Time Clock,
the Power Management Controller and the Memory Controller. When a system interrupt
occurs, the service routine must first distinguish the cause of the interrupt. This is performed
by reading successively the status registers of the above mentioned system peripherals.
The interrupt sources 2 to 31 can either be connected to the interrupt outputs of an embedded
user peripheral or to external interrupt lines. The external interrupt lines can be connected
directly, or through the PIO Controller.
The PIO Controllers are considered as user peripherals in the scope of interrupt handling.
Accordingly, the PIO Controller interrupt lines are connected to the Interrupt Sources 2 to 31.
The peripheral identification defined at the product level corresponds to the interrupt source
number (as well as the bit number controlling the clock of the peripheral). Consequently, to
simplify the description of the functional operations and the user interface, the interrupt
sources are named FIQ, SYS, and PID2 to PID31.
233
1768BATARM08/03
Functional Description
Interrupt Source
Control
Interrupt Source Mode The Advanced Interrupt Controller independently programs each interrupt source. The SRC-
TYPE field of the corresponding AIC_SMR (Source Mode Register) selects the interrupt
condition of each source.
The internal interrupt sources wired on the interrupt outputs of the embedded peripherals can
be programmed either in level-sensitive mode or in edge-triggered mode. The active level of
the internal interrupts is not important for the user.
The external interrupt sources can be programmed either in high level-sensitive or low level-
sensitive modes, or in positive edge-triggered or negative edge-triggered modes.
Interrupt Source Each interrupt source, including the FIQ in source 0, can be enabled or disabled by using the
Enabling command registers; AIC_IECR (Interrupt Enable Command Register) and AIC_IDCR (Inter-
rupt Disable Command Register). This set of registers conducts enabling or disabling in one
instruction. The interrupt mask can be read in the AIC_IMR register. A disabled interrupt does
not affect servicing of other interrupts.
Interrupt Clearing and All interrupt sources programmed to be edge-triggered (including the FIQ in source 0) can be
Setting individually set or cleared by writing respectively the AIC_ISCR and AIC_ICCR registers.
Clearing or setting interrupt sources programmed in level-sensitive mode has no effect.
The clear operation is perfunctory, as the software must perform an action to reinitialize the
memorization circuitry activated when the source is programmed in edge-triggered mode.
However, the set operation is available for auto-test or software debug purposes. It can also
be used to execute an AIC-implementation of a software interrupt.
The AIC features an automatic clear of the current interrupt when the AIC_IVR (Interrupt Vec-
tor Register) is read. Only the interrupt source being detected by the AIC as the current
interrupt is affected by this operation. (See Priority Controller on page 237.) The automatic
clear reduces the operations required by the interrupt service routine entry code to reading the
AIC_IVR. Note that the automatic interrupt clear is disabled if the interrupt source has the Fast
Forcing feature enabled as it is considered uniquely as a FIQ source. (For further details, See
Fast Forcing on page 241.)
The automatic clear of the interrupt source 0 is performed when AIC_FVR is read.
Interrupt Status For each interrupt, the AIC operation originates in AIC_IPR (Interrupt Pending Register) and
its mask in AIC_IMR (Interrupt Mask Register). AIC_IPR enables the actual activity of the
sources, whether masked or not.
The AIC_ISR register reads the number of the current interrupt (see Priority Controller on
page 237) and the register AIC_CISR gives an image of the signals nIRQ and nFIQ driven on
the processor.
Each status referred to above can be used to optimize the interrupt handling of the systems.
234 AT91RM9200
1768BATARM08/03
AT91RM9200
nIRQ
Peripheral Interrupt
Becomes Active
Level/ AIC_IPR
Edge
AIC_IMR
Source i
Fast Interrupt Controller
or
Priority Controller
Pos./Neg. AIC_IECR
Edge
Detector FF
Set Clear
AIC_ISCR AIC_IDCR
AIC_ICCR
235
1768BATARM08/03
Interrupt Latencies Global interrupt latencies depend on several parameters, including:
The time the software masks the interrupts.
Occurrence, either at the processor level or at the AIC level.
The execution time of the instruction in progress when the interrupt occurs.
The treatment of higher priority interrupts and the resynchronization of the hardware
signals.
This section addresses only the hardware resynchronizations. It gives details of the latency
times between the event on an external interrupt leading in a valid interrupt (edge or level) or
the assertion of an internal interrupt source and the assertion of the nIRQ or nFIQ line on the
processor. The resynchronization time depends on the programming of the interrupt source
and on its type (internal or external). For the standard interrupt, resynchronization times are
given assuming there is no higher priority in progress.
The PIO Controller multiplexing has no effect on the interrupt latencies of the external interrupt
sources.
IRQ or FIQ
(Negative Edge)
nIRQ
Maximum IRQ Latency = 4 Cycles
nFIQ
Maximum FIQ Latency = 4 Cycles
IRQ or FIQ
(High Level)
IRQ or FIQ
(Low Level)
nIRQ
Maximum IRQ
Latency = 3 Cycles
nFIQ
Maximum FIQ
Latency = 3 cycles
236 AT91RM9200
1768BATARM08/03
AT91RM9200
Internal Interrupt Edge Figure 114. Internal Interrupt Edge Triggered Source
Triggered Source
MCK
nIRQ
Peripheral Interrupt
Becomes Active
Internal Interrupt Level Figure 115. Internal Interrupt Level Sensitive Source
Sensitive Source
MCK
nIRQ
Peripheral Interrupt
Becomes Active
Normal Interrupt
Priority Controller An 8-level priority controller drives the nIRQ line of the processor, depending on the interrupt
conditions occurring on the interrupt sources 1 to 31 (except for those programmed in Fast
Forcing).
Each interrupt source has a programmable priority level of 7 to 0, which is user-definable by
writing the PRIOR field of the corresponding AIC_SMR (Source Mode Register). Level 7 is the
highest priority and level 0 the lowest.
As soon as an interrupt condition occurs, as defined by the SRCTYPE field of the AIC_SVR
(Source Vector Register), the nIRQ line is asserted. As a new interrupt condition might have
happened on other interrupt sources since the nIRQ has been asserted, the priority controller
determines the current interrupt at the time the AIC_IVR (Interrupt Vector Register) is read.
The read of AIC_IVR is the entry point of the interrupt handling which allows the AIC to
consider that the interrupt has been taken into account by the software.
The current priority level is defined as the priority level of the current interrupt.
If several interrupt sources of equal priority are pending and enabled when the AIC_IVR is
read, the interrupt with the lowest interrupt source number is serviced first.
The nIRQ line can be asserted only if an interrupt condition occurs on an interrupt source with
a higher priority. If an interrupt condition happens (or is pending) during the interrupt treatment
in progress, it is delayed until the software indicates to the AIC the end of the current service
by writing the AIC_EOICR (End of Interrupt Command Register). The write of AIC_EOICR is
the exit point of the interrupt handling.
237
1768BATARM08/03
Interrupt Nesting The priority controller utilizes interrupt nesting in order for the highest priority interrupt to be
handled during the service of lower priority interrupts. This requires the interrupt service rou-
tines of the lower interrupts to re-enable the interrupt at the processor level.
When an interrupt of a higher priority happens during an already occurring interrupt service
routine, the nIRQ line is re-asserted. If the interrupt is enabled at the core level, the current
execution is interrupted and the new interrupt service routine should read the AIC_IVR. At this
time, the current interrupt number and its priority level are pushed into an embedded hardware
stack, so that they are saved and restored when the higher priority interrupt servicing is fin-
ished and the AIC_EOICR is written.
The AIC is equipped with an 8-level wide hardware stack in order to support up to eight inter-
rupt nestings pursuant to having eight priority levels.
Interrupt Vectoring The interrupt handler addresses corresponding to each interrupt source can be stored in the
registers AIC_SVR1 to AIC_SVR31 (Source Vector Register 1 to 31). When the processor
reads AIC_IVR (Interrupt Vector Register), the value written into AIC_SVR corresponding to
the current interrupt is returned.
This feature offers a way to branch in one single instruction to the handler corresponding to
the current interrupt, as AIC_IVR is mapped at the absolute address 0xFFFF F100 and thus
accessible from the ARM interrupt vector at address 0x0000 0018 through the following
instruction:
LDR PC,[PC,# -&F20]
When the processor executes this instruction, it loads the read value in AIC_IVR in its program
counter, thus branching the execution on the correct interrupt handler.
This feature is often not used when the application is based on an operating system (either
real time or not). Operating systems often have a single entry point for all the interrupts and
the first task performed is to discern the source of the interrupt.
However, it is strongly recommended to port the operating system on AT91 products by sup-
porting the interrupt vectoring. This can be performed by defining all the AIC_SVR of the
interrupt source to be handled by the operating system at the address of its interrupt handler.
When doing so, the interrupt vectoring permits a critical interrupt to transfer the execution on a
specific very fast handler and not onto the operating systems general interrupt handler. This
facilitates the support of hard real-time tasks (input/outputs of voice/audio buffers and software
peripheral handling) to be handled efficiently and independently of the application running
under an operating system.
Interrupt Handlers This section gives an overview of the fast interrupt handling sequence when using the AIC. It
is assumed that the programmer understands the architecture of the ARM processor, and
especially the processor interrupt modes and the associated status bits.
It is assumed that:
1. The Advanced Interrupt Controller has been programmed, AIC_SVR registers are
loaded with corresponding interrupt service routine addresses and interrupts are
enabled.
2. The instruction at the ARM interrupt exception vector address is required to work with
the vectoring
LDR PC, [PC, # -&F20]
238 AT91RM9200
1768BATARM08/03
AT91RM9200
0x18. In the following cycle during fetch at address 0x1C, the ARM core adjusts
R14_irq, decrementing it by four.
2. The ARM core enters Interrupt mode, if it has not already done so.
3. When the instruction loaded at address 0x18 is executed, the program counter is
loaded with the value read in AIC_IVR. Reading the AIC_IVR has the following effects:
Sets the current interrupt to be the pending and enabled interrupt with the highest
priority. The current level is the priority level of the current interrupt.
De-asserts the nIRQ line on the processor. Even if vectoring is not used, AIC_IVR
must be read in order to de-assert nIRQ.
Automatically clears the interrupt, if it has been programmed to be edge-triggered.
Pushes the current level and the current interrupt number on to the stack.
Returns the value written in the AIC_SVR corresponding to the current interrupt.
4. The previous step has the effect of branching to the corresponding interrupt service
routine. This should start by saving the link register (R14_irq) and SPSR_IRQ. The link
register must be decremented by four when it is saved if it is to be restored directly into
the program counter at the end of the interrupt. For example, the instruction SUB PC,
LR, #4 may be used.
5. Further interrupts can then be unmasked by clearing the I bit in CPSR, allowing re-
assertion of the nIRQ to be taken into account by the core. This can happen if an inter-
rupt with a higher priority than the current interrupt occurs.
6. The interrupt handler can then proceed as required, saving the registers that will be
used and restoring them at the end. During this phase, an interrupt of higher priority
than the current level will restart the sequence from step 1.
Note: If the interrupt is programmed to be level sensitive, the source of the interrupt must be cleared
during this phase.
7. The I bit in CPSR must be set in order to mask interrupts before exiting to ensure that
the interrupt is completed in an orderly manner.
8. The End of Interrupt Command Register (AIC_EOICR) must be written in order to indi-
cate to the AIC that the current interrupt is finished. This causes the current level to be
popped from the stack, restoring the previous current level if one exists on the stack. If
another interrupt is pending, with lower or equal priority than the old current level but
with higher priority than the new current level, the nIRQ line is re-asserted, but the
interrupt sequence does not immediately start because the I bit is set in the core.
SPSR_irq is restored. Finally, the saved value of the link register is restored directly
into the PC. This has effect of returning from the interrupt to whatever was being exe-
cuted before, and of loading the CPSR with the stored SPSR, masking or unmasking
the interrupts depending on the state saved in SPSR_irq.
Note: The I bit in SPSR is significant. If it is set, it indicates that the ARM core was on the verge of
masking an interrupt when the mask instruction was interrupted. Hence, when SPSR is
restored, the mask instruction is completed (interrupt is masked).
Fast Interrupt
Fast Interrupt Source The interrupt source 0 is the only source which can raise a fast interrupt request to the proces-
sor except if fast forcing is used. The interrupt source 0 is generally connected to a FIQ pin of
the product, either directly or through a PIO Controller.
239
1768BATARM08/03
reads what has been written. The field SRCTYPE of AIC_SMR0 enables programming the
fast interrupt source to be positive-edge triggered or negative-edge triggered or high-level sen-
sitive or low-level sensitive
Writing 0x1 in the AIC_IECR (Interrupt Enable Command Register) and AIC_IDCR (Interrupt
Disable Command Register) respectively enables and disables the fast interrupt. The bit 0 of
AIC_IMR (Interrupt Mask Register) indicates whether the fast interrupt is enabled or disabled.
Fast Interrupt The fast interrupt handler address can be stored in AIC_SVR0 (Source Vector Register 0).
Vectoring The value written into this register is returned when the processor reads AIC_FVR (Fast Vec-
tor Register). This offers a way to branch in one single instruction to the interrupt handler, as
AIC_FVR is mapped at the absolute address 0xFFFF F104 and thus accessible from the ARM
fast interrupt vector at address 0x0000 001C through the following instruction:
LDR PC,[PC,# -&F20]
When the processor executes this instruction it loads the value read in AIC_FVR in its pro-
gram counter, thus branching the execution on the fast interrupt handler. It also automatically
performs the clear of the fast interrupt source if it is programmed in edge-triggered mode.
Fast Interrupt This section gives an overview of the fast interrupt handling sequence when using the AIC. It
Handlers is assumed that the programmer understands the architecture of the ARM processor, and
especially the processor interrupt modes and associated status bits.
Assuming that:
1. The Advanced Interrupt Controller has been programmed, AIC_SVR0 is loaded with
the fast interrupt service routine address, and the interrupt source 0 is enabled.
2. The Instruction at address 0x1C (FIQ exception vector address) is required to vector
the fast interrupt:
LDR PC, [PC, # -&F20]
3. The user does not need nested fast interrupts.
When nFIQ is asserted if the bit "F" of CPSR is 0, the sequence is:
1. The CPSR is stored in SPSR_fiq, the current value of the program counter is loaded in
the FIQ link register (R14_FIQ) and the program counter (R15) is loaded with 0x1C. In
the following cycle, during fetch at address 0x20, the ARM core adjusts R14_fiq, decre-
menting it by four.
2. The ARM core enters FIQ mode.
3. When the instruction loaded at address 0x1C is executed, the program counter is
loaded with the value read in AIC_FVR. Reading the AIC_FVR has effect of automati-
cally clearing the fast interrupt, if it has been programmed to be edge triggered. In this
case only, it de-asserts the nFIQ line on the processor.
4. The previous step enables branching to the corresponding interrupt service routine. It
is not necessary to save the link register R14_fiq and SPSR_fiq if nested fast interrupts
are not needed.
5. The Interrupt Handler can then proceed as required. It is not necessary to save regis-
ters R8 to R13 because FIQ mode has its own dedicated registers and the user R8 to
R13 are banked. The other registers, R0 to R7, must be saved before being used, and
restored at the end (before the next step). Note that if the fast interrupt is programmed
to be level sensitive, the source of the interrupt must be cleared during this phase in
order to de-assert the interrupt source 0.
6. Finally, the Link Register R14_fiq is restored into the PC after decrementing it by four
(with instruction SUB PC, LR, #4 for example). This has the effect of returning from
the interrupt to whatever was being executed before, loading the CPSR with the SPSR
240 AT91RM9200
1768BATARM08/03
AT91RM9200
and masking or unmasking the fast interrupt depending on the state saved in the
SPSR.
Note: The "F" bit in SPSR is significant. If it is set, it indicates that the ARM core was just about to
mask FIQ interrupts when the mask instruction was interrupted. Hence when the SPSR is
restored, the interrupted instruction is completed (FIQ is masked).
Another way to handle the fast interrupt is to map the interrupt service routine at the address
of the ARM vector 0x1C. This method does not use the vectoring, so that reading AIC_FVR
must be performed at the very beginning of the handler operation. However, this method
saves the execution of a branch instruction.
Fast Forcing The Fast Forcing feature of the advanced interrupt controller provides redirection of any nor-
mal Interrupt source on the fast interrupt controller.
Fast Forcing is enabled or disabled by writing to the Fast Forcing Enable Register
(AIC_FFER) and the Fast Forcing Disable Register (AIC_FFDR). Writing to these registers
results in an update of the Fast Forcing Status Register (AIC_FFSR) that controls the feature
for each internal or external interrupt source.
When Fast Forcing is disabled, the interrupt sources are handled as described in the previous
pages.
When Fast Forcing is enabled, the edge/level programming and, in certain cases, edge detec-
tion of the interrupt source is still active but the source cannot trigger a normal interrupt to the
processor and is not seen by the priority handler.
If the interrupt source is programmed in level-sensitive mode and an active level is sampled,
Fast Forcing results in the assertion of the nFIQ line to the core.
If the interrupt source is programmed in edge-triggered mode and an active edge is detected,
Fast Forcing results in the assertion of the nFIQ line to the core.
The Fast Forcing feature does not affect the Source 0 pending bit in the Interrupt Pending
Register (AIC_IPR).
The Fast Interrupt Vector Register (AIC_FVR) reads the contents of the Source Vector Regis-
ter 0 (AIC_SVR0), whatever the source of the fast interrupt may be. The read of the FVR does
not clear the Source 0 when the fast forcing feature is used and the interrupt source should be
cleared by writing to the Interrupt Clear Command Register (AIC_ICCR).
All enabled and pending interrupt sources that have the fast forcing feature enabled and that
are programmed in edge-triggered mode must be cleared by writing to the Interrupt Clear
Command Register. In doing so, they are cleared independently and thus lost interrupts are
prevented.
The read of AIC_IVR does not clear the source that has the fast forcing feature enabled.
The source 0, reserved to the fast interrupt, continues operating normally and becomes one of
the Fast Interrupt sources.
241
1768BATARM08/03
Figure 116. Fast Forcing
Source 0 _ FIQ AIC_IPR
Input Stage
AIC_FFSR
Source n AIC_IPR
Input Stage
Priority
Manager
Automatic Clear AIC_IMR nIRQ
Protect Mode The Protect Mode permits reading the Interrupt Vector Register without performing the associ-
ated automatic operations. This is necessary when working with a debug system. When a
debugger, working either with a Debug Monitor or the ARM processor's ICE, stops the applica-
tions and updates the opened windows, it might read the AIC User Interface and thus the IVR.
This has undesirable consequences:
If an enabled interrupt with a higher priority than the current one is pending, it is stacked.
If there is no enabled pending interrupt, the spurious vector is returned.
In either case, an End of Interrupt command is necessary to acknowledge and to restore the
context of the AIC. This operation is generally not performed by the debug system as the
debug system would become strongly intrusive and cause the application to enter an undes-
ired state.
This is avoided by using the Protect Mode. Writing DBGM in AIC_DCR (Debug Control Regis-
ter) at 0x1 enables the Protect Mode.
When the Protect Mode is enabled, the AIC performs interrupt stacking only when a write
access is performed on the AIC_IVR. Therefore, the Interrupt Service Routines must write
(arbitrary data) to the AIC_IVR just after reading it. The new context of the AIC, including the
value of the Interrupt Status Register (AIC_ISR), is updated with the current interrupt only
when AIC_IVR is written.
An AIC_IVR read on its own (e.g., by a debugger), modifies neither the AIC context nor the
AIC_ISR. Extra AIC_IVR reads perform the same operations. However, it is recommended to
not stop the processor between the read and the write of AIC_IVR of the interrupt service rou-
tine to make sure the debugger does not modify the AIC context.
To summarize, in normal operating mode, the read of AIC_IVR performs the following opera-
tions within the AIC:
1. Calculates active interrupt (higher than current or spurious).
2. Determines and returns the vector of the active interrupt.
3. Memorizes the interrupt.
4. Pushes the current priority level onto the internal stack.
5. Acknowledges the interrupt.
242 AT91RM9200
1768BATARM08/03
AT91RM9200
However, while the Protect Mode is activated, only operations 1 to 3 are performed when
AIC_IVR is read. Operations 4 and 5 are only performed by the AIC when AIC_IVR is written.
Software that has been written and debugged using the Protect Mode runs correctly in Normal
Mode without modification. However, in Normal Mode the AIC_IVR write has no effect and can
be removed to optimize the code.
Spurious Interrupt The Advanced Interrupt Controller features protection against spurious interrupts. A spurious
interrupt is defined as being the assertion of an interrupt source long enough for the AIC to
assert the nIRQ, but no longer present when AIC_IVR is read. This is most prone to occur
when:
An external interrupt source is programmed in level-sensitive mode and an active level
occurs for only a short time.
An internal interrupt source is programmed in level sensitive and the output signal of the
corresponding embedded peripheral is activated for a short time. (As in the case for the
Watchdog.)
An interrupt occurs just a few cycles before the software begins to mask it, thus resulting in
a pulse on the interrupt source.
The AIC detects a spurious interrupt at the time the AIC_IVR is read while no enabled interrupt
source is pending. When this happens, the AIC returns the value stored by the programmer in
AIC_SPU (Spurious Vector Register). The programmer must store the address of a spurious
interrupt handler in AIC_SPU as part of the application, to enable an as fast as possible return
to the normal execution flow. This handler writes in AIC_EOICR and performs a return from
interrupt.
General Interrupt The AIC features a General Interrupt Mask bit to prevent interrupts from reaching the proces-
Mask sor. Both the nIRQ and the nFIQ lines are driven to their inactive state if the bit GMSK in
AIC_DCR (Debug Control Register) is set. However, this mask does not prevent waking up
the processor if it has entered Idle Mode. This function facilitates synchronizing the processor
on a next event and, as soon as the event occurs, performs subsequent operations without
having to handle an interrupt. It is strongly recommended to use this mask with caution.
243
1768BATARM08/03
Advanced Interrupt Controller (AIC) User Interface
Base Address The AIC is mapped at the address 0xFFFF F000. It has a total 4-Kbyte addressing space. This
permits the vectoring feature, as the PC-relative load/store instructions of the ARM processor
supports only an 4-Kbyte offset.
244 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
SRCTYPE PRIOR
23 22 21 20 19 18 17 16
VECTOR
15 14 13 12 11 10 9 8
VECTOR
7 6 5 4 3 2 1 0
VECTOR
245
1768BATARM08/03
Access Type: Read-only
Reset Value: 0
31 30 29 28 27 26 25 24
IRQV
23 22 21 20 19 18 17 16
IRQV
15 14 13 12 11 10 9 8
IRQV
7 6 5 4 3 2 1 0
IRQV
23 22 21 20 19 18 17 16
FIQV
15 14 13 12 11 10 9 8
FIQV
7 6 5 4 3 2 1 0
FIQV
246 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
IRQID
23 22 21 20 19 18 17 16
PID23 PID22 PID21 PID20 PID19 PID18 PID17 PID16
15 14 13 12 11 10 9 8
PID15 PID14 PID13 PID12 PID11 PID10 PID9 PID8
7 6 5 4 3 2 1 0
PID7 PID6 PID5 PID4 PID3 PID2 SYS FIQ
247
1768BATARM08/03
AIC Interrupt Mask Register
Register Name: AIC_IMR
Access Type: Read-only
Reset Value: 0
31 30 29 28 27 26 25 24
PID31 PID30 PID29 PID28 PID27 PID26 PID25 PID24
23 22 21 20 19 18 17 16
PID23 PID22 PID21 PID20 PID19 PID18 PID17 PID16
15 14 13 12 11 10 9 8
PID15 PID14 PID13 PID12 PID11 PID10 PID9 PID8
7 6 5 4 3 2 1 0
PID7 PID6 PID5 PID4 PID3 PID2 SYS FIQ
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
NIRQ NIFQ
248 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
PID23 PID22 PID21 PID20 PID19 PID18 PID17 PID16
15 14 13 12 11 10 9 8
PID15 PID14 PID13 PID12 PID11 PID10 PID9 PID8
7 6 5 4 3 2 1 0
PID7 PID6 PID5 PID4 PID3 PID2 SYS FIQ
23 22 21 20 19 18 17 16
PID23 PID22 PID21 PID20 PID19 PID18 PID17 PID16
15 14 13 12 11 10 9 8
PID15 PID14 PID13 PID12 PID11 PID10 PID9 PID8
7 6 5 4 3 2 1 0
PID7 PID6 PID5 PID4 PID3 PID2 SYS FIQ
249
1768BATARM08/03
AIC Interrupt Clear Command Register
Register Name: AIC_ICCR
Access Type: Write-only
31 30 29 28 27 26 25 24
PID31 PID30 PID29 PID28 PID27 PID26 PID25 PID24
23 22 21 20 19 18 17 16
PID23 PID22 PID21 PID20 PID19 PID18 PID17 PID16
15 14 13 12 11 10 9 8
PID15 PID14 PID13 PID12 PID11 PID10 PID9 PID8
7 6 5 4 3 2 1 0
PID7 PID6 PID5 PID4 PID3 PID2 SYS FIQ
23 22 21 20 19 18 17 16
PID23 PID22 PID21 PID20 PID19 PID18 PID17 PID16
15 14 13 12 11 10 9 8
PID15 PID14 PID13 PID12 PID11 PID10 PID9 PID8
7 6 5 4 3 2 1 0
PID7 PID6 PID5 PID4 PID3 PID2 SYS FIQ
250 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
The End of Interrupt Command Register is used by the interrupt routine to indicate that the interrupt treatment is complete.
Any value can be written because it is only necessary to make a write to this register location to signal the end of interrupt
treatment.
23 22 21 20 19 18 17 16
SIQV
15 14 13 12 11 10 9 8
SIQV
7 6 5 4 3 2 1 0
SIQV
251
1768BATARM08/03
AIC Debug Control Register
Register Name: AIC_DEBUG
Access Type: Read/write
Reset Value: 0
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
GMSK PROT
252 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
PID23 PID22 PID21 PID20 PID19 PID18 PID17 PID16
15 14 13 12 11 10 9 8
PID15 PID14 PID13 PID12 PID11 PID10 PID9 PID8
7 6 5 4 3 2 1 0
PID7 PID6 PID5 PID4 PID3 PID2 SYS
23 22 21 20 19 18 17 16
PID23 PID22 PID21 PID20 PID19 PID18 PID17 PID16
15 14 13 12 11 10 9 8
PID15 PID14 PID13 PID12 PID11 PID10 PID9 PID8
7 6 5 4 3 2 1 0
PID7 PID6 PID5 PID4 PID3 PID2 SYS
253
1768BATARM08/03
AIC Fast Forcing Status Register
Register Name: AIC_FFSR
Access Type: Read-only
31 30 29 28 27 26 25 24
PID31 PID30 PID29 PID28 PID27 PID26 PID25 PID24
23 22 21 20 19 18 17 16
PID23 PID22 PID21 PID20 PID19 PID18 PID17 PID16
15 14 13 12 11 10 9 8
PID15 PID14 PID13 PID12 PID11 PID10 PID9 PID8
7 6 5 4 3 2 1 0
PID7 PID6 PID5 PID4 PID3 PID2 SYS
254 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Power Management Controller (PMC) generates all the clocks of a system thanks
to the integration of two oscillators and two PLLs.
The PMC provides clocks to the embedded processor and enables the idle mode by
stopping the processor clock until the next interrupt.
The PMC independently provides and controls up to thirty peripheral clocks and four
programmable clocks that can be used as outputs on pins to feed external devices. The
integration of the PLLs supplies the USB devices and host ports with a 48 MHz clock, as
required by the bus speed, and the rest of the system with a clock at another frequency.
Thus, the fully-featured Power Management Controller optimizes power consumption of
the whole system and supports the Normal, Idle, Slow Clock and Standby operating
modes.
The main features of the PMC are:
Optimize the Power Consumption of the Whole System
Embeds and Controls:
One Main Oscillator and One Slow Clock Oscillator (32.768 kHz)
Two Phase Locked Loops (PLLs) and Dividers
Clock Prescalers
Provides:
the Processor Clock PCK
the Master Clock MCK
the USB Clocks, UHPCK and UDPCK, Respectively for the USB Host Port
and the USB Device Port
Programmable Automatic PLL Switch-off in USB Device Suspend Conditions
up to Thirty Peripheral Clocks
up to Four Programmable Clock Outputs
Four Operating Modes:
Normal Mode, Idle Mode, Slow Clock Mode, Standby Mode
255
1768BATARM08/03
Product Dependencies
I/O Lines The Power Management Controller is capable of handling up to four Programmable
Clocks, PCK0 to PCK3.
A Programmable Clock is generally multiplexed on a PIO Controller. The user must first
program the PIO controllers to assign the pins of the Programmable Clock to its periph-
eral function.
Interrupt The Power Management Controller has an interrupt line connected to the Advanced
Interrupt Controller (AIC). Handling the PMC interrupt requires programming the AIC
before configuring the PMC.
Oscillator and PLL The electrical characteristics of the embedded oscillators and PLLs are product-depen-
Characteristics dent, even if the way to control them is similar.
All of the parameters for both oscillators and the PLLs are given in the DC Characteris-
tics section of the product datasheet. These figures are used not only for the hardware
design, as they affect the external components to be connected to the pins, but also the
software configuration, as they determine the waiting time for the startup and lock times
to be programmed.
Peripheral Clocks The Power Management Controller provides and controls up to thirty peripheral clocks.
The bit number permitting the control of a peripheral clock is the Peripheral ID of the
embedded peripheral.
When the Peripheral ID does not correspond to a peripheral, either because this is an
external interrupt or because there are less than thirty peripherals, the control bits of the
Peripheral ID are not implemented in the PMC and programming them has no effect on
the behavior of the PMC.
USB Clocks The Power Management Controller provides and controls two USB Clocks, the UHPCK
for the USB Host Port, and the UDPCK for the USB Device.
If the product does not embed either the USB Host Port or the USB Device Port, the
associated control bits and registers are not implemented in the PMC and programming
them has no effect on the behavior of the PMC.
256 AT91RM9200
1768BATARM08/03
AT91RM9200
Block Diagram
Figure 117. Power Management Controller Block Diagram
Processor
Power Management Controller Clock ARM7
Processor Processor
Clock
Controller
Processor
Clock Generator Idle Mode Clock ARM920T
Processor
XIN32 Slow
Slow Clock Clock IRQ or FIQ
Oscillator SLCK
XOUT32 Master Clock Controller
PMCIRQ AIC
SLCK Divider
Main Clock Prescaler /1,/2,/3,/4
XIN
Main Main PLLA Clock /2,/4,...,/64 ARM9-systems MCK
Oscillator Clock PLLB Clock only (Continuous)
XOUT
Memory Controller
UDPCK
UDP
USB Clock
PLLB Controller Suspend
Clock
ON/OFF UHPCK
UHP
ST
Slow
Clock
User Interface SLCK
SLCK
RTC
APB
257
1768BATARM08/03
Functional Description
Operating Modes The following operating modes are supported by the PMC and offer different power con-
Definition sumption levels and event response latency times:
Normal Mode: The ARM processor clock is enabled and peripheral clocks are
enabled depending on application requirements.
Idle Mode: The ARM processor clock is disabled and waiting for the next interrupt
(or a main reset). The peripheral clocks are enabled depending on application
requirements. PDC transfers are still possible.
Slow Clock Mode: Slow clock mode is similar to normal mode, but the main
oscillator and the PLL are switched off to save power and the processor and the
peripherals run in Slow Clock mode. Note that slow clock mode is the mode
selected after the reset.
Standby Mode: Standby mode is a combination of Slow Clock mode and Idle Mode.
It enables the processor to respond quickly to a wake-up event by keeping power
consumption very low.
Clock Definitions The Power Management Controller provides the following clocks:
Slow Clock (SLCK), typically at 32.768 kHz, is the only permanent clock within the
system.
Master Clock (MCK), programmable from a few hundred Hz to the maximum
operating frequency of the device. It is available to the modules running
permanently, such as the AIC and the Memory Controller.
Processor Clock (PCK), typically the Master Clock for ARM7-based systems and a
faster clock on ARM9-based systems, switched off when entering idle mode.
Peripheral Clocks, typically MCK, provided to the embedded peripherals (USART,
SSC, SPI, TWI, TC, MCI, etc.) and independently controllable. In order to reduce the
number of clock names in a product, the Peripheral Clocks are named MCK in the
product datasheet.
UDP Clock (UDPCK), typically at 48 MHz, required by the USB Device Port
operations.
UHP Clock (UHPCK), typically at 48 MHz, required by the USB Host Port
operations.
Programmable Clock Outputs (PCK0 to PCK3) can be selected from the clocks
provided by the clock generator and driven on the PCK0 to PCK3 pins.
258 AT91RM9200
1768BATARM08/03
AT91RM9200
Clock Generator
XIN32 Slow
Slow Clock
Oscillator Clock
XOUT32 SLCK
XIN
Main
Main
Oscillator
Clock
XOUT
Main Clock
Frequency PLLA
Divider A PLL A
Counter Clock
PLLRCA
/2 PLLB
Divider B PLL B (optional)
Clock
PLLRCB
Slow Clock Oscillator The Clock Generator integrates a low-power 32.768 kHz oscillator. The XIN32 and
Connection XOUT32 pins must be connected to a 32.768 kHz crystal. Two external capacitors must
be wired as shown in Figure 119.
CL1 CL2
Slow Clock Oscillator Startup The startup time of the Slow Clock Oscillator is given in the DC Characteristics section
Time of the product datasheet. As it is often higher than 500 ms and the processor requires
an assertion of the reset until it has stabilized, the user must implement an external
reset supervisor covering this startup time. However, this startup is only required in case
of cold reset, i.e., in case of system power-up. When a warm reset occurs, the length of
the reset pulse may be much lower. For further details, see AT91RM9200 Reset Con-
troller on page 119.
259
1768BATARM08/03
Main Oscillator Figure 120 shows the Main Oscillator block diagram.
MOSCEN
XIN
Main Main
Oscillator Clock
XOUT
OSCOUNT
Main
Slow
Oscillator MOSCS
Clock
Counter
Main Oscillator Connections The Clock Generator integrates a Main Oscillator that is designed for a 3 to 20 MHz fun-
damental crystal. The typical crystal connection is illustrated in Figure 121. The 1 k
resistor is only required for crystals with frequencies lower than 8 MHz. The oscillator
contains twenty-five pF capacitors on each XIN and XOUT pin. Consequently, CL1 and
CL2 can be removed when a crystal with a load capacitance of 12.5 pF is used. For fur-
ther details on the electrical characteristics of the Main Oscillator, see the DC
Characteristics section of the product datasheet.
1K
CL1 CL2
Main Oscillator Startup Time The startup time of the Main Oscillator is given in the DC Characteristics section of the
product datasheet. The startup time depends on the crystal frequency and increases
when the frequency rises.
Main Oscillator Control To minimize the power required to start up the system, the Main Oscillator is disabled
after reset and the Slow Clock mode is selected.
The software enables or disables the Main Oscillator so as to reduce power consump-
tion by clearing the MOSCEN bit in the Main Oscillator Register (CKGR_MOR). When
disabling the Main Oscillator by clearing the MOSCEN bit in CKGR_MOR, the MOSCS
bit in PMC_SR is automatically cleared indicating the Main Clock is off.
260 AT91RM9200
1768BATARM08/03
AT91RM9200
When enabling the Main Oscillator, the user must initiate the Main Oscillator counter
with a value corresponding to the startup time of the oscillator. This startup time
depends on the crystal frequency connected to the main oscillator. When the MOSCEN
bit and the OSCOUNT are written in CKGR_MOR to enable the Main Oscillator, the
MOSCS bit is cleared and the counter starts counting down on the Slow Clock divided
by 8 from the OSCOUNT value. Since the OSCOUNT value is coded with 8 bits, the
maximum startup time is about 62 ms.
When the counter reaches 0, the MOSCS bit is set, indicating that the Main Clock is
valid. Setting the MOSCS bit in PMC_IMR can trigger an interrupt to the processor on
this event.
Main Clock Frequency The Main Oscillator features a Main Clock frequency counter that provides the quartz
Counter frequency connected to the Main Oscillator. Generally, this value is known by the sys-
tem designer; however, it is useful for the boot program to configure the device with the
correct clock speed, independently of the application.
The Main Clock frequency counter starts incrementing at the Main Clock speed after the
next rising edge of the Slow Clock as soon as the Main Oscillator is stable, i.e., as soon
as the MOSCS bit is set. Then, at the 16th falling edge of Slow Clock, the bit MAINRDY
in CKGR_MCFR (Main Clock Frequency Register) is set and the counter stops count-
ing. Its value can be read in the MAINF field of CKGR_MCFR and gives the number of
Main Clock cycles during 16 periods of Slow Clock, so that the frequency of the crystal
connected on the Main Oscillator can be determined.
Main Oscillator Bypass The user can input a clock on the device instead of connecting a crystal. In this case, the
user has to provide the external clock signal on the pin XIN. The input characteristics of
the XIN pin under these conditions are given in the product Electrical Characteristics
section. The programmer has to be sure not to modify the MOSCEN bit in the Main
Oscillator Register (CKGR_MOR). This bit must remain at 0, its reset value, for the
external clock to operate properly. While this bit is at 0, the pin XIN is tied low to prevent
any internal oscillation regardless of pin connected.
The external clock signal must meet the requirements relating to the power supply
VDDPLL (i.e., between 1.65V and 1.95V) and cannot exceed 50 MHz.
261
1768BATARM08/03
Divider and PLL Blocks The Clock Generator features two Divider/PLL Blocks that generates a wide range of
frequencies. Additionally, they provide a 48 MHz signal to the embedded USB device
and/or host ports, regardless of the frequency of the Main Clock.
Figure 122 shows the block diagram of the divider and PLL blocks.
PLL B
Main
Divider B PLL B Output
Clock
PLLRCB
PLLRCA
PLLBCOUNT
PLL B
LOCKB
Counter
PLLACOUNT
Slow
PLL A
Clock LOCKA
Counter
PLL Filters The two PLLs require connection to an external second-order filter through the pins
PLLRC. Figure 123 shows a schematic of these filters.
C2
C1
GND
262 AT91RM9200
1768BATARM08/03
AT91RM9200
PLL Source Clock The source of PLLs A and B is respectively the output of Divider A, i.e. the Main Clock
divided by DIVA, and the output of Divider B, i.e. the Main Clock divided by DIVB.
As the input frequency of the PLLs is limited, the user has to make sure that the pro-
gramming of DIVA and DIVB are compliant with the input frequency range of the PLLs,
which is given in the DC Characteristics section of the product datasheet.
Divider and Phase Lock Loop The two dividers increase the accuracy of the PLLA and the PLLB clocks independently
Programming of the input frequency.
The Main Clock can be divided by programming the DIVB field in CKGR_PLLBR and
the DIVA field in CKGR_PLLAR. Each divider can be set between 1 and 255 in steps of
1. When the DIVA and DIVB fields are set to 0, the output of the divider and the PLL out-
puts A and B are a continuous signal at level 0. On reset, the DIVA and DIVB fields are
set to 0, thus both PLL input clocks are set to 0.
The two PLLs of the clock generator allow multiplication of the dividers outputs. The
PLLA and the PLLB clock signals have a frequency that depends on the respective
source signal frequency and on the parameters DIV (DIVA, DIVB) and MUL (MULA,
MULB). The factor applied to the source signal frequency is (MUL + 1)/DIV. When
MULA or MULB is written to 0, the corresponding PLL is disabled and its power con-
sumption is saved. Re-enabling the PLLA or the PLLB can be performed by writing a
value higher than 0 in the MULA or MULB field, respectively.
Whenever a PLL is re-enabled or one of its parameters is changed, the LOCKA or
LOCKB bit in PMC_SR is automatically cleared. The values written in the PLLACOUNT
or PLLBCOUNT fields in CKGR_PPLAR and CKGR_PLLBR, respectively, are loaded in
the corresponding PLL counter. The PLL counter then decrements at the speed of the
Slow Clock until it reaches 0. At this time, the corresponding LOCK bit is set in PMC_SR
and can trigger an interrupt to the processor. The user has to load the number of Slow
Clock cycles required to cover the PLL transient time into the PLLACOUNT and PLLB-
COUNT field. The transient time depends on the PLL filters. The initial state of the PLL
and its target frequency can be calculated using a specific tool provided by Atmel.
PLLB Divider by 2 In ARM9-based systems, the PLLB clock may be divided by two. This divider can be
enabled by setting the bit USB_96M of CKGR_PLLBR. In this case, the divider by 2 is
enabled and the PLLB must be programmed to output 96 MHz and not 48 MHz, thus
ensuring correct operation of the USB bus.
Clock Controllers The Power Management Controller provides the clocks to the different peripherals of the
system, either internal or external. It embeds the following elements:
the Master Clock Controller, which selects the Master Clock.
the Processor Clock Controller, which implements the Idle Mode.
the Peripheral Clock Controller, which provides power saving by controlling clocks of
the embedded peripherals.
the USB Clock Controller, which distributes the 48 MHz clock to the USB controllers.
the Programmable Clock Controller, which allows generation of up to four
programmable clock signals on external pins.
Master Clock Controller The Master Clock Controller provides selection and division of the Master Clock (MCK).
MCK is the clock provided to all the peripherals and the memory controller.
The Master Clock is selected from one of the clocks provided by the Clock Generator.
Selecting the Slow Clock enables Slow Clock Mode by providing a 32.768 kHz signal to
the whole device. Selecting the Main Clock saves power consumption of both PLLs, but
263
1768BATARM08/03
prevents using the USB ports. Selecting the PLLB Clock saves the power consumption
of the PLLA by running the processor and the peripheral at 48 MHz required by the USB
ports. Selecting the PLLA Clock runs the processor and the peripherals at their maxi-
mum speed while running the USB ports at 48 MHz.
The Master Clock Controller is made up of a clock selector and a prescaler, as shown in
Figure 124. It also contains an optional Master Clock divider in products integrating an
ARM9 processor. This allows the processor clock to be faster than the Master Clock.
The Master Clock selection is made by writing the CSS field (Clock Source Selection) in
PMC_MCKR (Master Clock Register). The prescaler supports the division by a power of
2 of the selected clock between 1 and 64. The PRES field in PMC_MCKR programs the
prescaler.
When the Master Clock divider is implemented, it can be programmed between 1 and 4
through the MDIV field in PMC_MCKR.
Each time PMC_MCKR is written to define a new Master Clock, the MCKRDY bit is
cleared in PMC_SR. It reads 0 until the Master Clock is established. Then, the MCK-
RDY bit is set and can trigger an interrupt to the processor. This feature is useful when
switching from a high-speed clock to a lower one to inform the software when the
change is actually done.
Note: A new value to be written in PMC_MCKR must not be the same as the current value in
PMC_MCKR.
Master
Clock MCK
CD PRES Divider
Processor Clock Controller The PMC features a Processor Clock Controller that implements the Idle Mode. The
Processor Clock can be enabled and disabled by writing the System Clock Enable
(PMC_SCER) and System Clock Disable Registers (PMC_SCDR). The status of this
clock (at least for debug purpose) can be read in the System Clock Status Register
(PMC_SCSR).
Processor Clock Source The clock provided to the processor is determined by the Master Clock controller. On
ARM7-based systems, the Processor Clock source is directly the Master Clock.
On ARM9-based systems, the Processor Clock source might be 2, 3 or 4 times the Mas-
ter Clock. This ratio value is determined by programming the field MDIV of the Master
Clock Register (PMC_MCKR).
Idle Mode The Processor Clock is enabled after a reset and is automatically re-enabled by any
enabled interrupt. The Idle Mode is achieved by disabling the Processor Clock, which is
automatically re-enabled by any enabled fast or normal interrupt, or by the reset of the
product.
264 AT91RM9200
1768BATARM08/03
AT91RM9200
When the Processor Clock is disabled, the current instruction is finished before the clock
is stopped, but this does not prevent data transfers from other masters of the system
bus.
Peripheral Clock Controller The PMC controls the clocks of each embedded peripheral. The user can individually
enable and disable the Master Clock on the peripherals by writing into the Peripheral
Clock Enable (PMC_PCER) and Peripheral Clock Disable (PMC_PCDR) registers. The
status of the peripheral clock activity can be read in the Peripheral Clock Status Register
(PMC_PCSR).
When a peripheral clock is disabled, the clock is immediately stopped. When the clock is
re-enabled, the peripheral resumes action where it left off. The peripheral clocks are
automatically disabled after a reset.
In order to stop a peripheral, it is recommended that the system software wait until the
peripheral has executed its last programmed operation before disabling the clock. This
is to avoid data corruption or erroneous behavior of the system.
The bit number within the Peripheral Clock Control registers (PMC_PCER,
PMC_PCDR, and PMC_PCSR) is the Peripheral Identifier defined at the product level.
Generally, the bit number corresponds to the interrupt source number assigned to the
peripheral.
USB Clock Controller If using one of the USB ports, the user has to program the Divider and PLL B block to
output a 48 MHz signal with an accuracy of 0.25%.
When the clock for the USB is stable, the USB device and host clocks, UDPCK and
UHPCK, can be enabled. They can be disabled when the USB transactions are finished,
so that the power consumption generated by the 48 MHz signal on these peripherals is
saved.
The USB ports require both the 48 MHz signal and the Master Clock. The Master Clock
may be controlled via the Peripheral Clock Controller.
USB Device Clock Control The USB Device Port clock UDPCK can be enabled by writing 1 at the UDP bit in
PMC_SCER (System Clock Enable Register) and disabled by writing 1 at the bit UDP in
PMC_SCDR (System Clock Disable Register). The activity of UDPCK is shown in the bit
UDP of PMC_SCSR (System Clock Status Register).
USB Device Port Suspend When the USB Device Port detects a suspend condition, the 48 MHz clock is automati-
cally disabled, i.e., the UDP bit in PMC_SCSR is cleared. It is also possible to
automatically disable the Master Clock provided to the USB Device Port on a suspend
condition. The MCKUDP bit in PMC_SCSR configures this feature and can be set or
cleared by writing one in the same bit of PMC_SCER and PMC_SCDR.
USB Host Clock Control The USB Host Port clock UHPCK can be enabled by writing 1 at the UHP bit in
PMC_SCER (System Clock Enable Register) and disabled by writing 1 at the UHP bit in
PMC_SCDR (System Clock Disable Register). The activity of UDPCK is shown in the bit
UHP of PMC_SCSR (System Clock Status Register).
Programmable Clock Output The PMC controls up to four signals to be output on external pins PCK0 to PCK3. Each
Controller signal can be independently programmed via the registers PMC_PCK0 to PMC_PCK3.
PCK0 to PCK3 can be independently selected between the four clocks provided by the
Clock Generator by writing the CSS field in PMC_PCK0 to PMC_PCK3. Each output
signal can also be divided by a power of 2 between 1 and 64 by writing the field PRES
(Prescaler) in PMC_PCK0 to PMC_PCK3.
265
1768BATARM08/03
Each output signal can be enabled and disabled by writing 1 in the corresponding bit
PCK0 to PCK3 of PMC_SCER and PMC_SCDR, respectively. Status of the active pro-
grammable output clocks are given in the bits PCK0 to PCK3 of PMC_SCSR (System
Clock Status Register).
Moreover, like the MCK, a status bit in PMC_SR indicates that the Programmable Clock
is actually what has been programmed in the Programmable Clock registers.
As the Programmable Clock Controller does not manage with glitch prevention when
switching clocks, it is strongly recommended to disable the Programmable Clock before
any configuration change and to re-enable it after the change is actually performed.
Note also that it is required to assign the pin to the Programmable Clock operation in the
PIO Controller to enable the signal to be driven on the pin.
266 AT91RM9200
1768BATARM08/03
AT91RM9200
Master Clock Switching Table 60 gives the worst case timing required for the Master Clock to switch from one
Timings selected clock to another one. This is in the event that the prescaler is de-activated.
When the prescaler is activated, an additional time of 64 clock cycles of the new
selected clock has to be added.
267
1768BATARM08/03
Clock Switching Waveforms
Figure 125. Switch Master Clock from Slow Clock to PLLA Clock
Slow Clock
PLLA Clock
LOCK A
MCKRDY
Master Clock
Write PMC_MCKR
Figure 126. Switch Master Clock from Main Clock to Slow Clock
Slow Clock
Main Clock
MCKRDY
Master Clock
Write PMC_MCKR
268 AT91RM9200
1768BATARM08/03
AT91RM9200
Slow Clock
PLLA Clock
LOCKA
MCKRDY
Master Clock
Slow Clock
Write CKGR_PLLAR
PCKRDY
PCKx Output
269
1768BATARM08/03
Power Management Controller (PMC) User Interface
Table 61. Register Mapping
Offset Register Name Access Reset Value
0x0000 System Clock Enable Register PMC_SCER Write-only
0x0004 System Clock Disable Register PMC_SCDR Write-only
0x0008 System Clock Status Register PMC _SCSR Read-only 0x01
0x000C Reserved
0x0010 Peripheral Clock Enable Register PMC _PCER Write-only
0x0014 Peripheral Clock Disable Register PMC_PCDR Write-only
0x0018 Peripheral Clock Status Register PMC_PCSR Read-only 0x0
0x001C Reserved
0x0020 Main Oscillator Register CKGR_MOR ReadWrite 0x0
0x0024 Main Clock Frequency Register CKGR_MCFR Read-only -
0x0028 PLL A Register CKGR_PLLAR ReadWrite 0x3F00
0x002C PLL B Register CKGR_PLLBR ReadWrite 0x3F00
0x0030 Master Clock Register PMC_MCKR Read/Write 0x00
0x0034 Reserved
0x0038 Reserved
0x003C Reserved
0x0040 Programmable Clock 0 Register PMC_PCK0 Read/Write 0x0
0x0044 Programmable Clock 1 Register PMC_PCK1 Read/Write 0x0
0x0048 Programmable Clock 2 Register PMC_PCK2 Read/Write 0x0
0x004C Programmable Clock 3 Register PMC_PCK3 Read/Write 0x0
0x0050 Reserved
0x0054 Reserved
0x0058 Reserved
0x005C Reserved
0x0060 Interrupt Enable Register PMC_IER Write-only --
0x0064 Interrupt Disable Register PMC_IDR Write-only --
0x0068 Status Register PMC_SR Read-only --
0x006C Interrupt Mask Register PMC_IMR Read-only 0x0
270 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
PCK3 PCK2 PCK1 PCK0
7 6 5 4 3 2 1 0
UHP MCKUDP UDP PCK
271
1768BATARM08/03
PMC System Clock Disable Register
Register Name: PMC_SCDR
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
PCK3 PCK2 PCK1 PCK0
7 6 5 4 3 2 1 0
UHP MCKUDP UDP PCK
272 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
PCK3 PCK2 PCK1 PCK0
7 6 5 4 3 2 1 0
UHP MCKUDP UDP PCK
273
1768BATARM08/03
PMC Peripheral Clock Enable Register
Register Name: PMC_PCER
Access Type: Write-only
31 30 29 28 27 26 25 24
PID31 PID30 PID29 PID28 PID27 PID26 PID25 PID24
23 22 21 20 19 18 17 16
PID23 PID22 PID21 PID20 PID19 PID18 PID17 PID16
15 14 13 12 11 10 9 8
PID15 PID14 PID13 PID12 PID11 PID10 PID9 PID8
7 6 5 4 3 2 1 0
PID7 PID6 PID5 PID4 PID3 PID2
23 22 21 20 19 18 17 16
PID23 PID22 PID21 PID20 PID19 PID18 PID17 PID16
15 14 13 12 11 10 9 8
PID15 PID14 PID13 PID12 PID11 PID10 PID9 PID8
7 6 5 4 3 2 1 0
PID7 PID6 PID5 PID4 PID3 PID2
274 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
PID23 PID22 PID21 PID20 PID19 PID18 PID17 PID16
15 14 13 12 11 10 9 8
PID15 PID14 PID13 PID12 PID11 PID10 PID9 PID8
7 6 5 4 3 2 1 0
PID7 PID6 PID5 PID4 PID3 PID2
275
1768BATARM08/03
PMC Clock Generator Main Oscillator Register
Register Name: CKGR_MOR
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
OSCOUNT
7 6 5 4 3 2 1 0
- MOSCEN
276 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
MAINRDY
15 14 13 12 11 10 9 8
MAINF
7 6 5 4 3 2 1 0
MAINF
277
1768BATARM08/03
PMC Clock Generator PLL A Register
Register Name: CKGR_PLLAR
Access Type: Read/Write
31 30 29 28 27 26 25 24
1 MULA
23 22 21 20 19 18 17 16
MULA
15 14 13 12 11 10 9 8
OUTA PLLACOUNT
7 6 5 4 3 2 1 0
DIVA
Possible limitations on PLL A input frequencies and multiplier factors should be checked before using the Clock Generator.
DIVA: Divider A
278 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
MULB
15 14 13 12 11 10 9 8
OUTB PLLBCOUNT
7 6 5 4 3 2 1 0
DIVB
DIVB: Divider B
279
1768BATARM08/03
PMC Master Clock Register
Register Name: PMC_MCKR
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
MDIV
7 6 5 4 3 2 1 0
PRES CSS
Note: Value to be written in PMC_MCKR must not be the same as current value in PMC_MCKR.
CSS: Master Clock Selection
280 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
PRES CSS
281
1768BATARM08/03
PMC Interrupt Enable Register
Register Name: PMC_IER
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
PCK3RDY PCK2RDY PCK1RDY PCK0RDY
7 6 5 4 3 2 1 0
MCKRDY LOCKB LOCKA MOSCS
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
PCK3RDY PCK2RDY PCK1RDY PCK0RDY
7 6 5 4 3 2 1 0
MCKRDY LOCKB LOCKA MOSCS
282 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
PCK3RDY PCK2RDY PCK1RDY PCK0RDY
7 6 5 4 3 2 1 0
MCKRDY LOCKB LOCKA MOSCS
283
1768BATARM08/03
PMC Interrupt Mask Register
Register Name: PMC_IMR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
PCK3RDY PCK2RDY PCK1RDY PCK0RDY
7 6 5 4 3 2 1 0
MCKRDY LOCKB LOCKA MOSCS
284 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The System Timer (ST) module integrates three different free-running timers:
A Period Interval Timer (PIT) that sets the time base for an operating system.
A Watchdog Timer (WDT) with system reset capabilities in case of software deadlock.
A Real-Time Timer (RTT) counting elapsed seconds.
These timers count using the Slow Clock provided by the Power Management Controller. Typ-
ically, this clock has a frequency of 32.768 kHz, but the System Timer might be configured to
support another frequency.
The System Timer provides an interrupt line connected to one of the sources of the Advanced
Interrupt Controller (AIC). Interrupt handling requires programming the AIC before configuring
the System Timer. Usually, the System Timer interrupt line is connected to the first interrupt
source line and shares this entry with the Debug Unit (DBGU) and the Real Time Clock (RTC).
This sharing requires the programmer to determine the source of the interrupt when the
source 1 is triggered.
Important features of the System Timer include:
One Period Interval Timer, 16-bit Programmable Counter
One Watchdog Timer, 16-bit Programmable Counter
One Real-time Timer, 20-bit Free-running Counter
Interrupt Generation on Event
Real-Time Timer
Power NWDOVF
Management SLCK
Controller Watchdog Timer
STIRQ
285
1768BATARM08/03
Product
Dependencies
Power The System Timer is continuously clocked at 32768 Hz. The power management controller
Management has no effect on the system timer behavior.
Interrupt Sources The System Timer interrupt is generally connected to the source 1 of the Advanced Interrupt
Controller. This interrupt line is the result of the OR-wiring of the system peripheral interrupt
lines (System Timer, Real Time Clock, Power Management Controller, Memory Controller).
When a system interrupt happens, the service routine must first determine the cause of the
interrupt. This is accomplished by reading successively the status registers of the above men-
tioned system peripherals.
Watchdog The System Timer is capable of driving the NWDOVF pin. This pin might be implemented or
Overflow not in a product. When it is implemented, this pin might or not be multiplexed on the PIO Con-
trollers even though it is recommended to dedicate a pin to the watchdog function. If the
NWDOVF is multiplexed on a PIO Controller, this last should be first programmed to assign
the pin to the watchdog function before using the pin as NWDOVF.
When it is not implemented, programming the associated bits and registers has no effect on
the behavior of the System Timer.
Functional
Description
System Timer The System Timer uses only the SLCK clock so that it is capable to provide periodic, watch-
dog, second change or alarm interrupt even if the Power Management Controller is
Clock
programmed to put the product in Slow Clock Mode. If the product has the capability to back
up the Slow Clock oscillator and the System Timer, the System Timer can continue to operate.
Period Interval The Period Interval Timer can be used to provide periodic interrupts for use by operating sys-
Timer (PIT) tems. The reset value of the PIT is 0 corresponding to the maximum value. It is built around a
16-bit down counter, which is preloaded by a value programmed in ST_PIMR (Period Interval
Mode Register). When the PIT counter reaches 0, the bit PITS is set in ST_SR (Status Regis-
ter), and an interrupt is generated if it is enabled.
The counter is then automatically reloaded and restarted. Writing to the ST_PIMR at any time
immediately reloads and restarts the down counter with the new programmed value.
Warning: If ST_PIMR is programmed with a period less or equal to the current MCK period,
the update of the PITS status bit and its associated interrupt generation are unpredictable.
PIV
16-bit PITS
Down Counter
SLCK
Slow Clock
286 AT91RM9200
1768BATARM08/03
AT91RM9200
Watchdog Timer The Watchdog Timer can be used to prevent system lock-up if the software becomes trapped
(WDT) in a deadlock. It is built around a 16-bit down counter loaded with the value defined in
ST_WDMR (Watchdog Mode Register).
At reset, the value of the ST_WDMR is 0x00020000, corresponding to the maximum value of
the counter. The watchdog overflow signal is tied low during 8 slow clock cycles when a
watchdog overflow occurs (EXTEN bit set in ST_WDMR).
It uses the Slow Clock divided by 128 to establish the maximum watchdog period to be 256
seconds (with a typical slow clock of 32.768 kHz).
In normal operation, the user reloads the Watchdog at regular intervals before the timer over-
flow occurs, by setting the bit WDRST in the ST_CR (Control Register).
If an overflow does occur, the watchdog timer:
Sets the WDOVF bit in ST_SR (Status Register), from which an interrupt can be
generated.
Generates a pulse for 8 slow clock cycles on the external signal watchdog overflow if the
bit EXTEN in ST_WDMR is set.
Generates an internal reset if the parameter RSTEN in ST_WDMR is set.
Reloads and restarts the down counter.
Writing the ST_WDMR does not reload or restart the down counter. When the ST_CR is writ-
ten the watchdog counter is immediately reloaded from ST_WDMR and restarted and the
Slow Clock 128 divider is also immediately reset and restarted.
WV
NWDOVF
Real-time Timer The Real-Time Timer is used to count elapsed seconds. It is built around a 20-bit counter fed
(RTT) by Slow Clock divided by a programmable value. At reset, this value is set to 0x8000, corre-
sponding to feeding the real-time counter with a 1 Hz signal when the Slow Clock is 32.768
Hz. The 20-bit counter can count up to 1048576 seconds, corresponding to more than 12
days, then roll over to 0.
The Real-Time Timer value can be read at any time in the register ST_CRTR (Current Real-
time Register). As this value can be updated asynchronously to the master clock, it is advis-
able to read this register twice at the same value to improve accuracy of the returned value.
This current value of the counter is compared with the value written in the alarm register
ST_RTAR (Real-time Alarm Register). If the counter value matches the alarm, the bit ALMS in
TC_SR is set. The alarm register is set to its maximum value, corresponding to 0, after a reset.
The bit RTTINC in ST_SR is set each time the 20-bit counter is incremented. This bit can be
used to start an interrupt, or generate a one-second signal.
287
1768BATARM08/03
Writing the ST_RTMR immediately reloads and restarts the clock divider with the new pro-
grammed value. This also resets the 20-bit counter.
Warning: If RTPRES is programmed with a period less or equal to the current MCK period, the update of
the RTTINC and ALMS status bits and their associated interrupt generation are unpredictable.
20-bit
Counter
= ALMS
ALMV
288 AT91RM9200
1768BATARM08/03
AT91RM9200
ST Control Register
Register Name: ST_CR
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
WDRST
289
1768BATARM08/03
ST Period Interval Mode Register
Register Name: ST_PIMR
Access Type: Read/Write
PIV
PIV
23 22 21 20 19 18 17 16
EXTEN RSTEN
15 14 13 12 11 10 9 8
WDV
7 6 5 4 3 2 1 0
WDV
290 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RTPRES
7 6 5 4 3 2 1 0
RTPRES
ST Status Register
Register Name: ST_SR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
ALMS RTTINC WDOVF PITS
291
1768BATARM08/03
ST Interrupt Enable Register
Register Name: ST_IER
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
ALMS RTTINC WDOVF PITS
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
ALMS RTTINC WDOVF PITS
292 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
ALMS RTTINC WDOVF PITS
23 22 21 20 19 18 17 16
ALMV
15 14 13 12 11 10 9 8
ALMV
7 6 5 4 3 2 1 0
ALMV
293
1768BATARM08/03
ST Current Real-Time Register
Register Name: ST_CRTR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
CRTV
15 14 13 12 11 10 9 8
CRTV
7 6 5 4 3 2 1 0
CRTV
294 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Real-time Clock (RTC) peripheral is designed for very low power consumption.
It combines a complete time-of-day clock with alarm and a two-hundred-year Gregorian calen-
dar, complemented by a programmable periodic interrupt. The alarm and calendar registers
are accessed by a 32-bit data bus.
The time and calendar values are coded in binary-coded decimal (BCD) format. The time for-
mat can be 24-hour mode or 12-hour mode with an AM/PM indicator.
Updating time and calendar fields and configuring the alarm fields are performed by a parallel
capture on the 32-bit data bus. An entry control is performed to avoid loading registers with
incompatible BCD format data or with an incompatible date according to the current
month/year/century.
Important features of the RTC include:
Low Power Consumption
Full Asynchronous Design
Two Hundred Year Calendar
Programmable Periodic Interrupt
Alarm and Update Parallel Load
Control of Alarm and Update Time/Calendar Data In
Block Diagram
Figure 134. RTC Block Diagram
Product
Dependencies
Power The Real-time Clock is continuously clocked at 32768 Hz. The Power Management Controller
Management has no effect on RTC behavior.
Interrupt The RTC Interrupt is connected to interrupt source 1 (IRQ1) of the advanced interrupt control-
ler. This interrupt line is due to the OR-wiring of the system peripheral interrupt lines (System
Timer, Real Time Clock, Power Management Controller, Memory Controller, etc.). When a
295
1768BATARM08/03
system interrupt occurs, the service routine must first determine the cause of the interrupt.
This is done by reading the status registers of the above system peripherals successively.
Functional The RTC provides a full binary-coded decimal (BCD) clock that includes century (19/20), year
(with leap years), month, date, day, hours, minutes and seconds.
Description
The valid year range is 1900 to 2099, a two-hundred-year Gregorian calendar achieving full
Y2K compliance.
The RTC can operate in 24-hour mode or in 12-hour mode with an AM/PM indicator.
Corrections for leap years are included (all years divisible by 4 being leap years, including
year 2000). This is correct up to the year 2099.
After hardware reset, the calendar is initialized to Thursday, January 1, 1998.
Reference Clock The reference clock is Slow Clock (SLCK). It can be driven by the Atmel cell OSC55 or OSC56
(or an equivalent cell) and an external 32.768 kHz crystal.
During low power modes of the processor (idle mode), the oscillator runs and power consump-
tion is critical. The crystal selection has to take into account the current consumption for power
saving and the frequency drift due to temperature effect on the circuit for time accuracy.
Timing The RTC is updated in real time at one-second intervals in normal mode for the counters of
seconds, at one-minute intervals for the counter of minutes and so on.
Due to the asynchronous operation of the RTC with respect to the rest of the chip, to be cer-
tain that the value read in the RTC registers (century, year, month, date, day, hours, minutes,
seconds) are valid and stable, it is necessary to read these registers twice. If the data is the
same both times, then it is valid. Therefore, a minimum of two and a maximum of three
accesses are required.
Alarm The RTC has five programmable fields: month, date, hours, minutes and seconds.
Each of these fields can be enabled or disabled to match the alarm condition:
If all the fields are enabled, an alarm flag is generated (the corresponding flag is asserted
and an interrupt generated if enabled) at a given month, date, hour/minute/second.
If only the seconds field is enabled, then an alarm is generated every minute.
Depending on the combination of fields enabled, a large number of possibilities are available
to the user ranging from minutes to 365/366 days.
Error Checking Verification on user interface data is performed when accessing the century, year, month,
date, day, hours, minutes, seconds and alarms. A check is performed on illegal BCD entries
such as illegal date of the month with regard to the year and century configured.
If one of the time fields is not correct, the data is not loaded into the register/counter and a flag
is set in the validity register. The user can not reset this flag. It is reset as soon as an accept-
able value is programmed. This avoids any further side effects in the hardware. The same
procedure is done for the alarm.
The following checks are performed:
1. Century (check if it is in range 19 - 20)
2. Year (BCD entry check)
3. Date (check range 01 - 31)
296 AT91RM9200
1768BATARM08/03
AT91RM9200
Updating To update any of the time/calendar fields, the user must first stop the RTC by setting the corre-
Time/Calendar sponding field in the Control Register. Bit UPDTIM must be set to update time fields (hour,
minute, second) and bit UPDCAL must be set to update calendar fields (century, year, month,
date, day).
Then the user must poll or wait for the interrupt (if enabled) of bit ACKUPD in the Status Reg-
ister. Once the bit reads 1, the user can write to the appropriate register.
Once the update is finished, the user must reset (0) UPDTIM and/or UPDCAL in the Control
Register.
When programming the calendar fields, the time fields remain enabled. This avoids a time slip
in case the user stays in the calendar update phase for several tens of seconds or more. In
successive update operations, the user must wait at least one second after resetting the
UPDTIM/UPDCAL bit in the RTC_CR (Control Register) before setting these bits again. This
is done by waiting for the SEC flag in the Status Register before setting UPDTIM/UPDCAL bit.
After resetting UPDTIM/UPDCAL, the SEC flag must also be cleared.
297
1768BATARM08/03
Real Time Controller (RTC) User Interface
Table 63. RTC Register Mapping
Offset Register Register Name Read/Write Reset
0x00 RTC Control Register RTC_CR Read/Write 0x0
0x04 RTC Mode Register RTC_MR Read/Write 0x0
0x08 RTC Time Register RTC_TIMR Read/Write 0x0
0x0C RTC Calendar Register RTC_CALR Read/Write 0x01819819
0x10 RTC Time Alarm Register RTC_TIMALR Read/Write 0x0
0x14 RTC Calendar Alarm Register RTC_CALALR Read/Write 0x01010000
0x18 RTC Status Register RTC_SR Read only 0x0
0x1C RTC Status Clear Command Register RTC_SCCR Write only ---
0x20 RTC Interrupt Enable Register RTC_IER Write only ---
0x24 RTC Interrupt Disable Register RTC_IDR Write only ---
0x28 RTC Interrupt Mask Register RTC_IMR Read only 0x0
0x2C RTC Valid Entry Register RTC_VER Read only 0x0
298 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
CALEVSEL
15 14 13 12 11 10 9 8
TIMEVSEL
7 6 5 4 3 2 1 0
UPDCAL UPDTIM
299
1768BATARM08/03
RTC Mode Register
Name: RTC_MR
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
HRMOD
300 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
AMPM HOUR
15 14 13 12 11 10 9 8
MIN
7 6 5 4 3 2 1 0
SEC
301
1768BATARM08/03
RTC Calendar Register
Name: RTC_CALR
Access Type: Read/Write
31 30 29 28 27 26 25 24
DATE
23 22 21 20 19 18 17 16
DAY MONTH
15 14 13 12 11 10 9 8
YEAR
7 6 5 4 3 2 1 0
CENT
302 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
HOUREN AMPM HOUR
15 14 13 12 11 10 9 8
MINEN MIN
7 6 5 4 3 2 1 0
SECEN SEC
303
1768BATARM08/03
RTC Calendar Alarm Register
Name: RTC_CALALR
Access Type: Read/Write
31 30 29 28 27 26 25 24
DATEEN DATE
23 22 21 20 19 18 17 16
MTHEN MONTH
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
304 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
CALEV TIMEV SEC ALARM ACKUPD
305
1768BATARM08/03
RTC Status Clear Command Register
Name: RTC_SCCR
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
CALCLR TIMCLR SECCLR ALRCLR ACKCLR
306 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
CALEN TIMEN SECEN ALREN ACKEN
307
1768BATARM08/03
RTC Interrupt Disable Register
Name: RTC_IDR
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
CALDIS TIMDIS SECDIS ALRDIS ACKDIS
308 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
CAL TIM SEC ALR ACK
309
1768BATARM08/03
RTC Valid Entry Register
Name: RTC_VER
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
NVCALAR NVTIMALR NVCAL NVTIM
310 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Debug Unit provides a single entry point from the processor for access to all the debug
capabilities of Atmels ARM-based systems.
The Debug Unit features a two-pin UART that can be used for several debug and trace pur-
poses and offers an ideal medium for in-situ programming solutions and debug monitor
communications. Moreover, the association with two peripheral data controller channels per-
mits packet handling for these tasks with processor time reduced to a minimum.
The Debug Unit also makes the Debug Communication Channel (DCC) signals provided by
the In-circuit Emulator of the ARM processor visible to the software. These signals indicate the
status of the DCC read and write registers and generate an interrupt to the ARM processor,
making possible the handling of the DCC under interrupt control.
Chip Identifier registers permit recognition of the device and its revision. These registers
inform as to the sizes and types of the on-chip memories, as well as the set of embedded
peripherals.
Finally, the Debug Unit features a Force NTRST capability that enables the software to decide
whether to prevent access to the system via the In-circuit Emulator. This permits protection of
the code, stored in ROM.
Important features of the Debug Unit are:
System Peripheral to Facilitate Debug of Atmels ARM-based Systems
Composed of Four Functions
Two-pin UART
Debug Communication Channel (DCC) Support
Chip ID Registers
ICE Access Prevention
Two-pin UART
Implemented Features are 100% Compatible with the Standard Atmel USART
Independent Receiver and Transmitter with a Common Programmable Baud Rate
Generator
Even, Odd, Mark or Space Parity Generation
Parity, Framing and Overrun Error Detection
Automatic Echo, Local Loopback and Remote Loopback Channel Modes
Interrupt Generation
Support for Two PDC Channels with Connection to Receiver and Transmitter
Debug Communication Channel Support
Offers Visibility of COMMRX and COMMTX Signals from the ARM Processor
Interrupt Generation
Chip ID Registers
Identification of the Device Revision, Sizes of the Embedded Memories, Set of
Peripherals
ICE Access Prevention
Enables Software to Prevent System Access Through the ARM Processors ICE
Prevention is Made by Asserting the NTRST Line of the ARM Processors ICE
311
1768BATARM08/03
Block Diagram
Figure 135. Debug Unit Functional Block Diagram
Peripheral
Bridge
Transmit
Power MCK Baud Rate Parallel
Management Generator Input/
Controller Output
Receive
DRXD
COMMRX
DCC
ARM Handler Chip ID
COMMTX
Processor
nTRST
ICE Interrupt
Access Control
Handler
NTRST(1)
DBGU Interupt
Force NTRST
Other Advanced
Source 1
System Interrupt
Interrupt Controller
Sources
Debug Unit
RS232 Drivers
312 AT91RM9200
1768BATARM08/03
AT91RM9200
Product
Dependencies
I/O Lines Depending on product integration, the Debug Unit pins may be multiplexed with PIO lines. In
this case, the programmer must first configure the corresponding PIO Controller to enable I/O
lines operations of the Debug Unit.
Power Depending on product integration, the Debug Unit clock may be controllable through the
Management Power Management Controller. In this case, the programmer must first configure the PMC to
enable the Debug Unit clock. Usually, the peripheral identifier used for this purpose is 1.
Interrupt Source Depending on product integration, the Debug Unit interrupt line is connected to one of the
interrupt sources of the Advanced Interrupt Controller. Interrupt handling requires program-
ming of the AIC before configuring the Debug Unit. Usually, the Debug Unit interrupt line
connects to the interrupt source 1 of the AIC, which may be shared with the real-time clock,
the system timer interrupt lines and other system peripheral interrupts, as shown in Figure
135. This sharing requires the programmer to determine the source of the interrupt when the
source 1 is triggered.
UART The Debug Unit operates as a UART, (asynchronous mode only) and supports only 8-bit char-
acter handling (with parity). It has no clock pin.
Operations
The Debug Unit's UART is made up of a receiver and a transmitter that operate independently,
and a common baud rate generator. Receiver timeout and transmitter time guard are not
implemented. However, all the implemented features are compatible with those of a standard
USART.
Baud Rate The baud rate generator provides the bit period clock named baud rate clock to both the
Generator receiver and the transmitter.
The baud rate clock is the master clock divided by 16 times the value (CD) written in
DBGU_BRGR (Baud Rate Generator Register). If DBGU_BRGR is set to 0, the baud rate
clock is disabled and the Debug Unit's UART remains inactive. The maximum allowable baud
rate is Master Clock divided by 16. The minimum allowable baud rate is Master Clock divided
by (16 x 65536).
MCK
Baud Rate = ---------------------
16 CD
313
1768BATARM08/03
Figure 137. Baud Rate Generator
CD
CD
Receiver
Receiver Reset, After device reset, the Debug Unit receiver is disabled and must be enabled before being
Enable and Disable used. The receiver can be enabled by writing the control register DBGU_CR with the bit RXEN
at 1. At this command, the receiver starts looking for a start bit.
The programmer can disable the receiver by writing DBGU_CR with the bit RXDIS at 1. If the
receiver is waiting for a start bit, it is immediately stopped. However, if the receiver has already
detected a start bit and is receiving the data, it waits for the stop bit before actually stopping its
operation.
The programmer can also put the receiver in its reset state by writing DBGU_CR with the bit
RSTRX at 1. In doing so, the receiver immediately stops its current operations and is disabled,
whatever its current state. If RSTRX is applied when data is being processed, this data is lost.
Start Detection and The Debug Unit only supports asynchronous operations, and this affects only its receiver. The
Data Sampling Debug Unit receiver detects the start of a received character by sampling the DRXD signal
until it detects a valid start bit. A low level (space) on DRXD is interpreted as a valid start bit if
it is detected for more than 7 cycles of the sampling clock, which is 16 times the baud rate.
Hence, a space that is longer than 7/16 of the bit period is detected as a valid start bit. A space
which is 7/16 of a bit period or shorter is ignored and the receiver continues to wait for a valid
start bit.
When a valid start bit has been detected, the receiver samples the DRXD at the theoretical
midpoint of each bit. It is assumed that each bit lasts 16 cycles of the sampling clock (1-bit
period) so the bit sampling point is eight cycles (0.5-bit period) after the start of the bit. The first
sampling point is therefore 24 cycles (1.5-bit periods) after the falling edge of the start bit was
detected.
Each subsequent bit is sampled 16 cycles (1-bit period) after the previous one.
DRXD
True Start D0
Detection
Baud Rate
Clock
314 AT91RM9200
1768BATARM08/03
AT91RM9200
DRXD
Receiver Ready When a complete character is received, it is transferred to the DBGU_RHR and the RXRDY
status bit in DBGU_SR (Status Register) is set. The bit RXRDY is automatically cleared when
the receive holding register DBGU_RHR is read.
DRXD S D0 D1 D2 D3 D4 D5 D6 D7 P S D0 D1 D2 D3 D4 D5 D6 D7 P
RXRDY
Read DBGU_RHR
Receiver Overrun If DBGU_RHR has not been read by the software (or the Peripheral Data Controller) since the
last transfer, the RXRDY bit is still set and a new character is received, the OVRE status bit in
DBGU_SR is set. OVRE is cleared when the software writes the control register DBGU_CR
with the bit RSTSTA (Reset Status) at 1.
RXRDY
OVRE
RSTSTA
Parity Error Each time a character is received, the receiver calculates the parity of the received data bits,
in accordance with the field PAR in DBGU_MR. It then compares the result with the received
parity bit. If different, the parity error bit PARE in DBGU_SR is set at the same time the
RXRDY is set. The parity bit is cleared when the control register DBGU_CR is written with the
bit RSTSTA (Reset Status) at 1. If a new character is received before the reset status com-
mand is written, the PARE bit remains at 1.
RXRDY
PARE
315
1768BATARM08/03
Receiver Framing When a start bit is detected, it generates a character reception when all the data bits have
Error been sampled. The stop bit is also sampled and when it is detected at 0, the FRAME (Framing
Error) bit in DBGU_SR is set at the same time the RXRDY bit is set. The bit FRAME remains
high until the control register DBGU_CR is written with the bit RSTSTA at 1.
RXRDY
FRAME
Transmitter
Transmitter Reset, After device reset, the Debug Unit transmitter is disabled and it must be enabled before being
Enable and Disable used. The transmitter is enabled by writing the control register DBGU_CR with the bit TXEN at
1. From this command, the transmitter waits for a character to be written in the Transmit Hold-
ing Register DBGU_THR before actually starting the transmission.
The programmer can disable the transmitter by writing DBGU_CR with the bit TXDIS at 1. If
the transmitter is not operating, it is immediately stopped. However, if a character is being pro-
cessed into the Shift Register and/or a character has been written in the Transmit Holding
Register, the characters are completed before the transmitter is actually stopped.
The programmer can also put the transmitter in its reset state by writing the DBGU_CR with
the bit RSTTX at 1. This immediately stops the transmitter, whether or not it is processing
characters.
Transmit Format The Debug Unit transmitter drives the pin DTXD at the baud rate clock speed. The line is
driven depending on the format defined in the Mode Register and the data stored in the Shift
Register. One start bit at level 0, then the 8 data bits, from the lowest to the highest bit, one
optional parity bit and one stop bit at 1 are consecutively shifted out as shown on the following
figure. The field PARE in the mode register DBGU_MR defines whether or not a parity bit is
shifted out. When a parity bit is enabled, it can be selected between an odd parity, an even
parity, or a fixed space or mark bit.
Baud Rate
Clock
DTXD
Transmitter Control When the transmitter is enabled, the bit TXRDY (Transmitter Ready) is set in the status regis-
ter DBGU_SR. The transmission starts when the programmer writes in the Transmit Holding
Register DBGU_THR, and after the written character is transferred from DBGU_THR to the
Shift Register. The bit TXRDY remains high until a second character is written in DBGU_THR.
316 AT91RM9200
1768BATARM08/03
AT91RM9200
As soon as the first character is completed, the last character written in DBGU_THR is trans-
ferred into the shift register and TXRDY rises again, showing that the holding register is
empty.
When both the Shift Register and the DBGU_THR are empty, i.e., all the characters written in
DBGU_THR have been processed, the bit TXEMPTY rises after the last stop bit has been
completed.
TXRDY
TXEMPTY
Peripheral Data Both the receiver and the transmitter of the Debug Unit's UART are generally connected to a
Controller Peripheral Data Controller (PDC) channel.
The peripheral data controller channels are programmed via registers that are mapped within
the Debug Unit user interface from the offset 0x100. The status bits are reported in the Debug
Unit status register DBGU_SR and can generate an interrupt.
The RXRDY bit triggers the PDC channel data transfer of the receiver. This results in a read of
the data in DBGU_RHR. The TXRDY bit triggers the PDC channel data transfer of the trans-
mitter. This results in a write of a data in DBGU_THR.
Test Modes The Debug Unit supports three tests modes. These modes of operation are programmed by
using the field CHMODE (Channel Mode) in the mode register DBGU_MR.
The Automatic Echo mode allows bit-by-bit retransmission. When a bit is received on the
DRXD line, it is sent to the DTXD line. The transmitter operates normally, but has no effect on
the DTXD line.
The Local Loopback mode allows the transmitted characters to be received. DTXD and DRXD
pins are not used and the output of the transmitter is internally connected to the input of the
receiver. The DRXD pin level has no effect and the DTXD line is held high, as in idle state.
The Remote Loopback mode directly connects the DRXD pin to the DTXD line. The transmit-
ter and the receiver are disabled and have no effect. This mode allows a bit-by-bit
retransmission.
317
1768BATARM08/03
Figure 146. Test Modes
Automatic Echo
Receiver RXD
Disabled
Transmitter TXD
Local Loopback
Disabled
Receiver RXD
VDD
Disabled
Transmitter TXD
Disabled
Transmitter TXD
318 AT91RM9200
1768BATARM08/03
AT91RM9200
Debug The Debug Unit handles the signals COMMRX and COMMTX that come from the Debug
Communication Communication Channel of the ARM Processor and are driven by the In-circuit Emulator.
Channel Support The Debug Communication Channel contains two registers that are accessible through the
ICE Breaker on the JTAG side and through the coprocessor 0 on the ARM Processor side.
As a reminder, the following instructions are used to read and write the Debug Communication
Channel:
MRC p14, 0, Rd, c1, c0, 0
Chip Identifier The Debug Unit features two chip identifier registers, DBGU_CIDR (Chip ID Register) and
DBGU_EXID (Extension ID). Both registers contain a hard-wired value that is read-only. The
first register contains the following fields:
EXT - shows the use of the extension identifier register
NVPTYP and NVPSIZ - identifies the type of embedded non-volatile memory and its size
ARCH - identifies the set of embedded peripheral
SRAMSIZ - indicates the size of the embedded SRAM
EPROC - indicates the embedded ARM processor
VERSION - gives the revision of the silicon
The second register is device-dependent and reads 0 if the bit EXT is 0.
ICE Access The Debug Unit allows blockage of access to the system through the ARM processor's ICE
Prevention interface. This feature is implemented via the register Force NTRST (DBGU_FNR), that allows
assertion of the NTRST signal of the ICE Interface. Writing the bit FNTRST (Force NTRST) to
1 in this register prevents any activity on the TAP controller.
On standard devices, the bit FNTRST resets to 0 and thus does not prevent ICE access.
This feature is especially useful on custom ROM devices for customers who do not want their
on-chip code to be visible.
319
1768BATARM08/03
Debug Unit User Interface
Table 65. Debug Unit Memory Map
Offset Register Name Access Reset Value
0x0000 Control Register DBGU_CR Write-only
0x0004 Mode Register DBGU_MR Read/Write 0x0
0x0008 Interrupt Enable Register DBGU_IER Write-only
0x000C Interrupt Disable Register DBGU_IDR Write-only
0x0010 Interrupt Mask Register DBGU_IMR Read-only 0x0
0x0014 Status Register DBGU_SR Read-only
0x0018 Receive Holding Register DBGU_RHR Read-only 0x0
0x001C Transmit Holding Register DBGU_THR Write-only
0x0020 Baud Rate Generator Register DBGU_BRGR Read/Write 0x0
0x0024 - 0x003C Reserved
0X0040 Chip ID Register DBGU_CIDR Read-only
0X0044 Chip ID Extension Register DBGU_EXID Read-only
0X0048 Force NTRST Register DBGU_FNR Read/Write 0x0
0x004C - 0x00FC Reserved
0x0100 - 0x0124 PDC Area
320 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RSTSTA
7 6 5 4 3 2 1 0
TXDIS TXEN RXDIS RXEN RSTTX RSTRX
321
1768BATARM08/03
Debug Unit Mode Register
Name: DBGU_MR
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
CHMODE PAR
7 6 5 4 3 2 1 0
322 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RXBUFF TXBUFE TXEMPTY
7 6 5 4 3 2 1 0
PARE FRAME OVRE ENDTX ENDRX TXRDY RXRDY
323
1768BATARM08/03
Debug Unit Interrupt Disable Register
Name: DBGU_IDR
Access Type: Write-only
31 30 29 28 27 26 25 24
COMMRX COMMTX
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RXBUFF TXBUFE TXEMPTY
7 6 5 4 3 2 1 0
PARE FRAME OVRE ENDTX ENDRX TXRDY RXRDY
324 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RXBUFF TXBUFE TXEMPTY
7 6 5 4 3 2 1 0
PARE FRAME OVRE ENDTX ENDRX TXRDY RXRDY
325
1768BATARM08/03
Debug Unit Status Register
Name: DBGU_SR
Access Type: Read-only
31 30 29 28 27 26 25 24
COMMRX COMMTX
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RXBUFF TXBUFE TXEMPTY
7 6 5 4 3 2 1 0
PARE FRAME OVRE ENDTX ENDRX TXRDY RXRDY
326 AT91RM9200
1768BATARM08/03
AT91RM9200
327
1768BATARM08/03
Debug Unit Receiver Holding Register
Name: DBGU_RHR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
RXCHR
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
TXCHR
328 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
CD
7 6 5 4 3 2 1 0
CD
329
1768BATARM08/03
Debug Unit Chip ID Register
Name: DBGU_CIDR
Access Type: Read-only
31 30 29 28 27 26 25 24
EXT NVPTYP ARCH
23 22 21 20 19 18 17 16
ARCH SRAMSIZ
15 14 13 12 11 10 9 8
0 0 0 0 NVPSIZ
7 6 5 4 3 2 1 0
EPROC VERSION
EPROC Processor
0 0 1 ARM946ES
0 1 0 ARM7TDMI
1 0 0 ARM920T
NVPSIZ Size
0 0 0 0 None
0 0 0 1 8K bytes
0 0 1 0 16K bytes
0 0 1 1 32K bytes
0 1 0 0 Reserved
0 1 0 1 64K bytes
0 1 1 0 Reserved
0 1 1 1 128K bytes
1 0 0 0 Reserved
1 0 0 1 256K bytes
1 0 1 0 Reserved
1 0 1 1 Reserved
1 1 0 0 Reserved
1 1 0 1 Reserved
1 1 1 0 Reserved
1 1 1 1 Reserved
330 AT91RM9200
1768BATARM08/03
AT91RM9200
SRAMSIZ Size
0 0 0 0 Reserved
0 0 0 1 1K bytes
0 0 1 0 2K bytes
0 0 1 1 Reserved
0 1 0 0 Reserved
0 1 0 1 4K bytes
0 1 1 0 Reserved
0 1 1 1 Reserved
1 0 0 0 8K bytes
1 0 0 1 16K bytes
1 0 1 0 32K bytes
1 0 1 1 64K bytes
1 1 0 0 128K bytes
1 1 0 1 256K bytes
1 1 1 0 96K bytes
1 1 1 1 512K bytes
ARCH
Hex Dec Architecture
0x40 0100 0000 AT91x40 Series
0x63 0110 0011 AT91x63 Series
0x55 0101 0101 AT91x55 Series
0x42 0100 0010 AT91x42 Series
0x92 1001 0010 AT91x92 Series
0x34 0011 0100 AT91x34 Series
NVPTYP Memory
0 0 0 ROM
0 0 1 ROMless or on-chip Flash
1 0 0 SRAM emulating ROM
331
1768BATARM08/03
Debug Unit Chip ID Extension Register
Name: DBGU_EXID
Access Type: Read-only
31 30 29 28 27 26 25 24
EXID
23 22 21 20 19 18 17 16
EXID
15 14 13 12 11 10 9 8
EXID
7 6 5 4 3 2 1 0
EXID
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
FNTRST
332 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Parallel Input/Output Controller (PIO) manages up to 32 fully programmable input/output
lines. Each I/O line may be dedicated as a general-purpose I/O or be assigned to a function of
an embedded peripheral. This assures effective optimization of the pins of a product.
Each I/O line is associated with a bit number in all of the 32-bit registers of the 32-bit wide
User Interface.
Each I/O line of the PIO Controller features:
An input change interrupt enabling level change on any I/O line.
A glitch filter providing rejection of pulses lower than one-half of clock cycle.
Multi-drive capability similar to an open drain I/O line.
Control of the the pull-up of the I/O line.
Input visibility and output control.
The PIO Controller also features a synchronous output providing up to 32 bits of data output in
a single write operation.
Important features of the PIO also include:
Up to 32 Programmable I/O Lines
Fully Programmable through Set/Clear Registers
Multiplexing of Two Peripheral Functions per I/O Line
For each I/O Line (Whether Assigned to a Peripheral or Used as General Purpose I/O)
Input Change Interrupt
Glitch Filter
Multi-drive Option Enables Driving in Open Drain
Programmable Pull Up on Each I/O Line
Pin Data Status Register, Supplies Visibility of the Level on the Pin at Any Time
Synchronous Output, Provides Set and Clear of Several I/O lines in a Single Write
333
1768BATARM08/03
Block Diagram Figure 147. Block Diagram
PIO Controller
Embedded
Peripheral
Embedded Up to 32
Peripheral peripheral IOs
Embedded
Peripheral
PIN
Embedded
Peripheral PIN Up to 32 pins
Embedded Up to 32
Peripheral peripheral IOs
Embedded
Peripheral PIN
APB
PIO Controller
334 AT91RM9200
1768BATARM08/03
AT91RM9200
Product Dependencies
Pin Multiplexing Each pin is configurable, according to product definition as either a general-purpose I/O line
only, or as an I/O line multiplexed with one or two peripheral I/Os. As the multiplexing is hard-
ware-defined and thus product-dependent, the hardware designer and programmer must
carefully determine the configuration of the PIO controllers required by their application. When
an I/O line is general-purpose only, i.e. not multiplexed with any peripheral I/O, programming
of the PIO Controller regarding the assignment to a peripheral has no effect and only the PIO
Controller can control how the pin is driven by the product.
External Interrupt The interrupt signals FIQ and IRQ0 to IRQn are most generally multiplexed through the PIO
Lines Controllers. However, it is not necessary to assign the I/O line to the interrupt function as the
PIO Controller has no effect on inputs and the interrupt lines (FIQ or IRQs) are used only as
inputs.
Power The Power Management Controller controls the PIO Controller clock in order to save power.
Management Writing any of the registers of the user interface does not require the PIO Controller clock to be
enabled. This means that the configuration of the I/O lines does not require the PIO Controller
clock to be enabled.
However, when the clock is disabled, not all of the features of the PIO Controller are available.
Note that the Input Change Interrupt and the read of the pin level require the clock to be
validated.
After a hardware reset, the PIO clock is disabled by default (see Power Management
Controller).
The user must configure the Power Management Controller before any access to the input line
information.
Interrupt For interrupt handling, the PIO Controllers are considered as user peripherals. This means
Generation that the PIO Controller interrupt lines are connected among the interrupt sources 2 to 31.
Refer to the PIO Controller peripheral identifier in the product description to identify the inter-
rupt sources dedicated to the PIO Controllers.
The PIO Controller interrupt can be generated only if the PIO Controller clock is enabled.
335
1768BATARM08/03
Functional The PIO Controller features up to 32 fully-programmable I/O lines. Most of the control logic
Description associated to each I/O is represented in Figure 149.
1 PIO_PUDR
Peripheral A
Output Enable 0
0
Peripheral B
0
Output Enable 1
PIO_ASR PIO_PER
PIO_ABSR PIO_PSR 1
Peripheral B 1
PIO_SODR 1
Output
PIO_ODSR 1 Pad
PIO_CODR 0
Peripheral A
Input
Peripheral B
PIO_PDSR PIO_ISR Input
0
Edge
Detector 1
Glitch 1
Filter
0 PIO Interrupt
PIO_IFER
PIO_IFSR PIO_IER
PIO_IFDR PIO_IMR
PIO_IDR
336 AT91RM9200
1768BATARM08/03
AT91RM9200
Pull-up Resistor Each I/O line is designed with an embedded pull-up resistor. The value of this resistor is about
Control 100 k (see the product electrical characteristics for more details about this value). The pull-
up resistor can be enabled or disabled by writing respectively PIO_PUER (Pull-up Enable
Register) and PIO_PUDR (Pull-up Disable Resistor). Writing in these registers results in set-
ting or clearing the corresponding bit in PIO_PUSR (Pull-up Status Register). Reading a 1 in
PIO_PUSR means the pull-up is disabled and reading a 0 means the pull-up is enabled.
Control of the pull-up resistor is possible regardless of the configuration of the I/O line.
After reset, all of the pull-ups are enabled, i.e. PIO_PUSR resets at the value 0x0.
I/O Line or When a pin is multiplexed with one or two peripheral functions, the selection is controlled with
Peripheral the registers PIO_PER (PIO Enable Register) and PIO_PDR (PIO Disable Register). The reg-
Function Selection ister PIO_PSR (PIO Status Register) is the result of the set and clear registers and indicates
whether the pin is controlled by the corresponding peripheral or by the PIO Controller. A value
of 0 indicates that the pin is controlled by the corresponding on-chip peripheral selected in the
PIO_ABSR (AB Select Status Register). A value of 1 indicates the pin is controlled by the PIO
controller.
If a pin is used as a general purpose I/O line (not multiplexed with an on-chip peripheral),
PIO_PER and PIO_PDR have no effect and PIO_PSR returns 1 for the corresponding bit.
After reset, most generally, the I/O lines are controlled by the PIO controller, i.e. PIO_PSR
resets at 1. However, in some events, it is important that PIO lines are controlled by the
peripheral (as in the case of memory chip select lines that must be driven inactive after reset
or for address lines that must be driven low for booting out of an external memory). Thus, the
reset value of PIO_PSR is defined at the product level, depending on the multiplexing of the
device.
Peripheral A or B The PIO Controller provides multiplexing of up to two peripheral functions on a single pin. The
Selection selection is performed by writing PIO_ASR (A Select Register) and PIO_BSR (Select B Regis-
ter). PIO_ABSR (AB Select Status Register) indicates which peripheral line is currently
selected. For each pin, the corresponding bit at level 0 means peripheral A is selected
whereas the corresponding bit at level 1 indicates that peripheral B is selected.
Note that multiplexing of peripheral lines A and B only affects the output line. The peripheral
input lines are always connected to the pin input.
After reset, PIO_ABSR is 0, thus indicating that all the PIO lines are configured on peripheral
A. However, peripheral A generally does not drive the pin as the PIO Controller resets in I/O
line mode.
Writing in PIO_ASR and PIO_BSR manages PIO_ABSR regardless of the configuration of the
pin. However, assignment of a pin to a peripheral function requires a write in the correspond-
ing peripheral selection register (PIO_ASR or PIO_BSR) in addition to a write in PIO_PDR.
Output Control When the I/0 line is assigned to a peripheral function, i.e. the corresponding bit in PIO_PSR is
at 0, the drive of the I/O line is controlled by the peripheral. Peripheral A or B, depending on
the value in PIO_ABSR, determines whether the pin is driven or not.
When the I/O line is controlled by the PIO controller, the pin can be configured to be driven.
This is done by writing PIO_OER (Output Enable Register) and PIO_PDR (Output Disable
Register). The results of these write operations are detected in PIO_OSR (Output Status Reg-
ister). When a bit in this register is at 0, the corresponding I/O line is used as an input only.
When the bit is at 1, the corresponding I/O line is driven by the PIO controller.
337
1768BATARM08/03
The level driven on an I/O line can be determined by writing in PIO_SODR (Set Output Data
Register) and PIO_CODR (Clear Output Data Register). These write operations respectively
set and clear PIO_ODSR (Output Data Status Register), which represents the data driven on
the I/O lines. Writing in PIO_OER and PIO_ODR manages PIO_OSR whether the pin is con-
figured to be controlled by the PIO controller or assigned to a peripheral function. This enables
configuration of the I/O line prior to setting it to be managed by the PIO Controller.
Similarly, writing in PIO_SODR and PIO_CODR effects PIO_ODSR. This is important as it
defines the first level driven on the I/O line.
Synchronous Data Using the write operations in PIO_SODR and PIO_CODR can require that several instructions
Output be executed in order to define values on several bits. Both clearing and setting I/O lines on an
8-bit port, for example, cannot be done at the same time, and thus might limit the application
covered by the PIO Controller.
To avoid these inconveniences, the PIO Controller features a Synchronous Data Output to
clear and set a number of I/O lines in a single write. This is performed by authorizing the writ-
ing of PIO_ODSR (Output Data Status Register) from the register set PIO_OWER (Output
Write Enable Register), PIO_OWDR (Output Write Disable Register) and PIO_OWSR (Output
Write Status Register). The value of PIO_OWSR register is user-definable by writing in
PIO_OWER and PIO_OWDR. It is used by the PIO Controller as a PIO_ODSR write authori-
zation mask. Authorizing the write of PIO_ODSR on a user-definable number of bits is
especially useful, as it guarantees that the unauthorized bit will not be changed when writing it
and thus avoids the need of a time consuming read-modify-write operation.
After reset, the synchronous data output is disabled on all the I/O lines as PIO_OWSR resets
at 0x0.
Multi Drive Control Each I/O can be independently programmed in Open Drain by using the Multi Drive feature.
(Open Drain) This feature permits several drivers to be connected on the I/O line which is driven low only by
each device. An external pull-up resistor (or enabling of the internal one) is generally required
to guarantee a high level on the line.
The Multi Drive feature is controlled by PIO_MDER (Multi-driver Enable Register) and
PIO_MDDR (Multi-driver Disable Register). The Multi Drive can be selected whether the I/O
line is controlled by the PIO controller or assigned to a peripheral function. PIO_MDSR (Multi-
driver Status Register) indicates the pins that are configured to support external drivers.
After reset, the Multi Drive feature is disabled on all pins, i.e. PIO_MDSR resets at value 0x0.
Output Line Figure 150 shows how the outputs are driven either by writing PIO_SODR or PIO_CODR, or
Timings by directly writing PIO_ODSR. This last case is valid only if the corresponding bit in
PIO_OWSR is set. Figure 150 also shows when the feedback in PIO_PDSR is available.
338 AT91RM9200
1768BATARM08/03
AT91RM9200
MCK
Write PIO_SODR
Write PIO_ODSR at 1 APB Access
Write PIO_CODR
Write PIO_ODSR at 0 APB Access
PIO_ODSR
2 Cycles 2 Cycles
PIO_PDSR
Inputs The level on each I/O line can be read through PIO_PDSR (Peripheral Data Status Register).
This register indicates the level of the I/O lines regardless of their configuration, whether
uniquely as an input or driven by the PIO controller or driven by a peripheral.
Reading the I/O line levels requires the clock of the PIO controller to be enabled, otherwise
PIO_PDSR reads the levels present on the I/O line at the time the clock was disabled.
Input Glitch Optional input glitch filters are independently programmable on each I/O line. When the glitch
Filtering filter is enabled, a glitch with a duration of less than 1/2 Master Clock (MCK) cycle is automat-
ically rejected, while a pulse with a duration of 1 Master Clock cycle or more is accepted. For
pulse durations between 1/2 Master Clock cycle and 1 Master Clock cycle the pulse may or
may not be taken into account, depending on the precise timing of its occurrence. Thus for a
pulse to be visible it must exceed 1 Master Clock cycle, whereas for a glitch to be reliably fil-
tered out, its duration must not exceed 1/2 Master Clock cycle. The filter introduces one
Master Clock cycle latency if the pin level change occurs before a rising edge. However, this
latency does not appear if the pin level change occurs before a falling edge. This is illustrated
in Figure 151.
The glitch filters are controlled by the register set; PIO_IFER (Input Filter Enable Register),
PIO_IFDR (Input Filter Disable Register) and PIO_IFSR (Input Filter Status Register). Writing
PIO_IFER and PIO_IFDR respectively sets and clears bits in PIO_IFSR. This last register
enables the glitch filter on the I/O lines.
When the glitch filter is enabled, it does not modify the behavior of the inputs on the peripher-
als. It acts only on the value read in PIO_PDSR and on the input change interrupt detection.
The glitch filters require that the PIO Controller clock is enabled.
MCK
Pin Level
1 cycle 1 cycle 1 cycle 1 cycle
PIO_PDSR
if PIO_IFSR = 0
2 cycles 1 cycle
PIO_PDSR
if PIO_IFSR = 1
339
1768BATARM08/03
Input Change The PIO Controller can be programmed to generate an interrupt when it detects an input
Interrupt change on an I/O line. The Input Change Interrupt is controlled by writing PIO_IER (Interrupt
Enable Register) and PIO_IDR (Interrupt Disable Register), which respectively enable and
disable the input change interrupt by setting and clearing the corresponding bit in PIO_IMR
(Interrupt Mask Register). As Input change detection is possible only by comparing two suc-
cessive samplings of the input of the I/O line, the PIO Controller clock must be enabled. The
Input Change Interrupt is available, regardless of the configuration of the I/O line, i.e. config-
ured as an input only, controlled by the PIO Controller or assigned to a peripheral function.
When an input change is detected on an I/O line, the corresponding bit in PIO_ISR (Interrupt
Status Register) is set. If the corresponding bit in PIO_IMR is set, the PIO Controller interrupt
line is asserted. The interrupt signals of the thirty-two channels are ORed-wired together to
generate a single interrupt signal to the Advanced Interrupt Controller.
When the software reads PIO_ISR, all the interrupts are automatically cleared. This signifies
that all the interrupts that are pending when PIO_ISR is read must be handled.
MCK
PIO_PDSR
PIO_ISR
340 AT91RM9200
1768BATARM08/03
AT91RM9200
I/O Lines The programing example shown in Table 66 below is used to define the following
Programming configuration.
4-bit output port on I/O lines 0 to 3, (should be written in a single write operation), open-
Example drain, with pull-up resistor
Four output signals on I/O lines 4 to 7 (to drive LEDs for example), driven high and low, no
pull-up resistor
Four input signals on I/O lines 8 to 11 (to read push-button states for example), with pull-
up resistors, glitch filters and input change interrupts
Four input signals on I/O line 12 to 15 to read an external device status (polled, thus no
input change interrupt), no pull-up resistor, no glitch filter
I/O lines 16 to 19 assigned to peripheral A functions with pull-up resistor
I/O lines 20 to 23 assigned to peripheral B functions, no pull-up resistor
I/O lines 24 to 27 assigned to peripheral A with Input Change Interrupt and pull-up resistor
341
1768BATARM08/03
Parallel Input/Output Controller (PIO) User Interface
Each I/O line controlled by the PIO Controller is associated with a bit in each of the PIO Controller User Interface registers.
Each register is 32 bits wide. If a parallel I/O line is not defined, writing to the corresponding bits has no effect. Undefined
bits read zero. If the I/O line is not multiplexed with any peripheral, the I/O line is controlled by the PIO Controller and
PIO_PSR returns 1 systematically.
342 AT91RM9200
1768BATARM08/03
AT91RM9200
343
1768BATARM08/03
PIO Enable Register
Name: PIO_PER
Access Type: Write-only
31 30 29 28 27 26 25 24
P31 P30 P29 P28 P27 P26 P25 P24
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
344 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
345
1768BATARM08/03
PIO Output Disable Register
Name: PIO_ODR
Access Type: Write-only
31 30 29 28 27 26 25 24
P31 P30 P29 P28 P27 P26 P25 P24
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
346 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
347
1768BATARM08/03
PIO Input Filter Status Register
Name: PIO_IFSR
Access Type: Read-only
31 30 29 28 27 26 25 24
P31 P30 P29 P28 P27 P26 P25 P24
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
348 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
349
1768BATARM08/03
PIO Pin Data Status Register
Name: PIO_PDSR
Access Type: Read-only
31 30 29 28 27 26 25 24
P31 P30 P29 P28 P27 P26 P25 P24
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
350 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
351
1768BATARM08/03
PIO Interrupt Status Register
Name: PIO_IMR
Access Type: Read-only
31 30 29 28 27 26 25 24
P31 P30 P29 P28 P27 P26 P25 P24
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
352 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
353
1768BATARM08/03
PIO Pull Up Disable Register
Name: PIO_PUDR
Access Type: Write-only
31 30 29 28 27 26 25 24
P31 P30 P29 P28 P27 P26 P25 P24
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
354 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
355
1768BATARM08/03
PIO Peripheral B Select Register
Name: PIO_BSR
Access Type: Write-only
31 30 29 28 27 26 25 24
P31 P30 P29 P28 P27 P26 P25 P24
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
356 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
357
1768BATARM08/03
PIO Output Write Status Register
Name: PIO_OWSR
Access Type: Read-only
31 30 29 28 27 26 25 24
P31 P30 P29 P28 P27 P26 P25 P24
23 22 21 20 19 18 17 16
P23 P22 P21 P20 P19 P18 P17 P16
15 14 13 12 11 10 9 8
P15 P14 P13 P12 P11 P10 P9 P8
7 6 5 4 3 2 1 0
P7 P6 P5 P4 P3 P2 P1 P0
358 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Serial Peripheral Interface (SPI) circuit is a synchronous serial data link that provides
communication with external devices in Master or Slave Mode. It also allows communication
between processors if an external processor is connected to the system.
The Serial Peripheral Interface is a shift register that serially transmits data bits to other SPIs.
During a data transfer, one SPI system acts as the master that controls the data flow, while the
other system acts as the slave, having data shifted into and out of it by the master. Different
CPUs can take turn being masters (Multiple Master Protocol versus Single Master Protocol
where one CPU is always the master while all of the others are always slaves), and one mas-
ter may simultaneously shift data into multiple slaves. However, only one slave may drive its
output to write data back to the master at any given time.
A slave device is selected when the master asserts its NSS signal. If multiple slave devices
exist, the master generates a separate slave select signal for each slave (NPCS).
The SPI system consists of two data lines and two control lines:
Master Out Slave In (MOSI): This data line supplies the output data from the master
shifted into the input(s) of the slave(s).
Master In Slave Out (MISO): This data line supplies the output data from a slave to the
input of the master. There may be no more than one slave transmitting data during any
particular transfer.
Serial Clock (SPCK): This control line is driven by the master and regulates the flow of the
data bits. The master may transmit data at a variety of baud rates; the SPCK line cycles
once for each bit that is transmitted.
Slave Select (NSS): This control line allows slaves to be turned on and off by hardware.
The main features of the SPI are:
Supports Communication with Serial External Devices
4 Chip Selects with External Decoder Support Allow Communication with Up to 15
Peripherals
Serial Memories, such as DataFlash and 3-wire EEPROMs
Serial Peripherals, such as ADCs, DACs, LCD Controllers, CAN Controllers and
Sensors
External Co-processors
Master or Slave Serial Peripheral Bus Interface
8- to 16-bit Programmable Data Length Per Chip Select
Programmable Phase and Polarity Per Chip Select
Programmable Transfer Delays Between Consecutive Transfers and Between
Clock and Data Per Chip Select
Programmable Delay Between Consecutive Transfers
Selectable Mode Fault Detection
Connection to PDC Channel Capabilities Optimizes Data Transfers
One Channel for the Receiver, One Channel for the Transmitter
Next Buffer Support
359
1768BATARM08/03
Block Diagram
Figure 153. Block Diagram
ASB
APB Bridge
PDC
APB
SPCK
MISO
MOSI
PMC MCK
NPCS1
NPCS2
Interrupt Control
NPCS3
SPI Interrupt
360 AT91RM9200
1768BATARM08/03
AT91RM9200
SPCK SPCK
MISO MISO
Slave 0
MOSI MOSI
SPCK
NPCS1
MISO
NPCS2 NC Slave 1
NPCS3 MOSI
NSS
SPCK
MISO
Slave 2
MOSI
NSS
361
1768BATARM08/03
Product Dependencies
I/O Lines The pins used for interfacing the compliant external devices may be multiplexed with PIO
lines. The programmer must first program the PIO controllers to assign the SPI pins to their
peripheral functions.
Power The SPI may be clocked through the Power Management Controller (PMC), thus the program-
Management mer must first have to configure the PMC to enable the SPI clock.
Interrupt The SPI interface has an interrupt line connected to the Advanced Interrupt Controller (AIC).
Handling the SPI interrupt requires programming the AIC before configuring the SPI.
Functional Description
Master Mode When configured in Master Mode, the Serial Peripheral Interface controls data transfers to and
Operations from the slave(s) connected to the SPI bus. The SPI drives the chip select(s) to the slave(s)
and the serial clock (SPCK). After enabling the SPI, a data transfer begins when the core
writes to the SPI_TDR (Transmit Data Register).
Transmit and Receive buffers maintain the data flow at a constant rate with a reduced require-
ment for high-priority interrupt servicing. When new data is available in the SPI_TDR, the SPI
continues to transfer data. If the SPI_RDR (Receive Data Register) has not been read before
new data is received, the Overrun Error (OVRES) flag is set.
Note: As long as this flag is set, no data is loaded in the SPI_RDR. The user has to read the status
register to clear it.
The programmable delay between the activation of the chip select and the start of the data
transfer (DLYBS), as well as the delay between each data transfer (DLYBCT), can be pro-
grammed for each of the four external chip selects. All data transfer characteristics, including
the two timing values, are programmed in registers SPI_CSR0 to SPI_CSR3 (Chip Select
Registers).
In Master Mode, the peripheral selection can be defined in two different ways:
Fixed Peripheral Select: SPI exchanges data with only one peripheral
Variable Peripheral Select: Data can be exchanged with more than one peripheral
Figure 159 and Figure 160 show the operation of the SPI in Master Mode. For details concern-
ing the flag and control bits in these diagrams, see the tables in the Programmers Model,
starting in Section .
Fixed Peripheral This mode is used for transferring memory blocks without the extra overhead in the transmit
Select data register to determine the peripheral.
Fixed Peripheral Select is activated by setting bit PS to zero in SPI_MR (Mode Register). The
peripheral is defined by the PCS field in SPI_MR.
This option is only available when the SPI is programmed in Master Mode.
Variable Peripheral Variable Peripheral Select is activated by setting bit PS to one. The PCS field in SPI_TDR is
Select used to select the destination peripheral. The data transfer characteristics are changed when
the selected peripheral changes, according to the associated chip select register.
The PCS field in the SPI_MR has no effect.
This option is only available when the SPI is programmed in Master Mode.
362 AT91RM9200
1768BATARM08/03
AT91RM9200
Chip Selects The Chip Select lines are driven by the SPI only if it is programmed in Master Mode. These
lines are used to select the destination peripheral. The PCSDEC field in SPI_MR (Mode Reg-
ister) selects one to four peripherals (PCSDEC = 0) or up to 15 peripherals (PCSDEC = 1).
If Variable Peripheral Select is active, the chip select signals are defined for each transfer in
the PCS field in SPI_TDR. Chip select signals can thus be defined independently for each
transfer.
If Fixed Peripheral Select is active, Chip Select signals are defined for all transfers by the field
PCS in SPI_MR. If a transfer with a new peripheral is necessary, the software must wait until
the current transfer is completed, then change the value of PCS in SPI_MR before writing new
data in SPI_TDR.
The value on the NPCS pins at the end of each transfer can be read in the SPI_RDR (Receive
Data Register).
By default, all NPCS signals are high (equal to one) before and after each transfer.
Clock Generation and The SPI Baud rate clock is generated by dividing the Master Clock (MCK) or the Master Clock
Transfer Delays divided by 32 (if DIV32 is set in the Mode Register) by a value between 4 and 510. The divisor
is defined in the SCBR field in each Chip Select Register. The transfer speed can thus be
defined independently for each chip select signal.
Figure 155 shows a chip select transfer change and consecutive transfers on the same chip
selects. Three delays can be programmed to modify the transfer waveforms:
Delay between chip selects, programmable only once for all the chip selects by writing the
field DLYBCS in the Mode Register. Allows insertion of a delay between release of one
chip select and before assertion of a new one.
Delay before SPCK, independently programmable for each chip select by writing the field
DLYBS. Allows the start of SPCK to be delayed until after the chip select has been
asserted.
Delay between consecutive transfers, independently programmable for each chip select by
writing the field DLYBCT. Allows insertion of a delay between two transfers occurring on
the same chip select
These delays allow the SPI to be adapted to the interfaced peripherals and their speed and
bus release time.
Chip Select 1
Chip Select 2
SPCK
363
1768BATARM08/03
Mode Fault Detection A mode fault is detected when the SPI is programmed in Master Mode and a low level is
driven by an external master on the NPCS0/NSS signal.
When a mode fault is detected, the MODF bit in the SPI_SR is set until the SPI_SR is read
and the SPI is disabled until re-enabled by bit SPIEN in the SPI_CR (Control Register).
By default, Mode Fault Detection is enabled. It is disabled by setting the MODFDIS bit in the
SPI Mode Register.
364 AT91RM9200
1768BATARM08/03
AT91RM9200
1
TDRE
0 Fixed peripheral
PS
1 Variable peripheral
Delay DLYBS
Serializer = SPI_TDR(TD)
TDRE = 1
Data Transfer
SPI_RDR(RD) = Serializer
RDRF = 1
Delay DLYBCT
0
TDRE
1 0 Fixed peripheral
PS
NPCS = 0xF
1 Variable peripheral
Delay DLYBCS
Same peripheral
SPI_TDR(PCS)
New peripheral
NPCS = 0xF
Delay DLYBCS
NPCS = SPI_TDR(PCS)
365
1768BATARM08/03
Master Mode Block Diagram
SPIDIS SPIEN
Q
R
SPI_RDR
PCS RD
LSB MSB
MISO Serializer MOSI
SPI_TDR
PCS TD
NPCS3
NPCS2
NPCS1
SPI_MR(PS)
NPCS0
SPI_MR(PCS) 0
SPI_MR(MSTR)
SPI_SR S
M T R O P
O D D V I
D R R R E
F E F E N
S
SPI_IER
SPI_IDR
SPI_IMR
SPI Interrupt
366 AT91RM9200
1768BATARM08/03
AT91RM9200
SPCK
NSS
SPIDIS SPIEN
Q
R
SPI_RDR
RD
LSB MSB
MOSI Serializer MISO
SPI_TDR
TD
SPI_SR S
P T R O
I D D V
E R R R
N E F E
S
SPI_IER
SPI_IDR
SPI_IMR
SPI Interrupt
367
1768BATARM08/03
Data Transfer Four modes are used for data transfers. These modes correspond to combinations of a pair of
parameters called clock polarity (CPOL) and clock phase (NCPHA) that determine the edges
of the clock signal on which the data are driven and sampled. Each of the two parameters has
two possible states, resulting in four possible combinations that are incompatible with one
another. Thus a master/slave pair must use the same parameter pair values to communicate.
If multiple slaves are used and fixed in different configurations, the master must reconfigure
itself each time it needs to communicate with a different slave.
Table 69 shows the four modes and corresponding parameter settings.
Table 69. SPI Bus Protocol Mode
SPI Mode CPOL NCPHA
0 0 0
1 0 1
2 1 0
3 1 1
SPCK
(CPOL = 0)
(Mode 1)
SPCK
(CPOL = 1)
(Mode 3)
MOSI
(from master)
MSB 6 5 4 3 2 1 LSB
MISO
(from slave) MSB 6 5 4 3 2 1 LSB *
368 AT91RM9200
1768BATARM08/03
AT91RM9200
SPCK
(CPOL = 0)
(Mode 0)
SPCK
(CPOL = 1)
(Mode 2)
MOSI
(from master)
MSB 6 5 4 3 2 1 LSB
MISO
(from slave) * MSB 6 5 4 3 2 1 LSB
369
1768BATARM08/03
Serial Peripheral Interface (SPI) User Interface
Table 70. SPI Register Mapping
Offset Register Register Name Access Reset
0x00 Control Register SPI_CR Write-only ---
0x04 Mode Register SPI_MR Read/write 0x0
0x08 Receive Data Register SPI_RDR Read-only 0x0
0x0C Transmit Data Register SPI_TDR Write-only ---
0x10 Status Register SPI_SR Read-only 0x000000F0
0x14 Interrupt Enable Register SPI_IER Write-only ---
0x18 Interrupt Disable Register SPI_IDR Write-only ---
0x1C Interrupt Mask Register SPI_IMR Read-only 0x0
0x20 - 0x2C Reserved
0x30 Chip Select Register 0 SPI_CSR0 Read/write 0x0
0x34 Chip Select Register 1 SPI_CSR1 Read/write 0x0
0x38 Chip Select Register 2 SPI_CSR2 Read/write 0x0
0x3C Chip Select Register 3 SPI_CSR3 Read/write 0x0
0x40 - 0xFF Reserved
0x100 - 0x124 Reserved for the PDC
370 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
SWRST SPIDIS SPIEN
371
1768BATARM08/03
SPI Mode Register
Name: SPI_MR
Access Type: Read/write
31 30 29 28 27 26 25 24
DLYBCS
23 22 21 20 19 18 17 16
PCS
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
LLB MODFDIS DIV32 PCSDEC PS MSTR
372 AT91RM9200
1768BATARM08/03
AT91RM9200
373
1768BATARM08/03
SPI Receive Data Register
Name: SPI_RDR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
PCS
15 14 13 12 11 10 9 8
RD
7 6 5 4 3 2 1 0
RD
23 22 21 20 19 18 17 16
PCS
15 14 13 12 11 10 9 8
TD
7 6 5 4 3 2 1 0
TD
374 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
SPIENS
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
TXBUFE RXBUFF ENDTX ENDRX OVRES MODF TDRE RDRF
375
1768BATARM08/03
SPI Interrupt Enable Register
Name: SPI_IER
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
TXBUFE RXBUFF ENDTX ENDRX OVRES MODF TDRE RDRF
376 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
TXBUFE RXBUFF ENDTX ENDRX OVRES MODF TDRE RDRF
377
1768BATARM08/03
SPI Interrupt Mask Register
Name: SPI_IMR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
TXBUFE RXBUFF ENDTX ENDRX OVRES MODF TDRE RDRF
378 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
DLYBS
15 14 13 12 11 10 9 8
SCBR
7 6 5 4 3 2 1 0
BITS NCPHA CPOL
379
1768BATARM08/03
SCBR: Serial Clock Baud Rate
In Master Mode, the SPI Interface uses a modulus counter to derive the SPCK baud rate from the Master Clock MCK. The
Baud rate is selected by writing a value from 2 to 255 in the field SCBR. The following equation determines the SPCK baud
rate:
If DIV32 is 0:
SPCK Baudrate = MCK ( 2 SCBR )
If DIV32 is 1:
SPCK Baudrate = MCK ( 64 SCBR )
Giving SCBR a value of zero or one disables the baud rate generator. SPCK is disabled and assumes its inactive state
value. No serial transfers may occur. At reset, baud rate is disabled.
DLYBS: Delay Before SPCK
This field defines the delay from NPCS valid to the first valid SPCK transition.
When DLYBS equals zero, the NPCS valid to SPCK transition is 1/2 the SPCK clock period.
Otherwise, the following equations determine the delay:
If DIV32 is 0:
Delay Before SPCK = DLYBS MCK
If DIV32 is 1:
Delay Before SPCK = 32 DLYBS MCK
DLYBCT: Delay Between Consecutive Transfers
This field defines the delay between two consecutive transfers with the same peripheral without removing the chip select.
The delay is always inserted after each transfer and before removing the chip select if needed.
When DLYBCT equals zero, a minimum delay of four MCK cycles are inserted (or 128 MCK cycles when DIV32 is set)
between two consecutive characters.
Otherwise, the following equation determines the delay:
If DIV32 is 0:
Delay Between Consecutive Transfers = 32 DLYBCT MCK
If DIV32 is 1:
Delay Between Consecutive Transfers = 1024 DLYBCT MCK
380 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Two-wire Interface (TWI) interconnects components on a unique two-wire bus, made up
of one clock line and one data line with speeds of up to 400 Kbits per second, based on a byte-
oriented transfer format. It can be used with any Atmel two-wire bus serial EEPROM. The TWI
is programmable as a master with sequential or single-byte access.
A configurable baud rate generator permits the output data rate to be adapted to a wide range
of core clock frequencies.
The main features of the TWI are:
Compatibility with standard two-wire serial memory
One, two or three bytes for slave address
Sequential read/write operations
Block Diagram
Figure 161. Block Diagram
APB Bridge
TWCK
PIO
Two-wire TWD
Interface
PMC MCK
TWI
Interrupt
AIC
R R
TWD
Host with
TWI
Interface TWCK
381
1768BATARM08/03
Table 71. I/O Lines Description
Pin Name Pin Description Type
TWD Two-wire Serial Data Input/Output
TWCK Two-wire Serial Clock Input/Output
Product Dependencies
I/O Lines Both TWD and TWCK are bi-directional lines, connected to a positive supply voltage via a cur-
rent source or pull-up resistor (see Figure 162 on page 381). When the bus is free, both lines
are high. The output stages of devices connected to the bus must have an open-drain or open-
collector to perform the wired-AND function.
TWD and TWCK pins may be multiplexed with PIO lines. To enable the TWI, the programmer
must perform the following steps:
Program the PIO controller to:
Dedicate TWD and TWCK as peripheral lines.
Define TWD and TWCK as open-drain.
Interrupt The TWI interface has an interrupt line connected to the Advanced Interrupt Controller (AIC).
In order to handle interrupts, the AIC must be programmed before configuring the TWI.
Functional Description
Transfer Format The data put on the TWD line must be eight bits long. Data is transferred MSB first; each byte
must be followed by an acknowledgement. The number of bytes per transfer is unlimited (see
Figure 164 on page 383).
Each transfer begins with a START condition and terminates with a STOP condition (see Fig-
ure 163 on page 382).
A high-to-low transition on the TWD line while TWCK is high defines the START condition.
A low-to-high transition on the TWD line while TWCK is high defines a STOP condition.
TWD
TWCK
Start Stop
382 AT91RM9200
1768BATARM08/03
AT91RM9200
TWD
TWCK
Transmitting Data After the master initiates a Start condition, it sends a 7-bit slave address, configured in the
Master Mode register (DADR in TWI_MMR), to notify the slave device. The bit following the
slave address indicates the transfer direction (write or read). If this bit is 0, it indicates a write
operation (transmit operation). If the bit is 1, it indicates a request for data read (receive
operation).
The TWI transfers require the slave to acknowledge each received byte. During the acknowl-
edge clock pulse, the master releases the data line (HIGH), enabling the slave to pull it down
in order to generate the acknowledge. The master polls the data line during this clock pulse
and sets the NAK bit in the status register if the slave does not acknowledge the byte. As with
the other status bits, an interrupt can be generated if enabled in the interrupt enable register
(TWI_IER). After writing in the transmit-holding register (TWI_THR), setting the START bit in
the control register starts the transmission. The data is shifted in the internal shifter and when
an acknowledge is detected, the TXRDY bit is set until a new write in the TWI_THR (see Fig-
ure 166 on page 384). The master generates a stop condition to end the transfer.
The read sequence begins by setting the START bit. When the RXRDY bit is set in the status
register, a character has been received in the receive-holding register (TWI_RHR). The
RXRDY bit is reset when reading the TWI_RHR.
The TWI interface performs various transfer formats (7-bit slave address, 10-bit slave
address). The three internal address bytes are configurable through the Master Mode register
(TWI_MMR). If the slave device supports only a 7-bit address, the IADRSZ must be set to 0.
For slave address higher than seven bits, the user must configure the address size (IADRSZ)
and set the other slave address bits in the internal address register (TWI_IADR).
Figure 165. Master Write with One, Two or Three Bytes Internal Address and One Data Byte
Three bytes internal address
TWD S DADR W A IADR(23:16) A IADR(15:8) A IADR(7:0) A DATA A P
383
1768BATARM08/03
Figure 166. Master Write with One Byte Internal Address and Multiple Data Bytes
TXCOMP
Write THR
TXRDY
Figure 167. Master Read with One, Two or Three Bytes Internal Address and One Data Byte
Three bytes internal address
TWD S DADR W A IADR(23:16) A IADR(15:8) A IADR(7:0) A S DADR R A
DATA N P
Figure 168. Master Read with One Byte Internal Address and Multiple Data Bytes
TXCOMP
RXRDY
S = Start
P = Stop
W = Write/read
A = Acknowledge
DADR= Device Address
IADR = Internal Address
Figure 169 shows a byte write to an Atmel AT24LC512 EEPROM. This demonstrates the use
of internal addresses to access the device.
384 AT91RM9200
1768BATARM08/03
AT91RM9200
M LR A M A LA A
S S / C S C SC C
B BW K B K BK K
Read/Write The following flowcharts shown in Figure 170 on page 386 and in Figure 171 on page 387 give
Flowcharts examples for read and write operations in Master Mode. A polling or interrupt method can be
used to check the status bits. The interrupt method requires that the interrupt enable register
(TWI_IER) be configured first.
385
1768BATARM08/03
Figure 170. TWI Write in Master Mode
START
Data to send?
Yes
TXCOMP = 0?
END
386 AT91RM9200
1768BATARM08/03
AT91RM9200
START
RXRDY = 0?
Yes
Data to read?
Yes
Yes
TXCOMP = 0?
END
387
1768BATARM08/03
Two-wire Interface (TWI) User Interface
Table 72. TWI Register Mapping
Offset Register Name Access Reset Value
0x0000 Control Register TWI_CR Write-only N/A
0x0004 Master Mode Register TWI_MMR Read/write 0x0000
0x0008 Reserved
0x000C Internal Address Register TWI_IADR Read/write 0x0000
0x0010 Clock Waveform Generator Register TWI_CWGR Read/write 0x0000
0x0020 Status Register TWI_SR Read-only 0x0008
0x0024 Interrupt Enable Register TWI_IER Write-only N/A
0x0028 Interrupt Disable Register TWI_IDR Write-only N/A
0x002C Interrupt Mask Register TWI_IMR Read-only 0x0000
0x0030 Receive Holding Register TWI_RHR Read-only 0x0000
0x0034 Transmit Holding Register TWI_THR Read/write 0x0000
388 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
SWRST MSDIS MSEN STOP START
389
1768BATARM08/03
TWI Master Mode Register
Register Name: TWI_MMR
Address Type: Read/write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
DADR
15 14 13 12 11 10 9 8
MREAD IADRSZ
7 6 5 4 3 2 1 0
IADRSZ[9:8]
0 0 No internal device address
0 1 One-byte internal device address
1 0 Two-byte internal device address
1 1 Three-byte internal device address
390 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
IADR
15 14 13 12 11 10 9 8
IADR
7 6 5 4 3 2 1 0
IADR
23 22 21 20 19 18 17 16
CKDIV
15 14 13 12 11 10 9 8
CHDIV
7 6 5 4 3 2 1 0
CLDIV
CKDIV
T low = ( ( CLDIV 2 ) + 3 ) T MCK
CKDIV
T high = ( ( CHDIV 2 ) + 3 ) T MCK
391
1768BATARM08/03
TWI Status Register
Register Name: TWI_SR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
NACK
7 6 5 4 3 2 1 0
UNRE OVRE TXRDY RXRDY TXCOMP
392 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
NACK
7 6 5 4 3 2 1 0
UNRE OVRE TXRDY RXRDY TXCOMP
393
1768BATARM08/03
TWI Interrupt Disable Register
Register Name: TWI_IDR
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
NACK
7 6 5 4 3 2 1 0
UNRE OVRE TXRDY RXRDY TXCOMP
394 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
NACK
7 6 5 4 3 2 1 0
UNRE OVRE TXRDY RXRDY TXCOMP
395
1768BATARM08/03
TWI Receive Holding Register
Register Name: TWI_RHR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
RXDATA
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
TXDATA
396 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Universal Synchronous Asynchronous Receiver Transceiver (USART) provides one full
duplex universal synchronous asynchronous serial link. Data frame format is widely program-
mable (data length, parity, number of stop bits) to support a maximum of standards. The
receiver implements parity error, framing error and overrun error detection. The receiver time-
out enables handling variable-length frames and the transmitter timeguard facilitates commu-
nications with slow remote devices. Multi-drop communications are also supported through
address bit handling in reception and transmission.
The USART features three test modes: remote loopback, local loopback and automatic echo.
The USART supports specific operating modes providing interfaces on RS485 busses, with
ISO7816 T = 0 or T = 1 smart card slots, infrared transceivers and connection to modem ports.
The hardware handshaking feature enables an out-of-band flow control by automatic manage-
ment of the pins RTS and CTS.
The USART supports the connection to the Peripheral Data Controller, which enables data
transfers to the transmitter and from the receiver. The PDC provides chained buffer manage-
ment without any intervention of the processor.
Important features of the USART are:
Programmable Baud Rate Generator
5- to 9-bit Full-duplex Synchronous or Asynchronous Serial Communications
1, 1.5 or 2 Stop Bits in Asynchronous Mode or 1 or 2 Stop Bits in Synchronous
Mode
Parity Generation and Error Detection
Framing Error Detection, Overrun Error Detection
MSB- or LSB-first
Optional Break Generation and Detection
By 8 or by-16 Over-sampling Receiver Frequency
Optional Hardware Handshaking RTS-CTS
Optional Modem Signal Management DTR-DSR-DCD-RI
Receiver Time-out and Transmitter Timeguard
Optional Multi-Drop Mode with Address Generation and Detection
RS485 with driver control signal
ISO7816, T = 0 or T = 1 Protocols for Interfacing with Smart Cards
NACK Handling, Error Counter with Repetition and Iteration Limit
IrDA Modulation and Demodulation
Communication at up to 115.2 Kbps
Test Modes
Remote Loopback, Local Loopback, Automatic Echo
Supports Connection of Two Peripheral Data Controller Channels (PDC)
Offer Buffer Transfer without Processor Intervention
397
1768BATARM08/03
Block Diagram Figure 172. USART Block Diagram
Peripheral Data
Controller
Channel Channel
PIO
USART Controller
RXD
Receiver
RTS
DTR
PMC Modem
MCK DSR
Signals
Control
MCK/DIV DCD
DIV
RI
User Interface
APB
398 AT91RM9200
1768BATARM08/03
AT91RM9200
USART
I/O Lines
Description
Table 73. I/O Line Description
Name Description Type Active Level
SCK Serial Clock I/O
TXD Transmit Serial Data I/O
RXD Receive Serial Data Input
RI Ring Indicator Input Low
DSR Data Set Ready Input Low
DCD Data Carrier Detect Input Low
DTR Data Terminal Ready Output Low
CTS Clear to Send Input Low
RTS Request to Send Output Low
Product
Dependencies
I/O Lines The pins used for interfacing the USART may be multiplexed with the PIO lines. The program-
mer must first program the PIO controller to assign the desired USART pins to their peripheral
function. If I/O lines of the USART are not used by the application, they can be used for other
purposes by the PIO Controller.
All the pins of the modems may or may not not be implemented on the USART within a prod-
uct. Frequently, only the USART1 is fully equipped with all the modem signals. For the other
399
1768BATARM08/03
USARTs of the product not equipped with the corresponding pin, the associated control bits
and statuses have no effect on the behavior of the USART.
Power The USART is not continuously clocked. The programmer must first enable the USART Clock
Management in the Power Management Controller (PMC) before using the USART. However, if the applica-
tion does not require USART operations, the USART clock can be stopped when not needed
and be restarted later. In this case, the USART will resume its operations where it left off.
Configuring the USART does not require the USART clock to be enabled.
Interrupt The USART interrupt line is connected on one of the internal sources of the Advanced Inter-
rupt Controller. Using the USART interrupt requires the AIC to be programmed first. Note that
it is not recommended to use the USART interrupt line in edge sensitive mode.
Functional The USART is capable of managing several types of serial synchronous or asynchronous
communications.
Description
It supports the following communication modes.
5- to 9-bit full-duplex asynchronous serial communication:
MSB- or LSB-first
1, 1.5 or 2 stop bits
Parity even, odd, marked, space or none
By-8 or by-16 over-sampling receiver frequency
Optional hardware handshaking
Optional modem signals management
Optional break management
Optional multi-drop serial communication
High-speed 5- to 9-bit full-duplex synchronous serial communication:
MSB- or LSB-first
1 or 2 stop bits
Parity even, odd, marked, space or none
by 8 or by-16 over-sampling frequency
Optional Hardware handshaking
Optional Modem signals management
Optional Break management
Optional Multi-Drop serial communication
RS485 with driver control signal
ISO7816, T0 or T1 protocols for interfacing with smart cards
NACK handling, error counter with repetition and iteration limit
InfraRed IrDA Modulation and Demodulation
Test modes
remote loopback, local loopback, automatic echo
Baud Rate The Baud Rate Generator provides the bit period clock named the Baud Rate Clock to both
Generator the receiver and the transmitter.
The Baud Rate Generator clock source can be selected by setting the USCLKS field in the
Mode Register (US_MR) between:
400 AT91RM9200
1768BATARM08/03
AT91RM9200
MCK CD
0 SCK
MCK/DIV
1
Reserved 16-bit Counter
SCK 2 FIDI
>1 SYNC
3 OVER
1 0
0 0 Sampling 0
Divider
Baud Rate
1 Clock
1
SYNC
Sampling
USCLKS = 3 Clock
Baud Rate in If the USART is programmed to operate in asynchronous mode, the selected clock is first
Asynchronous Mode divided by CD, which is field programmed in the Baud Rate Generator Register (US_BRGR).
The resulting clock is provided to the receiver as a sampling clock and then divided by 16 or 8,
depending on the programming of the OVER bit in US_MR.
If OVER is set to 1, the receiver sampling is 8 times higher than the baud rate clock. If OVER
is cleared, the sampling is performed at 16 times the baud rate clock.
The following formula performs the calculation of the Baud Rate.
SelectedClock
Baudrate = --------------------------------------------
( 8 ( 2 Over )CD )
This gives a maximum baud rate of MCK divided by 8, assuming that MCK is the highest pos-
sible clock and that OVER is programmed at 1.
401
1768BATARM08/03
Baud Rate Calculation Table 74 shows calculations of CD to obtain a baud rate at 38400 bauds for different source
Example clock frequencies. This table also shows the actual resulting baud rate and the error.
Error = 1 ---------------------------------------------------
ExpectedBaudRate
ActualBaudRate
Baud Rate in If the USART is programmed to operate in synchronous mode, the selected clock is simply
Synchronous Mode divided by the field CD in US_BRGR.
SelectedClock
BaudRate = --------------------------------------
CD
402 AT91RM9200
1768BATARM08/03
AT91RM9200
In synchronous mode, if the external clock is selected (USCLKS = 3), the clock is provided
directly by the signal on the USART SCK pin. No division is active. The value written in
US_BRGR has no effect. The external clock frequency must be at least 4.5 times lower than
the system clock.
When either the external clock SCK or the internal clock divided (MCK/DIV) is selected, the
value programmed in CD must be even if the user has to ensure a 50:50 mark/space ratio on
the SCK pin. If the internal clock MCK is selected, the Baud Rate Generator ensures a 50:50
duty cycle on the SCK pin, even if the value programmed in CD is odd.
Baud Rate in ISO 7816 The ISO7816 specification defines the bit rate with the following formula:
Mode Di
B = ------ f
Fi
where:
B is the bit rate
Di is the bit-rate adjustment factor
Fi is the clock frequency division factor
f is the ISO7816 clock frequency (Hz)
Di is a binary value encoded on a 4-bit field, named DI, as represented in Table 75.
Fi is a binary value encoded on a 4-bit field, named FI, as represented in Table 76.
Table 77 shows the resulting Fi/Di Ratio, which is the ratio between the ISO7816 clock and the
baud rate clock..
If the USART is configured in ISO7816 Mode, the clock selected by the USCLKS field in the
Mode Register (US_MR) is first divided by the value programmed in the field CD in the Baud
403
1768BATARM08/03
Rate Generator Register (US_BRGR). The resulting clock can be provided to the SCK pin to
feed the smart card clock inputs. This means that the CLKO bit can be set in US_MR.
This clock is then divided by the value programmed in the FI_DI_RATIO field in the
FI_DI_Ratio register (US_FIDI). This is performed by the Sampling Divider, which performs a
division by up to 2047 in ISO7816 Mode. The non-integer values of the Fi/Di Ratio are not sup-
ported and the user must program the FI_DI_RATIO field to a value as close as possible to the
expected value.
The FI_DI_RATIO field resets to the value 0x174 (372 in decimal) and is the most common
divider between the ISO7816 clock and the bit rate (Fi = 372, Di = 1).
Figure 175 shows the relation between the Elementary Time Unit, corresponding to a bit time,
and the ISO 7816 clock.
ISO7816 Clock
on SCK
1 ETU
Receiver and After reset, the receiver is disabled. The user must enable the receiver by setting the RXEN bit
Transmitter in the Control Register (US_CR). However, the receiver registers can be programmed before
Control the receiver clock is enabled.
After reset, the transmitter is disabled. The user must enable it by setting the TXEN bit in the
Control Register (US_CR). However, the transmitter registers can be programmed before
being enabled.
The Receiver and the Transmitter can be enabled together or independently.
At any time, the software can perform a reset on the receiver or the transmitter of the USART
by setting the corresponding bit, RSTRX and RSTTX respectively, in the Control Register
(US_CR). The reset commands have the same effect as a hardware reset on the correspond-
ing logic. Regardless of what the receiver or the transmitter is performing, the communication
is immediately stopped.
The user can also independently disable the receiver or the transmitter by setting RXDIS and
TXDIS respectively in US_CR. If the receiver is disabled during a character reception, the
USART waits until the end of reception of the current character, then the reception is stopped.
If the transmitter is disabled while it is operating, the USART waits the end of transmission of
both the current character and character being stored in the Transmit Holding Register
(US_THR). If a time guard is programmed, it is handled normally.
404 AT91RM9200
1768BATARM08/03
AT91RM9200
Synchronous and
Asynchronous
Modes
Transmitter The transmitter performs the same in both synchronous and asynchronous operating modes
Operations (SYNC = 0 or SYNC = 1). One start bit, up to 9 data bits, one optional parity bit and up to two
stop bits are successively shifted out on the TXD pin at each falling edge of the programmed
serial clock.
The number of data bits is selected by the CHRL field and the MODE9 bit in the Mode Regis-
ter (US_MR). Nine bits are selected by setting the MODE 9 bit regardless of the CHRL field.
The parity bit is set according to the PAR field in US_MR. The even, odd, space, marked or
none parity bit can be configured. The MSBF field in US_MR configures which data bit is sent
first. If written at 1, the most significant bit is sent first. At 0, the less significant bit is sent first.
The number of stop bits is selected by the NBSTOP field in US_MR. The 1.5 stop bit is sup-
ported in asynchronous mode only.
Baud Rate
Clock
TXD
The characters are sent by writing in the Transmit Holding Register (US_THR). The transmit-
ter reports two status bits in the Channel Status Register (US_CSR): TXRDY (Transmitter
Ready), which indicates that US_THR is empty and TXEMPTY, which indicates that all the
characters written in US_THR have been processed. When the current character processing
is completed, the last character written in US_THR is transferred into the Shift Register of the
transmitter and US_THR becomes empty, thus TXRDY raises.
Both TXRDY and TXEMPTY bits are low since the transmitter is disabled. Writing a character
in US_THR while TXRDY is active has no effect and the written character is lost.
TXD
Start Parity Stop Start Parity Stop
D0 D1 D2 D3 D4 D5 D6 D7 D0 D1 D2 D3 D4 D5 D6 D7
Bit Bit Bit Bit Bit Bit
Write
US_THR
TXRDY
TXEMPTY
405
1768BATARM08/03
Asynchronous If the USART is programmed in asynchronous operating mode (SYNC = 0), the receiver over-
Receiver samples the RXD input line. The oversampling is either 16 or 8 times the Baud Rate clock,
depending on the OVER bit in the Mode Register (US_MR).
The receiver samples the RXD line. If the line is sampled during one half of a bit time at 0, a
start bit is detected and data, parity and stop bits are successively sampled on the bit rate
clock.
If the oversampling is 16, (OVER at 0), a start is detected at the eighth sample at 0. Then, data
bits, parity bit and stop bit are sampled on each 16 sampling clock cycle. If the oversampling is
8 (OVER at 1), a start bit is detected at the fourth sample at 0. Then, data bits, parity bit and
stop bit are sampled on each 8 sampling clock cycle.
The number of data bits, first bit sent and parity mode are selected by the same fields and bits
as the transmitter, i.e. respectively CHRL, MODE9, MSBF and PAR. The number of stop bits
has no effect on the receiver as it considers only one stop bit, regardless of the field NBSTOP,
so that resynchronization between the receiver and the transmitter can occur. Moreover, as
soon as the stop bit is sampled, the receiver starts looking for a new start bit so that resynchro-
nization can also be accomplished when the transmitter is operating with one stop bit.
Figure 178 and Figure 179 illustrate start detection and character reception when USART
operates in asynchronous mode.
Sampling
Clock (x16)
RXD
Sampling
1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
D0
Start Sampling
Detection
RXD
Sampling
1 2 3 4 5 6 7 0 1 2 3 4
Start
Rejection
Baud Rate
Clock
RXD
Start 16 16 16 16 16 16 16 16 16 16
Detection samples samples samples samples samples samples samples samples samples samples
D0 D1 D2 D3 D4 D5 D6 D7 Parity Stop
Bit Bit
406 AT91RM9200
1768BATARM08/03
AT91RM9200
Synchronous Receiver In synchronous mode (SYNC = 1), the receiver samples the RXD signal on each rising edge of
the Baud Rate Clock. If a low level is detected, it is considered as a start. All data bits, the par-
ity bit and the stop bits are sampled and the receiver waits for the next start bit. Synchronous
mode operations provide a high speed transfer capability.
Configuration fields and bits are the same as in asynchronous mode.
Figure 180 illustrates a character reception in synchronous mode.
Baud Rate
Clock
RXD
Sampling
Start D0 D1 D2 D3 D4 D5 D6 D7 Stop Bit
Parity Bit
Receiver Operations When a character reception is completed, it is transferred to the Receive Holding Register
(US_RHR) and the RXRDY bit in the Status Register (US_CSR) rises. If a character is com-
pleted while the RXRDY is set, the OVRE (Overrun Error) bit is set. The last character is
transferred into US_RHR and overwrites the previous one. The OVRE bit is cleared by writing
the Control Register (US_CR) with the RSTSTA (Reset Status) bit at 1.
Baud Rate
Clock
RXD
Start Parity Stop Start Parity Stop
D0 D1 D2 D3 D4 D5 D6 D7 D0 D1 D2 D3 D4 D5 D6 D7
Bit Bit Bit Bit Bit Bit
RSTSTA = 1
Write
US_CR
Read
US_RHR
RXRDY
OVRE
Parity The USART supports five parity modes selected by programming the PAR field in the Mode
Register (US_MR). The PAR field also enables the Multidrop mode, which is discussed in a
separate paragraph. Even and odd parity bit generation and error detection are supported.
If even parity is selected, the parity generator of the transmitter drives the parity bit at 1 if a
number of 1s in the character data bit is even, and at 0 if the number of 1s is odd. Accordingly,
the receiver parity checker counts the number of received 1s and reports a parity error if the
sampled parity bit does not correspond. If the odd parity is selected, the parity generator of the
407
1768BATARM08/03
transmitter drives the parity bit at 0 if a number of 1s in the character data bit is even, and at 1
if the number of 1s is odd. Accordingly, the receiver parity checker counts the number of
received 1s and reports a parity error if the sampled parity bit does not correspond. If the mark
parity is used, the parity generator of the transmitter drives the parity bit at 1 for all characters.
The receiver parity checker reports an error if the parity bit is sampled at 0.If the space parity is
used, the parity generator of the transmitter drives the parity bit at 0 for all characters. The
receiver parity checker reports an error if the parity bit is sampled at 1. If parity is disabled, the
transmitter does not generate any parity bit and the receiver does not report any parity error.
Table 78 shows an example of the parity bit for the character 0x41 (character ASCII A)
depending on the configuration of the USART. Because there are two bits at 1, 1 bit is added
when a parity is odd, or 0 is added when a parity is even. I
When the receiver detects a parity error, it sets the PARE (Parity Error) bit in the Channel Sta-
tus Register (US_CSR). The PARE bit can be cleared by writing the Control Register (US_CR)
with the RSTSTA bit at 1. Figure 182 illustrates the parity bit status setting and clearing.
Baud Rate
Clock
RXD
Start Bad Stop
D0 D1 D2 D3 D4 D5 D6 D7
Bit Parity Bit
Bit RSTSTA = 1
Write
US_CR
PARE
RXRDY
Multi-drop Mode If the PAR field in the Mode Register (US_MR) is programmed to the value 0x3, the USART
runs in Multi-drop mode. This mode differentiates the data characters and the address charac-
ters. Data is transmitted with the parity bit at 0 and addresses are transmitted with the parity bit
at 1.
If the USART is configured in multi-drop mode, the receiver sets the PARE parity error bit
when the parity bit is high and the transmitter is able to send a character with the parity bit high
when the Control Register is written with the SENDA bit at 1.
408 AT91RM9200
1768BATARM08/03
AT91RM9200
To handle parity error, the PARE bit is cleared when the Control Register is written with the bit
RSTSTA at 1.
The transmitter sends an address byte (parity bit set) when SENDA is written to US_CR. In
this case, the next byte written to US_THR is transmitted as an address. Any character written
in US_THR without having written the command SENDA is transmitted normally with the parity
at 0.
Transmitter Timeguard The timeguard feature enables the USART interface with slow remote devices.
The timeguard function enables the transmitter to insert an idle state on the TXD line between
two characters. This idle state actually acts as a long stop bit.
The duration of the idle state is programmed in the TG field of the Transmitter Timeguard Reg-
ister (US_TTGR). When this field is programmed at zero no timeguard is generated.
Otherwise, the transmitter holds a high level on TXD after each transmitted byte during the
number of bit periods programmed in TG in addition to the number of stop bits.
As illustrated in Figure 183, the behavior of TXRDY and TXEMPTY status bits is modified by
the programming of a timeguard. TXRDY rises only when the start bit of the next character is
sent, and thus remains at 0 during the timeguard transmission if a character has been written
in US_THR. TXEMPTY remains low until the timeguard transmission is completed as the time-
guard is part of the current character being transmitted.
TXD
Start Parity Stop Start Parity Stop
D0 D1 D2 D3 D4 D5 D6 D7 D0 D1 D2 D3 D4 D5 D6 D7
Bit Bit Bit Bit Bit Bit
Write
US_THR
TXRDY
TXEMPTY
Table 79 indicates the maximum length of a timeguard period that the transmitter can handle
in relation to the function of the Baud Rate.
409
1768BATARM08/03
Table 79. Maximum Timeguard Length Depending on Baud Rate (Continued)
Baud Rate Bit time Timeguard
56000 17.9 4.55
57600 17.4 4.43
115200 8.7 2.21
Receiver Time-out The Receiver Time-out provides support in handling variable-length frames. This feature
detects an idle condition on the RXD line. When a time-out is detected, the bit TIMEOUT in the
Channel Status Register (US_CSR) rises and can generate an interrupt, thus indicating to the
driver an end of frame.
The time-out delay period (during which the receiver waits for a new character) is programmed
in the TO field of the Receiver Time-out Register (US_RTOR). If the TO field is programmed at
0, the Receiver Time-out is disabled and no time-out is detected. The TIMEOUT bit in
US_CSR remains at 0. Otherwise, the receiver loads a 16-bit counter with the value pro-
grammed in TO. This counter is decremented at each bit period and reloaded each time a new
character is received. If the counter reaches 0, the TIMEOUT bit in the Status Register rises.
The user can either:
Obtain an interrupt when a time-out is detected after having received at least one
character. This is performed by writing the Control Register (US_CR) with the STTTO
(Start Time-out) bit at 1.
Obtain a periodic interrupt while no character is received. This is performed by writing
US_CR with the RETTO (Reload and Start Time-out) bit at 1.
If STTTO is performed, the counter clock is stopped until a first character is received. The idle
state on RXD before the start of the frame does not provide a time out. This prevents having to
obtain a periodic interrupt and enables a wait of the end of frame when the idle state on RXD
is detected.
If RETTO is performed, the counter starts counting down immediately from the value TO. This
enables generation of a periodic interrupt so that a user time-out can be handled, for example
when no key is pressed on a keyboard.
Figure 184 shows the block diagram of the Receiver Time out feature.
Baud Rate TO
Clock
16-bit
Value
1 D Q Clock 16-bit Time-out
Counter
STTTO = TIMEOUT
Load 0
Clear
Character
Received
RETTO
410 AT91RM9200
1768BATARM08/03
AT91RM9200
Table 80 gives the maximum time-out period for some standard baud rates.t
Framing Error The receiver is capable of detecting framing errors. A framing error happens when the stop bit
of a received character is detected at level 0. This can occur if the receiver and the transmitter
are fully desynchronized.
A framing error is reported on the FRAME bit of the Channel Status Register (US_CSR). The
FRAME bit is asserted in the middle of the stop bit as soon as the framing error is detected. It
is cleared by writing the Control Register (US_CR) with the RSTSTA bit at 1.
Baud Rate
Clock
RXD
Start Parity Stop
D0 D1 D2 D3 D4 D5 D6 D7
Bit Bit Bit
RSTSTA = 1
Write
US_CR
FRAME
RXRDY
Transmit Break The user can request the transmitter to generate a break condition on the TXD line. A break
condition drives the TXD line low during at least one complete character. It appears the same
as a 0x00 character sent with the parity and the stop bits at 0. However, the transmitter holds
the TXD line at least during one character until the user requests the break condition to be
removed.
411
1768BATARM08/03
A break is transmitted by writing the Control Register (US_CR) with the STTBRK bit at 1. This
can be performed at any time, either while the transmitter is empty (no character in either the
Shift Register or in US_THR) or when a character is being transmitted. If a break is requested
while a character is being shifted out, the character is first completed before the TXD line is
held low.
Once STTBRK command is requested further STTBRK commands are ignored until the end of
the break is completed.
The break condition is removed by writing US_CR with the STPBRK bit at 1. If the STPBRK is
requested before the end of the minimum break duration (one character, including start, data,
parity and stop bits), the transmitter ensures that the break condition completes.
The transmitter considers the break as though it is a character, i.e. the STTBRK and STPBRK
commands are taken into account only if the TXRDY bit in US_CSR is at 1 and the start of the
break condition clears the TXRDY and TXEMPTY bits as if a character is processed.
Writing US_CR with the both STTBRK and STPBRK bits at 1 can lead to an unpredictable
result. All STPBRK commands requested without a previous STTBRK command are ignored.
A byte written into the Transmit Holding Register while a break is pending, but not started, is
ignored.
After the break condition, the transmitter returns the TXD line to 1 for a minimum of 12 bit
times. Thus, the transmitter ensures that the remote receiver detects correctly the end of
break and the start of the next character. If the timeguard is programmed with a value higher
than 12, the TXD line is held high for the timeguard period.
After holding the TXD line for this period, the transmitter resumes normal operations.
Figure 186 illustrates the effect of both the Start Break (STTBRK) and Stop Break (STP BRK)
commands on the TXD line.
Baud Rate
Clock
TXD
Start Parity Stop
Bit
D0 D1 D2 D3 D4 D5 D6 D7
Bit Bit
Break Transmission End of Break
STTBRK = 1 STPBRK = 1
Write
US_CR
TXRDY
TXEMPTY
Receive Break The receiver detects a break condition when all data, parity and stop bits are low. This corre-
sponds to detecting a framing error with data at 0x00, but FRAME remains low.
When the low stop bit is detected, the receiver asserts the RXBRK bit in US_CSR. This bit
may be cleared by writing the Control Register (US_CR) with the bit RSTSTA at 1.
An end of receive break is detected by a high level for at least 2/16 of a bit period in asynchro-
nous operating mode or one sample at high level in synchronous operating mode. The end of
break detection also asserts the RXBRK bit.
412 AT91RM9200
1768BATARM08/03
AT91RM9200
Hardware The USART features a hardware handshaking out-of-band flow control. The RTS and CTS
Handshaking pins are used to connect with the remote device, as shown in Figure 187.
USART Remote
Device
TXD RXD
RXD TXD
CTS RTS
RTS CTS
Setting the USART to operate with hardware handshaking is performed by writing the
USART_MODE field in the Mode Register (US_MR) to the value 0x2.
The USART behavior when hardware handshaking is enabled is the same as the behavior in
standard synchronous or asynchronous mode, except that the receiver drives the RTS pin as
described below and the level on the CTS pin modifies the behavior of the transmitter as
described below. Using this mode requires using the PDC channel for reception. The transmit-
ter can handle hardware handshaking in any case.
Figure 188 shows how the receiver operates if hardware handshaking is enabled. The RTS
pin is driven high if the receiver is disabled and if the status RXBUFF (Receive Buffer Full)
coming from the PDC channel is high. Normally, the remote device does not start transmitting
while its CTS pin (driven by RTS) is high. As soon as the Receiver is enabled, the RTS falls,
indicating to the remote device that it can start transmitting. Defining a new buffer to the PDC
clears the status bit RXBUFF and, as a result, asserts the pin RTS low.
RXD
RXEN = 1 RXDIS = 1
Write
US_CR
RTS
RXBUFF
Figure 189 shows how the transmitter operates if hardware handshaking is enabled. The CTS
pin disables the transmitter. If a character is being processing, the transmitter is disabled only
after the completion of the current character and transmission of the next character happens
as soon as the pin CTS falls.
CTS
TXD
413
1768BATARM08/03
ISO7816 Mode The USART features an ISO7816-compatible operating mode. This mode permits interfacing
with smart cards and Security Access Modules (SAM) communicating through an ISO7816
link. Both T = 0 and T = 1 protocols defined by the ISO7816 specification are supported.
Setting the USART in ISO7816 mode is performed by writing the USART_MODE field in the
Mode Register (US_MR) to the value 0x4 for protocol T = 0 and to the value 0x5 for protocol T
= 1.
ISO7816 Mode The ISO7816 is a half duplex communication on only one bidirectional line. The baud rate is
overview determined by a division of the clock provided to the remote device (see Baud Rate Genera-
tor on page 400).
The USART connects to a smart card. as shown in Figure 190. The TXD line becomes bidirec-
tional and the Baud Rate Generator feeds the ISO7816 clock on the SCK pin. As the TXD pin
becomes bidirectional, its output remains driven by the output of the transmitter but only when
the transmitter is active while its input is directed to the input of the receiver. The USART is
considered as the master of the communication as it generates the clock.
USART
CLK
SCK Smart
Card
I/O
TXD
When operating in ISO7816, either in T = 0 or T = 1 modes, the character format is fixed. The
configuration is 8 data bits, even parity and 1 or 2 stop bits, regardless of the values pro-
grammed in the CHRL, MODE9, PAR and CHMODE fields. MSBF can be used to transmit
LSB or MSB first.
The USART cannot operate concurrently in both receiver and transmitter modes as the com-
munication is unidirectional at a time. It has to be configured according to the required mode
by enabling or disabling either the receiver or the transmitter as desired. Enabling both the
receiver and the transmitter at the same time in ISO7816 mode may lead to unpredictable
results.
The ISO7816 specification defines an inverse transmission format. Data bits of the character
must be transmitted on the I/O line at their negative value. The USART does not support this
format and the user has to perform an exclusive OR on the data before writing it in the Trans-
mit Holding Register (US_THR) or after reading it in the Receive Holding Register (US_RHR).
Protocol T = 0 In T = 0 protocol, a character is made up of one start bit, eight data bits, one parity bit and one
guard time, which lasts two bit times. The transmitter shifts out the bits and does not drive the
I/O line during the guard time.
If no parity error is detected, the I/O line remains at 1 during the guard time and the transmitter
can continue with the transmission of the next character, as shown in Figure 191.
If a parity error is detected by the receiver, it drives the I/O line at 0 during the guard time, as
shown in Figure 192. This error bit is also named NACK, for Non Acknowledge. In this case,
the character lasts 1 bit time more, as the guard time length is the same and is added to the
error bit time which lasts 1 bit time.
414 AT91RM9200
1768BATARM08/03
AT91RM9200
When the USART is the receiver and it detects an error, it does not load the erroneous charac-
ter in the Receive Holding Register (US_RHR). It appropriately sets the PARE bit in the Status
Register (US_SR) so that the software can handle the error.
RXD
I/O Error
Receive Error Counter The USART receiver also records the total number of errors. This can be read in the Number
of Error (US_NER) register. The NB_ERRORS field can record up to 255 errors. Reading
US_NER automatically clears the NB_ERRORS field.
Receive NACK Inhibit The USART can also be configured to inhibit an error. This can be achieved by setting the
INACK bit in the Mode Register (US_MR). If INACK is at 1, no error signal is driven on the I/O
line even if a parity bit is detected, but the INACK bit is set in the Status Register (US_SR).
The INACK bit can be cleared by writing the Control Register (US_CR) with the RSTNACK bit
at 1.
Moreover, if INACK is set, the erroneous received character is stored in the Receive Holding
Register, as if no error occurred. However, the RXRDY bit does not raise.
Transmit Character When the USART is transmitting a character and gets a NACK, it can automatically repeat the
Repetition character before moving on to the next one. Repetition is enabled by writing the
MAX_ITERATION field in the Mode Register (US_MR) at a value higher than 0. Each charac-
ter can be transmitted up to eight times; the first transmission plus seven repetitions.
If MAX_ITERATION does not equal zero, the USART repeats the character as many times as
the value loaded in MAX_ITERATION.
When the USART repetition number reaches MAX_ITERATION, the ITERATION bit is set in
the Channel Status Register (US_CSR). If the repetition of the character is acknowledged by
the receiver, the repetitions are stopped and the iteration counter is cleared.
The ITERATION bit in US_CSR can be cleared by writing the Control Register with the RSIT
bit at 1.
Disable Successive The receiver can limit the number of successive NACKs sent back to the remote transmitter.
Receive NACK This is programmed by setting the bit DSNACK in the Mode Register (US_MR). The maximum
number of NACK transmitted is programmed in the MAX_ITERATION field. As soon as
415
1768BATARM08/03
MAX_ITERATION is reached, the character is considered as correct, an acknowledge is sent
on the line and the ITERATION bit in the Channel Status Register is set.
Protocol T = 1 When operating in ISO7816 protocol T = 1, the transmission is similar to an asynchronous for-
mat with only one stop bit. The parity is generated when transmitting and checked when
receiving. Parity error detection sets the PARE bit in the Channel Status Register (US_CSR).
IrDA Mode The USART features an IrDA mode supplying half-duplex point-to-point wireless communica-
tion. It embeds the modulator and demodulator which allows a glueless connection to the
infrared transceivers, as shown in Figure 193. The modulator and demodulator are compliant
with the IrDA specification version 1.1 and support data transfer speeds ranging from 2,4 Kbps
to 115,2 Kbps.
The USART IrDA mode is enabled by setting the USART_MODE field in the Mode Register
(US_MR) to the value 0x8. The IrDA Filter Register (US_IF) allows configuring the demodula-
tor filter. The USART transmitter and receiver operate in a normal asynchronous mode and all
parameters are accessible. Note that the modulator and the demodulator are activated.
USART IrDA
Transceivers
Receiver Demodulator RXD RX
TX
Transmitter Modulator TXD
The receiver and the transmitter must be enabled or disabled according to the direction of the
transmission to be managed.
IrDA Modulation For baud rates up to and including 115.2 Kbits/sec, the RZI modulation scheme is used. "0" is
represented by a light pulse of 3/16th of a bit time. Some examples of signal pulse duration
are shown in Table 81..
416 AT91RM9200
1768BATARM08/03
AT91RM9200
TXD
Bit Period 3
16 Bit Period
IrDA Baud Rate Table 82 gives some examples of CD values, baud rate error and pulse duration. Note that the
requirement on the maximum acceptable error of +/- 1.87% must be met.
417
1768BATARM08/03
IrDA Demodulator The demodulator is based on the IrDA Receive filter comprised of an 8-bit down counter which
is loaded with the value programmed in US_IF. When a falling edge is detected on the RXD
pin, the Filter Counter starts counting down at the Master Clock (MCK) speed. If a rising edge
is detected on the RXD pin, the counter stops and is reloaded with US_IF. If no rising edge is
detected when the counter reaches 0, the input of the receiver is driven low during one bit
time.
Figure 195 illustrates the operations of the IrDA demodulator.
RXD
Counter Pulse
Value 6 5 4 3 2 6 6 5 4 3 2 1 0 Accepted
Pulse
Rejected
Receiver
Input Driven Low During 16 Baud Rate Clock Cycles
As the IrDA mode uses the same logic as the ISO7816, note that the FI_DI_RATIO field in
US_FIDI must be set to a value higher than 0 in order to assure IrDA communications operate
correctly.
418 AT91RM9200
1768BATARM08/03
AT91RM9200
RS485 Mode The USART features the RS485 mode to enable line driver control. While operating in RS485
mode, the USART behaves as though in asynchronous or synchronous mode and configura-
tion of all the parameters are possible. The difference is that the RTS pin is driven low when
the transmitter is operating. The behavior of the RTS pin is controlled by the TXEMPTY bit. A
typical connection of the USART to a RS485 bus is shown in Figure 196.
USART
RXD
Differential
TXD Bus
RTS
The USART is set in RS485 mode by programming the USART_MODE field in the Mode Reg-
ister (US_MR) to the value 0x1.
The RTS pin is at a level inverse of the TXEMPTY bit. Significantly, the RTS pin remains low
when a timeguard is programmed so that the line can remain driven after the last character
completion. Figure 197 gives an example of the RTS waveform during a character transmis-
sion when the timeguard is enabled.
TXD
Start Parity Stop
D0 D1 D2 D3 D4 D5 D6 D7
Bit Bit Bit
Write
US_THR
TXRDY
TXEMPTY
RTS
419
1768BATARM08/03
Modem Mode The USART features modem mode, which enables control of the signals: DTR (Data Terminal
Ready), DSR (Data Set Ready), RTS (Request to Send), CTS (Clear to Send), DCD (Data
Carrier Detect) and RI (Ring Indicator). While operating in modem mode, the USART behaves
as a DTE (Data Terminal Equipment) as it drives DTR and RTS and can detect level change
on DSR, DCD, CTS and RI.
Setting the USART in modem mode is performed by writing the USART_MODE field in the
Mode Register (US_MR) to the value 0x3. While operating in modem mode the USART
behaves as though in asynchronous mode and all the parameter configurations are available.
Table 83 gives the correspondence of the USART signals with modem connection standards.
The control of the RTS and DTR output pins is performed by witting the Control Register
(US_CR) with the RTSDIS, RTSEN, DTRDIS and DTREN bits respectively at 1. The disable
command forces the corresponding pin to its inactive level, i.e. high. The enable commands
force the corresponding pin to its active level, i.e. low.
The level changes are detected on the RI, DSR, DCD and CTS pins. If an input change is
detected, the RIIC, DSRIC, DCDIC and CTSIC bits in the Channel Status Register (US_CSR)
are set respectively and can trigger an interrupt. The status is automatically cleared when
US_CSR is read. Furthermore, the CTS automatically disables the transmitter when it is
detected at its inactive state. If a character is being transmitted when the CTS rises, the char-
acter transmission is completed before the transmitter is actually disabled.
Test Modes The USART can be programmed to operate in three different test modes. The internal loop-
back capability allows on-board diagnostics. In the loopback mode the USART interface pins
are disconnected or not and reconfigured for loopback internally or externally.
Normal Mode As a reminder, the normal mode simply connects the RXD pin on the receiver input and the
transmitter output on the TXD pin.
TXD
Transmitter
420 AT91RM9200
1768BATARM08/03
AT91RM9200
Automatic Echo Automatic echo mode allows bit-by-bit retransmission. When a bit is received on the RXD pin,
it is sent to the TXD pin, as shown in Figure 199. Programming the transmitter has no effect on
the TXD pin. The RXD pin is still connected to the receiver input, thus the receiver remains
active.
TXD
Transmitter
Local Loopback The local loopback mode connects the output of the transmitter directly to the input of the
receiver, as shown in Figure 200. The TXD and RXD pins are not used. The RXD pin has no
effect on the receiver and the TXD pin is continuously driven high, as in idle state.
TXD
Transmitter 1
Remote Loopback Remote loopback mode directly connects the RXD pin to the TXD pin, as shown in Figure 201.
The transmitter and the receiver are disabled and have no effect. This mode allows bit-by-bit
retransmission.
TXD
Transmitter
421
1768BATARM08/03
USART User Interface
Table 84. USART Memory Map
Offset Register Name Access Reset State
0x0000 Control Register US_CR Write-only
0x0004 Mode Register US_MR Read/Write
0x0008 Interrupt Enable Register US_IER Write-only
0x000C Interrupt Disable Register US_IDR Write-only
0x0010 Interrupt Mask Register US_IMR Read-only 0
0x0014 Channel Status Register US_CSR Read-only
0x0018 Receiver Holding Register US_RHR Read-only 0
0x001C Transmitter Holding Register US_THR Write-only
0x0020 Baud Rate Generator Register US_BRGR Read/Write 0
0x0024 Receiver Time-out Register US_RTOR Read/Write 0
0x0028 Transmitter Timeguard Register US_TTGR Read/Write 0
0x2C
to Reserved
0x3C
0x0040 FI DI Ratio Register US_FIDI Read/Write 0x174
0x0044 Number of Errors Register US_NER Read-only
0x0048 Reserved
0x004C IrDA Filter Register US_IF Read/Write 0
0x5C
to Reserved
0xFC
0x100
to Reserved for PDC Registers
0x128
422 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
RTSDIS RTSEN DTRDIS DTREN
15 14 13 12 11 10 9 8
RETTO RSTNACK RSTIT SENDA STTTO STPBRK STTBRK RSTSTA
7 6 5 4 3 2 1 0
TXDIS TXEN RXDIS RXEN RSTTX RSTRX
423
1768BATARM08/03
STTTO: Start Time-out
0 = No effect
1 = Starts waiting for a character before clocking the time-out counter.
SENDA: Send Address
0 = No effect.
1 = In Multi-drop Mode only, the next character written to the US_THR is sent with the address bit set.
RSTIT: Reset Iterations
0 = No effect.
1 = Resets ITERATION in US_CSR. No effect if the ISO7816 is not enabled.
RSTNACK: Reset Non Acknowledge
0 = No effect
1 = Resets NACK in US_CSR.
RETTO: Rearm Time-out
0 = No effect
1 = Restart Time-out
DTREN: Data Terminal Ready Enable
0 = No effect.
1 = Drives the pin DTR at 0.
DTRDIS: Data Terminal Ready Disable
0 = No effect.
1 = Drives the pin DTR to 1.
RTSEN: Request to Send Enable
0 = No effect.
1 = Drives the pin RTS to 0.
RTSDIS: Request to Send Disable
0 = No effect.
1 = Drives the pin RTS to 1.
424 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
DSNACK INACK OVER CLKO MODE9 MSBF
15 14 13 12 11 10 9 8
CHMODE NBSTOP PAR SYNC
7 6 5 4 3 2 1 0
CHRL USCLKS USART_MODE
USART_MODE
425
1768BATARM08/03
SYNC: Synchronous Mode Select
0 = USART operates in Asynchronous Mode.
1 = USART operates in Synchronous Mode
PAR: Parity Type
426 AT91RM9200
1768BATARM08/03
AT91RM9200
427
1768BATARM08/03
USART Interrupt Enable Register
Name: US_IER
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
CTSIC DCDIC DSRIC RIIC
15 14 13 12 11 10 9 8
NACK RXBUFF TXBUFE ITERATION TXEMPTY TIMEOUT
7 6 5 4 3 2 1 0
PARE FRAME OVRE ENDTX ENDRX RXBRK TXRDY RXRDY
428 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
CTSIC DCDIC DSRIC RIIC
15 14 13 12 11 10 9 8
NACK RXBUFF TXBUFE ITERATION TXEMPTY TIMEOUT
7 6 5 4 3 2 1 0
PARE FRAME OVRE ENDTX ENDRX RXBRK TXRDY RXRDY
429
1768BATARM08/03
USART Interrupt Mask Register
Name: US_IMR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
CTSIC DCDIC DSRIC RIIC
15 14 13 12 11 10 9 8
NACK RXBUFF TXBUFE ITERATION TXEMPTY TIMEOUT
7 6 5 4 3 2 1 0
PARE FRAME OVRE ENDTX ENDRX RXBRK TXRDY RXRDY
430 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
CTS DCD DSR RI CTSIC DCDIC DSRIC RIIC
15 14 13 12 11 10 9 8
NACK RXBUFF TXBUFE ITERATION TXEMPTY TIMEOUT
7 6 5 4 3 2 1 0
PARE FRAME OVRE ENDTX ENDRX RXBRK TXRDY RXRDY
431
1768BATARM08/03
TXEMPTY: Transmitter Empty
0 = There are characters in either US_THR or the Transmit Shift Register, or the transmitter is disabled.
1 = There is at least one character in either US_THR or the Transmit Shift Register.
ITERATION: Max number of Repetitions Reached
0 = Maximum number of repetitions has not been reached since the last RSIT.
1 = Maximum number of repetitions has been reached since the last RSIT.
TXBUFE: Transmission Buffer Empty
0 = The signal Buffer Empty from the Transmit PDC channel is inactive.
1 = The signal Buffer Empty from the Transmit PDC channel is active.
RXBUFF: Reception Buffer Full
0 = The signal Buffer Full from the Receive PDC channel is inactive.
1 = The signal Buffer Full from the Receive PDC channel is active.
NACK: Non Acknowledge
0 = No Non Acknowledge has not been detected since the last RSTNACK.
1 = At least one Non Acknowledge has been detected since the last RSTNACK.
RIIC: Ring Indicator Input Change Flag
0 = No input change has been detected on the RI pin since the last read of US_CSR.
1 = At least one input change has been detected on the RI pin since the last read of US_CSR.
DSRIC: Data Set Ready Input Change Flag
0 = No input change has been detected on the DSR pin since the last read of US_CSR.
1 = At least one input change has been detected on the DSR pin since the last read of US_CSR.
DCDIC: Data Carrier Detect Input Change Flag
0 = No input change has been detected on the DCD pin since the last read of US_CSR.
1 = At least one input change has been detected on the DCD pin since the last read of US_CSR.
CTSIC: Clear to Send Input Change Flag
0 = No input change has been detected on the CTS pin since the last read of US_CSR.
1 = At least one input change has been detected on the CTS pin since the last read of US_CSR.
RI: Image of RI Input
0 = RI is at 0.
1 = RI is at 1.
DSR: Image of DSR Input
0 = DSR is at 0
1 = DSR is at 1.
DCD: Image of DCD Input
0 = DCD is at 0.
1 = DCD is at 1.
CTS: Image of CTS Input
0 = CTS is at 0.
1 = CTS is at 1.
432 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RXCHR
7 6 5 4 3 2 1 0
RXCHR
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
TXCHR
7 6 5 4 3 2 1 0
TXCHR
433
1768BATARM08/03
USART Baud Rate Generator Register
Name: US_BRGR
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
CD
7 6 5 4 3 2 1 0
CD
USART_MODE ISO7816
USART_MODE =
CD SYNC = 0 SYNC = 1
ISO7816
OVER = 0 OVER = 1
0 Baud Rate Clock Disabled
1 to 65535 Baud Rate = Baud Rate = Baud Rate = Selected Baud Rate = Selected
Selected Clock/16/CD Selected Clock/8/CD Clock /CD Clock/CD/FI_DI_RATIO
434 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
TO
7 6 5 4 3 2 1 0
TO
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
TG
435
1768BATARM08/03
USART FI DI RATIO Register
Name: US_FIDI
Access Type: Read/Write
Reset Value: 0x174
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
FI_DI_RATIO
7 6 5 4 3 2 1 0
FI_DI_RATIO
436 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
NB_ERRORS
437
1768BATARM08/03
USART IrDA FILTER Register
Name: US_IF
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
IRDA_FILTER
438 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Atmel Synchronous Serial Controller (SSC) provides a synchronous communication link
with external devices. It supports many serial synchronous communication protocols generally
used in audio and telecom applications such as I2S, Short Frame Sync, Long Frame Sync,
etc.
The SSC contains an independent receiver and transmitter and a common clock divider. The
receiver and the transmitter each interface with three signals: the TD/RD signal for data, the
TK/RK signal for the clock and the TF/RF signal for the Frame Sync. Transfers contain up to
16 data of up to 32 bits. they can be programmed to start automatically or on different events
detected on the Frame Sync signal.
The SSCs high-level of programmability and its two dedicated PDC channels of up to 32 bits
permit a continuous high bit rate data transfer without processor intervention.
Featuring connection to two PDC channels, the SSC permits interfacing with low processor
overhead to the following:
CODECs in master or slave mode
DAC through dedicated serial interface, particularly I2S
Magnetic card reader
Features of the SSC are:
Provides Serial Synchronous Communication Links Used in Audio and Telecom
Applications
Contains an Independent Receiver and Transmitter and a Common Clock Divider
Interfaced with Two PDC Channels (DMA Access) to Reduce Processor Overhead
Offers a Configurable Frame Sync and Data Length
Receiver and Transmitter can be Programmed to Start Automatically or on Detection of
Different Event on the Frame Sync Signal
Receiver and Transmitter Include a Data Signal, a Clock Signal and a Frame
Synchronization Signal
441
1768BATARM08/03
Block Diagram Figure 202. Block Diagram
ASB
APB Bridge
PDC
APB
TF
TK
TD
MCK
PMC
SSC Interface PIO
RF
RK
Interrupt Control
RD
SSC Interrupt
SSC
442 AT91RM9200
1768BATARM08/03
AT91RM9200
Product Dependencies
I/O Lines The pins used for interfacing the compliant external devices may be multiplexed with PIO
lines.
Before using the SSC receiver, the PIO controller must be configured to dedicate the SSC
receiver I/O lines to the SSC peripheral mode.
Before using the SSC transmitter, the PIO controller must be configured to dedicate the SSC
transmitter I/O lines to the SSC peripheral mode.
Power The SSC is not continuously clocked. The SSC interface may be clocked through the Power
Management Management Controller (PMC), therefore the programmer must first configure the PMC to
enable the SSC clock.
Interrupt The SSC interface has an interrupt line connected to the Advanced Interrupt Controller (AIC).
Handling interrupts requires programming the AIC before configuring the SSC.
All SSC interrupts can be enabled/disabled configuring the SSC Interrupt mask register. Each
pending and unmasked SSC interrupt will assert the SSC interrupt line. The SSC interrupt ser-
vice routine can get the interrupt origin by reading the SSC interrupt status register.
443
1768BATARM08/03
Functional This chapter contains the functional description of the following: SSC Functional Block, Clock
Description Management, Data format, Start, Transmitter, Receiver and Frame Sync.
The receiver and transmitter operate separately. However, they can work synchronously by
programming the receiver to use the transmit clock and/or to start a data transfer when trans-
mission starts. Alternatively, this can be done by programming the transmitter to use the
receive clock and/or to start a data transfer when reception starts. The transmitter and the
receiver can be programmed to operate with the clock signals provided on either the TK or RK
pins. This allows the SSC to support many slave-mode data transfers. The maximum clock
speed allowed on the TK and RK pins is the master clock divided by 2. Each level of the clock
must be stable for at least two master clock periods.
Transmitter
Clock Output
TK
Controller
TK Input
MCK Clock Transmit Clock TX clock Frame Sync TF
Divider Controller Controller
RX clock
TF
Start
RF Transmit Shift Register TD
Selector
TX PDC Transmit Holding Transmit Sync
Register Holding Register
APB
Load Shift
User
Interface
RK Input
Receive Clock RX Clock Frame Sync
Controller RF
Controller
TX Clock
RF
Start
TF Receive Shift Register RD
Selector
RX PDC Receive Holding Receive Sync
Register Holding Register
PDC Interrupt Control
Load Shift
AIC
444 AT91RM9200
1768BATARM08/03
AT91RM9200
SSC_CMR
The Master Clock divider is determined by the 12-bit field DIV counter and comparator (so its
maximal value is 4095) in the Clock Mode Register SSC_CMR, allowing a Master Clock divi-
sion by up to 8190. The Divided Clock is provided to both the Receiver and Transmitter. When
this field is programmed to 0, the Clock Divider is not used and remains inactive.
When DIV is set to a value equal or greater to 1, the Divided Clock has a frequency of Master
Clock divided by 2 times DIV. Each level of the Divided Clock has a duration of the Master
Clock multiplied by DIV. This ensures a 50% duty cycle for the Divided Clock regardless if the
DIV value is even or odd.
Divided Clock
DIV = 1
Master Clock
Divided Clock
DIV = 3
445
1768BATARM08/03
Transmitter Clock The transmitter clock is generated from the receiver clock or the divider clock or an external
Management clock scanned on the TK I/O pad. The transmitter clock is selected by the CKS field in
SSC_TCMR (Transmit Clock Mode Register). Transmit Clock can be inverted independently
by the CKI bits in SSC_TCMR.
The transmitter can also drive the TK I/O pad continuously or be limited to the actual data
transfer. The clock output is configured by the SSC_TCMR register. The Transmit Clock Inver-
sion (CKI) bits have no effect on the clock outputs. Programming the TCMR register to select
TK pin (CKS field) and at the same time Continuous Transmit Clock (CKO field) might lead to
unpredictable results.
Receiver Clock TK
Divider Clock
0 Transmitter Clock
SSC_TCMR.CKI
Receiver Clock The receiver clock is generated from the transmitter clock or the divider clock or an external
Management clock scanned on the RK I/O pad. The Receive Clock is selected by the CKS field in
SSC_RCMR (Receive Clock Mode Register). Receive Clocks can be inverted independently
by the CKI bits in SSC_RCMR.
The receiver can also drive the RK I/O pad continuously or be limited to the actual data trans-
fer. The clock output is configured by the SSC_RCMR register. The Receive Clock Inversion
(CKI) bits have no effect on the clock outputs. Programming the RCMR register to select RK
pin (CKS field) and at the same time Continuous Receive Clock (CKO field) might lead to
unpredictable results.
Transmitter Clock RK
Divider Clock
0 Receiver Clock
SSC_RCMR.CKI
446 AT91RM9200
1768BATARM08/03
AT91RM9200
Transmitter A transmitted frame is triggered by a start event and can be followed by synchronization data
Operations before data transmission.
The start event is configured by setting the Transmit Clock Mode Register (SSC_TCMR). See
Start on page 448.
The frame synchronization is configured setting the Transmit Frame Mode Register
(SSC_TFMR). See Frame Sync on page 450.
To transmit data, the transmitter uses a shift register clocked by the transmitter clock signal
and the start mode selected in the SSC_TCMR. Data is written by the application to the
SSC_THR register then transferred to the shift register according to the data format selected.
When both the SSC_THR and the transmit shift register are empty, the status flag TXEMPTY
is set in SSC_SR. When the Transmit Holding register is transferred in the Transmit shift reg-
ister, the status flag TXRDY is set in SSC_SR and additional data can be loaded in the holding
register.
SSC_CR.TXDIS
SSC_TFMR.DATDEF SSC_TCMR.STTDLY
SSC_TFMR.FSDEN
SSC_TFMR.DATNB
1
TD
SSC_TFMR.MSBF 0
RF TF
Transmitter Clock
Start
Transmit Shift Register
Selector
SSC_TFMR.FSDEN 0 1
SSC_TCMR.STTDLY
447
1768BATARM08/03
Receiver A received frame is triggered by a start event and can be followed by synchronization data
Operations before data transmission.
The start event is configured setting the Receive Clock Mode Register (SSC_RCMR). See
Start on page 448.
The frame synchronization is configured setting the Receive Frame Mode Register
(SSC_RFMR). See Frame Sync on page 450.
The receiver uses a shift register clocked by the receiver clock signal and the start mode
selected in the SSC_RCMR. The data is transferred from the shift register in function of data
format selected.
When the receiver shift register is full, the SSC transfers this data in the holding register, the
status flag RXRDY is set in SSC_SR and the data can be read in the receiver holding register,
if another transfer occurs before read the RHR register, the status flag OVERUN is set in
SSC_SR and the receiver shift register is transferred in the RHR register.
SSC_CR.RXEN
SSC_SR.RXEN
SSC_CR.RXDIS
SSC_RFMR.MSBF SSC_RFMR.DATNB
RF TF
Receiver Clock
Start
Receive Shift Register RD
Selector
SSC_RSHR SSC_RHR
SSC_RCMR.STTDLY
SSC_RFMR.FSLEN SSC_RFMR.DATLEN
Start The transmitter and receiver can both be programmed to start their operations when an event
occurs, respectively in the Transmit Start Selection (START) field of SSC_TCMR and in the
Receive Start Selection (START) field of SSC_RCMR.
Under the following conditions the start event is independently programmable:
Continuous. In this case, the transmission starts as soon as a word is written in SSC_THR
and the reception starts as soon as the Receiver is enabled.
Synchronously with the transmitter/receiver
On detection of a falling/rising edge on TK/RK
On detection of a low level/high level on TK/RK
On detection of a level change or an edge on TK/RK
A start can be programmed in the same manner on either side of the Transmit/Receive Clock
Register (RCMR/TCMR). Thus, the start could be on TF (Transmit) or RF (Receive).
Detection on TF/RF input/output is done through the field FSOS of the Transmit / Receive
Frame Mode Register (TFMR/RFMR).
448 AT91RM9200
1768BATARM08/03
AT91RM9200
Generating a Frame Sync signal is not possible without generating it on its related output.
TF
(Input)
RK
RF
(Input)
RD
Start = Any Edge on RF X BO B1 BO B1
(Input)
STTDLY
449
1768BATARM08/03
Frame Sync The Transmitter and Receiver Frame Sync pins, TF and RF, can be programmed to generate
different kinds of frame synchronization signals. The Frame Sync Output Selection (FSOS)
field in the Receive Frame Mode Register (SSC_RFMR) and in the Transmit Frame Mode
Register (SSC_TFMR) are used to select the required waveform.
Programmable low or high levels during data transfer are supported.
Programmable high levels before the start of data transfers or toggling are also supported.
If a pulse waveform is selected, the Frame Sync Length (FSLEN) field in SSC_RFMR and
SSC_TFMR programs the length of the pulse, from 1-bit time up to 16-bit time.
The periodicity of the Receive and Transmit Frame Sync pulse output can be programmed
through the Period Divider Selection (PERIOD) field in SSC_RCMR and SSC_TCMR.
Frame Sync Data Frame Sync Data transmits or receives a specific tag during the Frame Synchro signal.
During the Frame Sync signal, the Receiver can sample the RD line and store the data in the
Receive Sync Holding Register and the transmitter can transfer Transmit Sync Holding Regis-
ter in the Shifter Register. The data length to be sampled/shifted out during the Frame Sync
signal is programmed by the FSLEN field in SSC_RFMR/SSC_TFMR.
Concerning the Receive Frame Sync Data operation, if the Frame Sync Length is equal to or
lower than the delay between the start event and the actual data reception, the data sampling
operation is performed in the Receive Sync Holding Register through the Receive Shift
Register.
The Transmit Frame Sync Operation is performed by the transmitter only if the bit Frame Sync
Data Enable (FSDEN) in SSC_TFMR is set. If the Frame Sync length is equal to or lower than
the delay between the start event and the actual data transmission, the normal transmission
has priority and the data contained in the Transmit Sync Holding Register is transferred in the
Transmit Register then shifted out.
Frame Sync Edge The Frame Sync Edge detection is programmed by the FSEDGE field in
Detection SSC_RFMR/SSC_TFMR. This sets the corresponding flags RXSYN/TXSYN in the SSC Sta-
tus Register (SSC_SR) on frame synchro edge detection (signals RF/TF).
Data Format The data framing format of both the transmitter and the receiver are largely programmable
through the Transmitter Frame Mode Register (SSC_TFMR) and the Receiver Frame Mode
Register (SSC_RFMR). In either case, the user can independently select:
The event that starts the data transfer (START).
The delay in number of bit periods between the start event and the first data bit (STTDLY).
The length of the data (DATLEN)
The number of data to be transferred for each start event (DATNB).
The length of Synchronization transferred for each start event (FSLEN).
The bit sense: most or lowest significant bit first (MSBF).
Additionally, the transmitter can be used to transfer Synchronization and select the level
driven on the TD pin while not in data transfer operation. This is done respectively by the
Frame Sync Data Enable (FSDEN) and by the Data Default Value (DATDEF) bits in
SSC_TFMR.
450 AT91RM9200
1768BATARM08/03
AT91RM9200
Figure 213. Transmit and Receive Frame Format in Edge/Pulse Start Modes
Start Start
PERIOD
TF/RF(1)
FSLEN
DATNB
451
1768BATARM08/03
Figure 214. Transmit Frame Format in Continuous Mode
Start
DATLEN DATLEN
Start: 1. TXEMPTY set to 1
2. Write to the SSC_THR
Note: 1. STTDLY is set to 0. In this example, SSC_THR is loaded twice. The value of FSDEN has no
effect on transmission. SyncData cannot be output in continuous mode.
RD Data Data
To SSC_RHR To SSC_RHR
DATLEN DATLEN
Loop Mode The receiver can be programmed to receive transmissions from the transmitter. This is done
by setting the Loop Mode (LOOP) bit in SSC_RFMR. In this case, RD is connected to TD, RF
is connected to TF and RK is connected to TK.
Interrupt Most bits in SSC_SR have a corresponding bit in interrupt management registers.
The SSC Controller can be programmed to generate an interrupt when it detects an event.
The Interrupt is controlled by writing SSC_IER (Interrupt Enable Register) and SSC_IDR
(Interrupt Disable Register), which respectively enable and disable the corresponding interrupt
by setting and clearing the corresponding bit in SSC_IMR (Interrupt Mask Register), which
controls the generation of interrupts by asserting the SSC interrupt line connected to the AIC.
452 AT91RM9200
1768BATARM08/03
AT91RM9200
SSC_IMR
SSC_IER SSC_IDR
PDC Set Clear
TXBUFE
ENDTX
Transmitter
TXRDY
TXEMPTY
TXSYNC SSC Interrupt
Interrupt
RXBUFF Control
ENDRX
Receiver
RXRDY
OVRUN
RXSYNC
SSC The SSC can support several serial communication modes used in audio or high speed serial
links. Some standard applications are shown in the following figures. All serial link applications
Application
supported by the SSC are not listed here.
Examples
Figure 217. Audio Application Block Diagram
Clock SCK
TK
Word Select WS
TF I2S
RECEIVER
Data SD
TD
SSC
RD Clock SCK
RF Word Select WS
453
1768BATARM08/03
Figure 218. Codec Application Block Diagram
Serial Data Clock (SCLK)
TK
Frame sync (FSYNC)
TF
CODEC
Serial Data Out
TD
SSC
Serial Data In
RD
RF
Serial Data Clock (SCLK)
Serial Data In
SSC
Data in
RD
RF
RK
CODEC
Second
Time Slot
Serial Data in
454 AT91RM9200
1768BATARM08/03
AT91RM9200
455
1768BATARM08/03
SSC Control Register
Name: SSC_CR
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
SWRST TXDIS TXEN
7 6 5 4 3 2 1 0
RXDIS RXEN
456 AT91RM9200
1768BATARM08/03
AT91RM9200
457
1768BATARM08/03
SSC Receive Clock Mode Register
Name: SSC_RCMR
Access Type: Read/Write
31 30 29 28 27 26 25 24
PERIOD
23 22 21 20 19 18 17 16
STTDLY
15 14 13 12 11 10 9 8
START
7 6 5 4 3 2 1 0
CKI CKO CKS
458 AT91RM9200
1768BATARM08/03
AT91RM9200
459
1768BATARM08/03
SSC Receive Frame Mode Register
Name: SSC_RFMR
Access Type: Read/Write
31 30 29 28 27 26 25 24
FSEDGE
23 22 21 20 19 18 17 16
FSOS FSLEN
15 14 13 12 11 10 9 8
DATNB
7 6 5 4 3 2 1 0
MSBF LOOP DATLEN
460 AT91RM9200
1768BATARM08/03
AT91RM9200
461
1768BATARM08/03
SSC Transmit Clock Mode Register
Name: SSC_TCMR
Access Type: Read/Write
31 30 29 28 27 26 25 24
PERIOD
23 22 21 20 19 18 17 16
STTDLY
15 14 13 12 11 10 9 8
START
7 6 5 4 3 2 1 0
CKI CKO CKS
462 AT91RM9200
1768BATARM08/03
AT91RM9200
463
1768BATARM08/03
SSC Transmit Frame Mode Register
Name: SSC_TFMR
Access Type: Read/Write
31 30 29 28 27 26 25 24
FSEDGE
23 22 21 20 19 18 17 16
FSDEN FSOS FSLEN
15 14 13 12 11 10 9 8
DATNB
7 6 5 4 3 2 1 0
MSBF DATDEF DATLEN
464 AT91RM9200
1768BATARM08/03
AT91RM9200
465
1768BATARM08/03
SSC Receive Holding Register
Name: SSC_RHR
Access Type: Read-only
31 30 29 28 27 26 25 24
RDAT
23 22 21 20 19 18 17 16
RDAT
15 14 13 12 11 10 9 8
RDAT
7 6 5 4 3 2 1 0
RDAT
466 AT91RM9200
1768BATARM08/03
AT91RM9200
467
1768BATARM08/03
SSC Status Register
Register Name: SSC_SR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
RXEN TXEN
15 14 13 12 11 10 9 8
RXSYN TXSYN
7 6 5 4 3 2 1 0
RXBUFF ENDRX OVRUN RXRDY TXBUFE ENDTX TXEMPTY TXRDY
468 AT91RM9200
1768BATARM08/03
AT91RM9200
469
1768BATARM08/03
SSC Interrupt Enable Register
Register Name: SSC_IER
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RXSYN TXSYN
7 6 5 4 3 2 1 0
RXBUFF ENDRX OVRUN RXRDY TXBUFE ENDTX TXEMPTY TXRDY
470 AT91RM9200
1768BATARM08/03
AT91RM9200
471
1768BATARM08/03
SSC Interrupt Mask Register
Register Name: SSC_IMR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RXSYN TXSYN
7 6 5 4 3 2 1 0
RXBUFF ENDRX OVRUN RXRDY TXBUFE ENDTX TXEMPTY TXRDY
472 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Timer Counter (TC) includes three identical 16-bit Timer Counter channels.
Each channel can be independently programmed to perform a wide range of functions includ-
ing frequency measurement, event counting, interval measurement, pulse generation, delay
timing and pulse width modulation.
Each channel has three external clock inputs, five internal clock inputs and two multi-purpose
input/output signals which can be configured by the user. Each channel drives an internal
interrupt signal which can be programmed to generate processor interrupts.
The Timer Counter block has two global registers which act upon all three TC channels.
The Block Control Register allows the three channels to be started simultaneously with the
same instruction.
The Block Mode Register defines the external clock inputs for each channel, allowing them to
be chained.
Key Features of the Timer Counter are:
Three 16-bit Timer Counter Channels
A Wide Range of Functions Including:
Frequency Measurement
Event Counting
Interval Measurement
Pulse Generation
Delay Timing
Pulse Width Modulation
Up/down Capabilities
Each Channel is User-configurable and Contains:
Three External Clock Inputs
Five Internal Clock Inputs
Two Multi-purpose Input/Output Signals
Internal Interrupt Signal
Two Global Registers that Act on All Three TC Channels
473
1768BATARM08/03
Block Diagram Figure 220. Timer Counter Block Diagram
Parallel I/O
TIMER_CLOCK1 Controller
TCLK0
TCLK0
TIMER_CLOCK2 TCLK1
TIOA1 TCLK2
TCLK0
TCLK2 SYNC
INT1
TC1XC1S
Timer Counter
Advanced
Interrupt
Controller
474 AT91RM9200
1768BATARM08/03
AT91RM9200
Product For further details on the Timer Counter hardware implementation, see the specific Product
Dependencies Properties document.
I/O Lines The pins used for interfacing the compliant external devices may be multiplexed with PIO
lines. The programmer must first program the PIO controllers to assign the TC pins to their
peripheral functions.
Power The TC must be clocked through the Power Management Controller (PMC), thus the program-
Management mer must first configure the PMC to enable the Timer Counter.
Interrupt The TC interface has an interrupt line connected to the Advanced Interrupt Controller (AIC).
Handling the TC interrupt requires programming the AIC before configuring the TC.
Functional
Description
TC Description The three channels of the Timer Counter are independent and identical in operation. The reg-
isters for channel programming are listed in Table 90 on page 475.
16-bit Counter Each channel is organized around a 16-bit counter. The value of the counter is incremented at
each positive edge of the selected clock. When the counter has reached the value 0xFFFF
and passes to 0x0000, an overflow occurs and the COVFS bit in TC_SR (Status Register) is
set.
The current value of the counter is accessible in real time by reading the Counter Value Regis-
ter, TC_CV. The counter can be reset by a trigger. In this case, the counter value passes to
0x0000 on the next valid edge of the selected clock.
Clock Selection At block level, input clock signals of each channel can either be connected to the external
inputs TCLK0, TCLK1 or TCLK2, or be connected to the configurable I/O signals TIOA0,
TIOA1 or TIOA2 for chaining by programming the TC_BMR (Block Mode). See Figure 221.
Each channel can independently select an internal or external clock source for its counter:
Internal clock signals: TIMER_CLOCK1, TIMER_CLOCK2, TIMER_CLOCK3,
TIMER_CLOCK4, TIMER_CLOCK5
External clock signals: XC0, XC1 or XC2
This selection is made by the TCCLKS bits in the TC Channel Mode Register (Capture Mode).
The selected clock can be inverted with the CLKI bit in TC_CMR (Capture Mode). This allows
counting on the opposite edges of the clock.
475
1768BATARM08/03
The burst function allows the clock to be validated when an external signal is high. The
BURST parameter in the Mode Register defines this signal (none, XC0, XC1, XC2).
Note: In all cases, if an external clock is used, the duration of each of its levels must be longer than the
master clock period. The external clock frequency must be at least 2.5 times lower than the
master clock
TCCLKS
TIMER_CLOCK1
TIMER_CLOCK2 CLKI
TIMER_CLOCK3
TIMER_CLOCK4
TIMER_CLOCK5
Selected
XC0 Clock
XC1
XC2
BURST
Clock Control The clock of each counter can be controlled in two different ways: it can be enabled/disabled
and started/stopped. See Figure 222.
The clock can be enabled or disabled by the user with the CLKEN and the CLKDIS
commands in the Control Register. In Capture Mode it can be disabled by an RB load
event if LDBDIS is set to 1 in TC_CMR. In Waveform Mode, it can be disabled by an RC
Compare event if CPCDIS is set to 1 in TC_CMR. When disabled, the start or the stop
actions have no effect: only a CLKEN command in the Control Register can re-enable the
clock. When the clock is enabled, the CLKSTA bit is set in the Status Register.
The clock can also be started or stopped: a trigger (software, synchro, external or
compare) always starts the clock. The clock can be stopped by an RB load event in
Capture Mode (LDBSTOP = 1 in TC_CMR) or a RC compare event in Waveform Mode
(CPCSTOP = 1 in TC_CMR). The start and the stop commands have effect only if the
clock is enabled.
476 AT91RM9200
1768BATARM08/03
AT91RM9200
Q S
R
Q S
R
Stop Disable
Counter Event Event
Clock
TC Operating Modes Each channel can independently operate in two different modes:
Capture Mode provides measurement on signals.
Waveform Mode provides wave generation.
The TC Operating Mode is programmed with the WAVE bit in the TC Channel Mode Register.
In Capture Mode, TIOA and TIOB are configured as inputs.
In Waveform Mode, TIOA is always configured to be an output and TIOB is an output if it is not
selected to be the external trigger.
Trigger A trigger resets the counter and starts the counter clock. Three types of triggers are common
to both modes, and a fourth external trigger is available to each mode.
The following triggers are common to both modes:
Software Trigger: Each channel has a software trigger, available by setting SWTRG in
TC_CCR.
SYNC: Each channel has a synchronization signal SYNC. When asserted, this signal has
the same effect as a software trigger. The SYNC signals of all channels are asserted
simultaneously by writing TC_BCR (Block Control) with SYNC set.
Compare RC Trigger: RC is implemented in each channel and can provide a trigger when
the counter value matches the RC value if CPCTRG is set in TC_CMR.
The channel can also be configured to have an external trigger. In Capture Mode, the external
trigger signal can be selected between TIOA and TIOB. In Waveform Mode, an external event
can be programmed on one of the following signals: TIOB, XC0, XC1 or XC2. This external
event can then be programmed to perform a trigger by setting ENETRG in TC_CMR.
If an external trigger is used, the duration of the pulses must be longer than the master clock
period in order to be detected.
Regardless of the trigger used, it will be taken into account at the following active edge of the
selected clock. This means that the counter value can be read differently from zero just after a
trigger, especially when a low frequency signal is selected as the clock.
477
1768BATARM08/03
Capture Operating This mode is entered by clearing the WAVE parameter in TC_CMR (Channel Mode Register).
Mode Capture Mode allows the TC channel to perform measurements such as pulse timing, fre-
quency, period, duty cycle and phase on TIOA and TIOB signals which are considered as
inputs.
Figure 223 shows the configuration of the TC channel when programmed in Capture Mode.
Capture Registers A Registers A and B (RA and RB) are used as capture registers. This means that they can be
and B loaded with the counter value when a programmable event occurs on the signal TIOA.
The LDRA parameter in TC_CMR defines the TIOA edge for the loading of register A, and the
LDRB parameter defines the TIOA edge for the loading of Register B.
RA is loaded only if it has not been loaded since the last trigger or if RB has been loaded since
the last loading of RA.
RB is loaded only if RA has been loaded since the last trigger or the last loading of RB.
Loading RA or RB before the read of the last value loaded sets the Overrun Error Flag
(LOVRS) in TC_SR (Status Register). In this case, the old value is overwritten.
Trigger Conditions In addition to the SYNC signal, the software trigger and the RC compare trigger, an external
trigger can be defined.
The ABETRG bit in TC_CMR selects TIOA or TIOB input signal as an external trigger. The
ETRGEDG parameter defines the edge (rising, falling or both) detected to generate an exter-
nal trigger. If ETRGEDG = 0 (none), the external trigger is disabled.
478 AT91RM9200
1768BATARM08/03
TCCLKS
CLKSTA CLKEN CLKDIS
CLKI
1768BATARM08/03
TIMER_CLOCK1
TIMER_CLOCK2
TIMER_CLOCK3
Q S
TIMER_CLOCK4
TIMER_CLOCK5 R
Figure 223. Capture Mode
Q S
XC0
R
XC1
XC2
LDBSTOP LDBDIS
BURST
Register C
Capture Capture
1 Register A Register B Compare RC =
16-bit Counter
SWTRG
CLK
OVF
RESET
SYNC
Trig
ABETRG
ETRGEDG CPCTRG
MTIOB Edge
Detector
TIOB
LDRA LDRB
CPCS
LDRAS
LDRBS
LOVRS
ETRGS
COVFS
TC1_SR
Timer/Counter Channel
INT
AT91RM9200
479
Waveform Waveform operating mode is entered by setting the WAVE parameter in TC_CMR (Channel
Operating Mode Mode Register).
In Waveform Operating Mode the TC channel generates 1 or 2 PWM signals with the same
frequency and independently programmable duty cycles, or generates different types of one-
shot or repetitive pulses.
In this mode, TIOA is configured as an output and TIOB is defined as an output if it is not used
as an external event (EEVT parameter in TC_CMR).
Figure 224 shows the configuration of the TC channel when programmed in Waveform Oper-
ating Mode.
Waveform Selection Depending on the WAVSEL parameter in TC_CMR (Channel Mode Register), the behavior of
TC_CV varies.
With any selection, RA, RB and RC can all be used as compare registers.
RA Compare is used to control the TIOA output, RB Compare is used to control the TIOB out-
put (if correctly configured) and RC Compare is used to control TIOA and/or TIOB outputs.
480 AT91RM9200
1768BATARM08/03
1768BATARM08/03
TCCLKS
CLKSTA CLKEN CLKDIS
TIMER_CLOCK1 ACPC
CLKI
TIMER_CLOCK2
TIMER_CLOCK3
Q S
TIMER_CLOCK4 CPCDIS MTIOA
TIMER_CLOCK5 R ACPA
Q S
XC0
R
XC1
XC2 CPCSTOP TIOA
AEEVT
Output Controller
Figure 224. Waveform Mode
BURST
Register A Register B Register C
WAVSEL
ASWTRG
1 Compare RA = Compare RB = Compare RC =
16-bit Counter
CLK
OVF
RESET
SWTRG
BCPC
SYNC
Trig
BCPB MTIOB
WAVSEL
EEVT
TIOB
BEEVT
EEVTEDG
ENETRG
Output Controller
CPAS
CPBS
CPCS
ETRGS
COVFS
Edge
TC1_SR
Detector BSWTRG
TIOB
TC1_IMR
Timer/Counter Channel
INT
AT91RM9200
481
WAVSEL = 00 When WAVSEL = 00, the value of TC_CV is incremented from 0 to 0xFFFF. Once 0xFFFF
has been reached, the value of TC_CV is reset. Incrementation of TC_CV starts again and the
cycle continues. See Figure 225.
An external event trigger or a software trigger can reset the value of TC_CV. It is important to
note that the trigger may occur at any time. See Figure 226.
RC Compare cannot be programmed to generate a trigger in this configuration. At the same
time, RC Compare can stop the counter clock (CPCSTOP = 1 in TC_CMR) and/or disable the
counter clock (CPCDIS = 1 in TC_CMR).
0xFFFF
RC
RB
RA
TIOB
TIOA
0xFFFF
RB
RA
Time
Waveform Examples
TIOB
TIOA
482 AT91RM9200
1768BATARM08/03
AT91RM9200
WAVSEL = 10 When WAVSEL = 10, the value of TC_CV is incremented from 0 to the value of RC, then auto-
matically reset on a RC Compare. Once the value of TC_CV has been reset, it is then
incremented and so on. See Figure 227.
It is important to note that TC_CV can be reset at any time by an external event or a software
trigger if both are programmed correctly. See Figure 228.
In addition, RC Compare can stop the counter clock (CPCSTOP = 1 in TC_CMR) and/or dis-
able the counter clock (CPCDIS = 1 in TC_CMR).
0xFFFF
Counter cleared by compare match with RC
RC
RB
RA
TIOB
TIOA
0xFFFF
Counter cleared by compare match with RC Counter cleared by trigger
RC
RB
RA
TIOB
TIOA
483
1768BATARM08/03
WAVSEL = 01 When WAVSEL = 01, the value of TC_CV is incremented from 0 to 0xFFFF. Once 0xFFFF is
reached, the value of TC_CV is decremented to 0, then re-incremented to 0xFFFF and so on.
See Figure 229.
A trigger such as an external event or a software trigger can modify TC_CV at any time. If a
trigger occurs while TC_CV is incrementing, TC_CV then decrements. If a trigger is received
while TC_CV is decrementing, TC_CV then increments. See Figure 230.
RC Compare cannot be programmed to generate a trigger in this configuration.
At the same time, RC Compare can stop the counter clock (CPCSTOP = 1) and/or disable the
counter clock (CPCDIS = 1).
0xFFFF
RC
RB
RA
TIOB
TIOA
0xFFFF
Counter decremented
by trigger
RC
RB
Counter incremented
by trigger
RA
TIOB
TIOA
484 AT91RM9200
1768BATARM08/03
AT91RM9200
WAVSEL = 11 When WAVSEL = 11, the value of TC_CV is incremented from 0 to RC. Once RC is reached,
the value of TC_CV is decremented to 0, then re-incremented to RC and so on. See Figure
231.
A trigger such as an external event or a software trigger can modify TC_CV at any time. If a
trigger occurs while TC_CV is incrementing, TC_CV then decrements. If a trigger is received
while TC_CV is decrementing, TC_CV then increments. See Figure 232.
RC Compare can stop the counter clock (CPCSTOP = 1) and/or disable the counter clock
(CPCDIS = 1).
0xFFFF
Counter decremented by compare match with RC
RC
RB
RA
TIOB
TIOA
0xFFFF
Counter decremented by compare match with RC
RC
Counter decremented
by trigger
RB
Counter incremented
by trigger
RA
TIOB
TIOA
485
1768BATARM08/03
External Event/Trigger An external event can be programmed to be detected on one of the clock sources (XC0, XC1,
Conditions XC2) or TIOB. The external event selected can then be used as a trigger.
The parameter EEVT parameter in TC_CMR selects the external trigger. The EEVTEDG
parameter defines the trigger edge for each of the possible external triggers (rising, falling or
both). If EEVTEDG is cleared (none), no external event is defined.
If TIOB is defined as an external event signal (EEVT = 0), TIOB is no longer used as an output
and the TC channel can only generate a waveform on TIOA.
When an external event is defined, it can be used as a trigger by setting bit ENETRG in
TC_CMR.
As in Capture Mode, the SYNC signal and the software trigger are also available as triggers.
RC Compare can also be used as a trigger depending on the parameter WAVSEL.
Output Controller The output controller defines the output level changes on TIOA and TIOB following an event.
TIOB control is used only if TIOB is defined as output (not as an external event).
The following events control TIOA and TIOB: software trigger, external event and RC com-
pare. RA compare controls TIOA and RB compare controls TIOB. Each of these events can
be programmed to set, clear or toggle the output as defined in the corresponding parameter in
TC_CMR.
486 AT91RM9200
1768BATARM08/03
AT91RM9200
TC_BCR (Block Control Register) and TC_BMR (Block Mode Register) control the whole TC block. TC channels are con-
trolled by the registers listed in Table 92. The offset of each of the channel registers in Table 92 is in relation to the offset of
the corresponding channel as mentioned in Table 92.
487
1768BATARM08/03
TC Block Control Register
Register Name: TC_BCR
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
SYNC
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
TC2XC2S TCXC1S TC0XC0S
488 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
SWTRG CLKDIS CLKEN
489
1768BATARM08/03
TC Channel Mode Register: Capture Mode
Register Name: TC_CMR
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
LDRB LDRA
15 14 13 12 11 10 9 8
WAVE = 0 CPCTRG ABETRG ETRGEDG
7 6 5 4 3 2 1 0
LDBDIS LDBSTOP BURST CLKI TCCLKS
BURST
0 0 The clock is not gated by an external signal.
0 1 XC0 is ANDed with the selected clock.
1 0 XC1 is ANDed with the selected clock.
1 1 XC2 is ANDed with the selected clock.
490 AT91RM9200
1768BATARM08/03
AT91RM9200
ETRGEDG Edge
0 0 none
0 1 rising edge
1 0 falling edge
1 1 each edge
LDRA Edge
0 0 none
0 1 rising edge of TIOA
1 0 falling edge of TIOA
1 1 each edge of TIOA
LDRB Edge
0 0 none
0 1 rising edge of TIOA
1 0 falling edge of TIOA
1 1 each edge of TIOA
491
1768BATARM08/03
TC Channel Mode Register: Waveform Mode
Register Name: TC_CMR
Access Type: Read/Write
31 30 29 28 27 26 25 24
BSWTRG BEEVT BCPC BCPB
23 22 21 20 19 18 17 16
ASWTRG AEEVT ACPC ACPA
15 14 13 12 11 10 9 8
WAVE = 1 WAVSEL ENETRG EEVT EEVTEDG
7 6 5 4 3 2 1 0
CPCDIS CPCSTOP BURST CLKI TCCLKS
BURST
0 0 The clock is not gated by an external signal.
0 1 XC0 is ANDed with the selected clock.
1 0 XC1 is ANDed with the selected clock.
1 1 XC2 is ANDed with the selected clock.
492 AT91RM9200
1768BATARM08/03
AT91RM9200
EEVTEDG Edge
0 0 none
0 1 rising edge
1 0 falling edge
1 1 each edge
WAVSEL Effect
0 0 UP mode without automatic trigger on RC Compare
1 0 UP mode with automatic trigger on RC Compare
0 1 UPDOWN mode without automatic trigger on RC Compare
1 1 UPDOWN mode with automatic trigger on RC Compare
WAVE = 1
0 = Waveform Mode is disabled (Capture Mode is enabled).
1 = Waveform Mode is enabled.
ACPA: RA Compare Effect on TIOA
ACPA Effect
0 0 none
0 1 set
1 0 clear
1 1 toggle
ACPC Effect
0 0 none
0 1 set
1 0 clear
1 1 toggle
493
1768BATARM08/03
AEEVT: External Event Effect on TIOA
AEEVT Effect
0 0 none
0 1 set
1 0 clear
1 1 toggle
ASWTRG Effect
0 0 none
0 1 set
1 0 clear
1 1 toggle
BCPB Effect
0 0 none
0 1 set
1 0 clear
1 1 toggle
BCPC Effect
0 0 none
0 1 set
1 0 clear
1 1 toggle
BEEVT Effect
0 0 none
0 1 set
1 0 clear
1 1 toggle
BSWTRG Effect
0 0 none
0 1 set
1 0 clear
1 1 toggle
494 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
CV
7 6 5 4 3 2 1 0
CV
TC Register A
Register Name: TC_RA
Access Type: Read-only if WAVE = 0, Read/Write if WAVE = 1
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RA
7 6 5 4 3 2 1 0
RA
RA: Register A
RA contains the Register A value in real time.
TC Register B
Register Name: TC_RB
Access Type: Read-only if WAVE = 0, Read/Write if WAVE = 1
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RB
7 6 5 4 3 2 1 0
RB
RB: Register B
RB contains the Register B value in real time.
495
1768BATARM08/03
TC Register C
Register Name: TC_RC
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RC
7 6 5 4 3 2 1 0
RC
RC: Register C
RC contains the Register C value in real time.
TC Status Register
Register Name: TC_SR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
MTIOB MTIOA CLKSTA
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
ETRGS LDRBS LDRAS CPCS CPBS CPAS LOVRS COVFS
496 AT91RM9200
1768BATARM08/03
AT91RM9200
0 = RA Load has not occurred since the last read of the Status Register or WAVE = 1.
1 = RA Load has occurred since the last read of the Status Register, if WAVE = 0.
LDRBS: RB Loading Status
0 = RB Load has not occurred since the last read of the Status Register or WAVE = 1.
1 = RB Load has occurred since the last read of the Status Register, if WAVE = 0.
ETRGS: External Trigger Status
0 = External trigger has not occurred since the last read of the Status Register.
1 = External trigger has occurred since the last read of the Status Register.
CLKSTA: Clock Enabling Status
0 = Clock is disabled.
1 = Clock is enabled.
MTIOA: TIOA Mirror
0 = TIOA is low. If WAVE = 0, this means that TIOA pin is low. If WAVE = 1, this means that TIOA is driven low.
1 = TIOA is high. If WAVE = 0, this means that TIOA pin is high. If WAVE = 1, this means that TIOA is driven high.
MTIOB: TIOB Mirror
0 = TIOB is low. If WAVE = 0, this means that TIOB pin is low. If WAVE = 1, this means that TIOB is driven low.
1 = TIOB is high. If WAVE = 0, this means that TIOB pin is high. If WAVE = 1, this means that TIOB is driven high.
497
1768BATARM08/03
TC Interrupt Enable Register
Register Name: TC_IER
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
ETRGS LDRBS LDRAS CPCS CPBS CPAS LOVRS COVFS
498 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
ETRGS LDRBS LDRAS CPCS CPBS CPAS LOVRS COVFS
499
1768BATARM08/03
TC Interrupt Mask Register
Register Name: TC_IMR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
ETRGS LDRBS LDRAS CPCS CPBS CPAS LOVRS COVFS
500 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The MultiMedia Card Interface (MCI) supports the MultiMediaCard (MMC) Specification V2.2
and the SD Memory Card Specification V1.0.
The MCI includes a command register, response registers, data registers, timeout counters
and error detection logic that automatically handle the transmission of commands and, when
required, the reception of the associated responses and data with limited processor overhead.
The MCI supports stream, block and multi-block data read and write, and is compatible with
the Peripheral Data Controller channels, minimizing processor intervention for large buffer
transfers.
The MCI operates at a rate of up to Master Clock divided by 2 and supports interfacing of up to
16 slots (depending on the product). Each slot may be used to interface with a MultiMediaCard
bus (up to 30 Cards) or with an SD Memory Card. Only one slot can be selected at a time
(slots are multiplexed). A bit in the Command Register performs this selection.
The SD Memory Card communication is based on a 9-pin interface (clock, command, four
data and three power lines) and the MultiMediaCard on a 7-pin interface (clock, command,
one data and three power lines).
The SD Memory Card interface also supports MultiMedia Card operations. The main differ-
ences between SD and MultiMedia Cards are the initialization process and the bus topology.
The main features of the MCI are:
Compatibility with MultiMedia Card Specification Version 2.2
Compatibility with SD Memory Card Specification Version 1.0
Cards clock rate up to Master Clock divided by 2
Embedded power management to slow down clock rate when not used
Supports up to sixteen multiplexed slots (product-dependent)
One slot for one MultiMediaCard bus (up to 30 cards) or one SD Memory Card
Support for stream, block and multi-block data read and write
Supports connection to Peripheral Data Controller
Minimizes processor intervention for large buffer transfers
501
1768BATARM08/03
Block Diagram
Figure 233. Block Diagram
ASB
APB Bridge
PDC
APB
MCCK
MCCDA
MCDA0
PMC MCK
MCDA1
MCDA2
MCCDB
MCDB0
MCDB1
MCDB2
Interrupt Control
MCDB3
MCI Interrupt
502 AT91RM9200
1768BATARM08/03
AT91RM9200
Application Layer
ex: File System, Audio, Security, etc.
Physical Layer
MCI Interface
1 2 3 4 5 6 78
1234567 9
SDCard
MMC
503
1768BATARM08/03
Product Dependencies
I/O Lines The pins used for interfacing the MultiMedia Cards or SD Cards may be multiplexed with PIO
lines. The programmer must first program the PIO controllers to assign the peripheral func-
tions to MCI pins.
Power The MCI may be clocked through the Power Management Controller (PMC), so the program-
Management mer must first to configure the PMC to enable the MCI clock.
Interrupt The MCI interface has an interrupt line connected to the Advanced Interrupt Controller (AIC).
Handling the MCI interrupt requires programming the AIC before configuring the MCI.
1234567
MMC
The MultiMedia Card communication is based on a 7-pin serial bus interface. It has three com-
munication lines and four supply lines.
MCDA0
MCCK
504 AT91RM9200
1768BATARM08/03
AT91RM9200
1 2 3 4 5 6 78
9
SD CARD
The SD Memory Card bus includes the signals listed in Table 95.
MCDA0 - MCDA3
MCCK SD CARD 1
MCCDA
9
1 2 3 4 5 6 78
MCDB0 - MCDB3
SD CARD 2
MCCDB
9
505
1768BATARM08/03
Figure 239. Mixing MultiMedia and SD Memory Cards
MCDA0
MCCDA
MCCK
1 2 3 4 5 6 78
MCDB0 - MCDB3
SD CARD
MCCDB
9
When the MCI is configured to operate with SD memory cards, the width of the data bus can
be selected in the MCI_SDCR register. Clearing the SDCBUS bit in this register means that
the width is one bit and setting it means that the width is four bits. In the case of multimedia
cards, only the data line 0 is used. The other data lines can be used as independent PIOs.
MultiMedia Card After a power-on reset, the cards are initialized by a special message-based MultiMedia Card
Operations bus protocol. Each message is represented by one of the following tokens:
Command: A command is a token that starts an operation. A command is sent from the
host either to a single card (addressed command) or to all connected cards (broadcast
command). A command is transferred serially on the CMD line.
Response: A response is a token which is sent from an addressed card or (synchronously)
from all connected cards to the host as an answer to a previously received command. A
response is transferred serially on the CMD line.
Data: Data can be transferred from the card to the host or vice versa. Data is transferred
via the data line.
Card addressing is implemented using a session address assigned during the initialization
phase by the bus controller to all currently connected cards. Their unique CID number identi-
fies individual cards.
The structure of commands, responses and data blocks is described in the MultiMedia-Card
System Specification Version 2.2. See also Table 96 on page 507.
MultiMediaCard bus data transfers are composed of these tokens.
There are different types of operations. Addressed operations always contain a command and
a response token. In addition, some operations have a data token; the others transfer their
information directly within the command or response structure. In this case, no data token is
present in an operation. The bits on the DAT and the CMD lines are transferred synchronous
to the clock MCCK.
Two types of data transfer commands are defined:
Sequential commands: These commands initiate a continuous data stream. They are
terminated only when a stop command follows on the CMD line. This mode reduces the
command overhead to an absolute minimum.
Block-oriented commands: These commands send a data block succeeded by CRC bits.
506 AT91RM9200
1768BATARM08/03
AT91RM9200
Both read and write operations allow either single or multiple block transmission. A multiple
block transmission is terminated when a stop command follows on the CMD line similarly to
the sequential read.
The MCI provides a set of registers to perform the entire range of MultiMediaCard operations.
Command- After reset the MCI is disabled and becomes valid after setting the MCIEN bit in the MCI_CR
response Control Register. The bit PWSEN allows saving power by dividing the MCI clock by 2 power
Operation PWSDIV (MCI_MR) when the bus is inactive.
The command and the response of the card are clocked out with the rising edge of the MCCK.
All the timings for MultiMediaCard are defined in the MultiMediaCard System Specification
Version 2.2.
The two bus modes (open drain and push/pull) needed to process all the operations are
defined in the MCI command register. The MCI_CMDR allows a command to be carried out.
For example, to perform an ALL_SEND_CID command:
The command ALL_SEND_CID and the fields and values for the MCI_CMDR Control Register
are described in Table 96 and Table 97.
507
1768BATARM08/03
command requires a response, it can be read in the MCI response register (MCI_RSPR). The
response size can be 48 bits up to 136 bits according to the command. The MCI embeds an
error detection to prevent any corrupted data during the transfer.
The following flowchart shows how to send a command to the card and read the response if
needed. In this example, the status register bits are polled but setting the appropriate bits in
the interrupt enable register (MCI_IER) allows using an interrupt method.
Read MCI_SR
RETURN ERROR
RETURN OK
Note: 1. If the command is SEND_OP_COND, the CRC error flag is always present (refer to R3
response in the MultiMediaCard specification).
Data Transfer The MultiMedia Card allows several read/write operations (single block, multiple blocks,
Operation stream, etc.).
These operations can be done using the Peripheral Data Controller (PDC) features. If the
PDCMODE bit is set in MCI_MR, then all reads and writes use the PDC facilities. In all cases,
the block length must be defined in the mode register.
508 AT91RM9200
1768BATARM08/03
AT91RM9200
Read Operation The following flowchart shows how to read a single block with or without use of PDC facilities.
In this example, a polling method is used to wait for the end of read. Similarly, the user can
configure the interrupt enable register (MCI_IER) to trigger an interrupt at the end of read.
These two methods can be applied for all MultiMediaCard read functions.
No Yes
Read with PDC
No
No
Read data = MCI_RDR
RETURN RETURN
509
1768BATARM08/03
Write Operation In write operation the MCI Mode Register (MCI_MR) is used to define the padding value when
writing non-multiple block size. If the bit PDCPADV is 0, then 0x00 value is used when pad-
ding data, otherwise 0xFF is used. If set, the bit PDCMODE enables PDC transfer.
The following flowchart shows how to write a single block with or without use of PDC facilities.
Polling or interrupt method can be used to wait for the end of write according to the contents of
the Interrupt Mask Register (MCI_IMR).
This flowchart can be adapted to perform all the MultiMedia Card write functions.
No Yes
Write using PDC
Yes
Number words write = Send command
BlockLength/4 WRITE_SINGLE_BLOCK
No
No
No
MCI_TDR = Data to write
RETURN RETURN
510 AT91RM9200
1768BATARM08/03
AT91RM9200
SD Card The MultiMedia Card Interface allows processing of SD Memory Card (Secure Digital Memory
Operations Card) commands. The SD Memory Card will include a copyright protection mechanism that
complies with the security requirements of the SDMI standard, is faster and applicable to
higher memory capacity.
The physical form factor, pin assignment and data transfer protocol are forward-com-patible
with the MultiMedia Card with some additions.
The SD Memory Card communication is based on a 9-pin interface (Clock, Command,
4 x Data and 3 x Power lines). The communication protocol is defined as a part of this specifi-
cation. The main difference between the SD Memory Card and the MultiMedia Card is the
initialization process.
The SD Card Control Register (MCI_SDCR) allows selection of the card slot and the data bus
width.
The SD Card bus allows dynamic configuration of the number of data lines. After power up, by
default, the SD Memory Card will use only DAT0 for data transfer. After initialization, the host
can change the bus width (number of active data lines).
511
1768BATARM08/03
MultiMedia Card (MCI) User Interface
Table 98. MCI Register Mapping
Offset Register Register Name Read/Write Reset
0x00 Control Register MCI_CR Write ---
0x04 Mode Register MCI_MR Read/write 0x0
0x08 Data Timeout Register MCI_DTOR Read/write 0x0
0x0C SD Card Register MCI_SDCR Read/write 0x0
0x10 Argument Register MCI_ARGR Read/write 0x0
0x14 Command Register MCI_CMDR Write ---
0x18 - 0x1C Reserved
0x20 Response Register(1) MCI_RSPR Read 0x0
0x24 Response Register(1) MCI_RSPR Read 0x0
(1)
0x28 Response Register MCI_RSPR Read 0x0
(1)
0x2C Response Register MCI_RSPR Read 0x0
0x30 Receive Data Register MCI_RDR Read 0x0
0x34 Transmit Data Register MCI_TDR Write ---
0x38 - 0x3C Reserved
0x40 Status Register MCI_SR Read 0xC0E5
0x44 Interrupt Enable Register MCI_IER Write ---
0x48 Interrupt Disable Register MCI_IDR Write ---
0x4C Interrupt Mask Register MCI_IMR Read 0x0
0x50-0xFF Reserved
0x100-0x124 Reserved for the PDC
Note: 1. The response register can be read by N accesses at the same MCI_RSPR or at consecutive addresses (0x20 to 0x2C).
N depends on the size of the response.
512 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
PWSDIS PWSEN MCIDIS MCIEN
513
1768BATARM08/03
MCI Mode Register
Name: MCI_MR
Access Type: Read/write
31 30 29 28 27 26 25 24
BLKLEN
23 22 21 20 19 18 17 16
BLKLEN 0 0
15 14 13 12 11 10 9 8
PDCMODE PDCPADV PWSDIV
7 6 5 4 3 2 1 0
CLKDIV
514 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
DTOMUL DTOCYC
DTOMUL Multiplier
0 0 0 1
0 0 1 16
0 1 0 128
0 1 1 256
1 0 0 1024
1 0 1 4096
1 1 0 65536
1 1 1 1048576
515
1768BATARM08/03
MCI SD Card Register
Name: MCI_SDCR
Access Type: Read/write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
SDCBUS SDCSEL
23 22 21 20 19 18 17 16
ARG
15 14 13 12 11 10 9 8
ARG
7 6 5 4 3 2 1 0
ARG
516 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
TRTYPE TRDIR TRCMD
15 14 13 12 11 10 9 8
MAXLAT OPDCMD SPCMD
7 6 5 4 3 2 1 0
RSPTYP CMDNB
This register is write-protected while CMDRDY is 0 in MCI_SR and in the case of a no Interrupt command sent (bit
SPCMD). This means that the current command execution cannot be interrupted or modified.
SPCMD CMD
0 0 0 Not a special CMD.
0 0 1 Initialization CMD:
74 clock cycles for initialization sequence.
0 1 0 Synchronized CMD:
Wait for the end of the current data block transfer before sending the pending
command.
0 1 1 Reserved.
1 0 0 Interrupt command:
Corresponds to the Interrupt Mode (CMD40).
1 0 1 Interrupt response:
Corresponds to the Interrupt Mode (CMD40).
517
1768BATARM08/03
MAXLAT: Max Latency for Command to Response
0 =5-cycle max latency
1 = 64-cycle max latency
TRCMD: Transfer Command
23 22 21 20 19 18 17 16
RSP
15 14 13 12 11 10 9 8
RSP
7 6 5 4 3 2 1 0
RSP
RSP: Response
518 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
DATA
15 14 13 12 11 10 9 8
DATA
7 6 5 4 3 2 1 0
DATA
23 22 21 20 19 18 17 16
DATA
15 14 13 12 11 10 9 8
DATA
7 6 5 4 3 2 1 0
DATA
519
1768BATARM08/03
MCI Status Register
Name: MCI_SR
Access Type: Read-only
31 30 29 28 27 26 25 24
UNRE OVRE
23 22 21 20 19 18 17 16
DTOE TCRCE RTOE RENDE RCRCE RDIRE RINDE
15 14 13 12 11 10 9 8
TXBUFE RXBUFF
7 6 5 4 3 2 1 0
ENDTX ENDRX NOTBUSY DTIP BLKE TXRDY RXRDY CMDRDY
520 AT91RM9200
1768BATARM08/03
AT91RM9200
521
1768BATARM08/03
MCI Interrupt Enable Register
Name: MCI_IER
Access Type: Write-only
31 30 29 28 27 26 25 24
UNRE OVRE
23 22 21 20 19 18 17 16
DTOE TCRCE RTOE RENDE RCRCE RDIRE RINDE
15 14 13 12 11 10 9 8
TXBUFE RXBUFF
7 6 5 4 3 2 1 0
ENDTX ENDRX NOTBUSY DTIP BLKE TXRDY RXRDY CMDRDY
522 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
DTOE TCRCE RTOE RENDE RCRCE RDIRE RINDE
15 14 13 12 11 10 9 8
TXBUFE RXBUFF
7 6 5 4 3 2 1 0
ENDTX ENDRX NOTBUSY DTIP BLKE TXRDY RXRDY CMDRDY
523
1768BATARM08/03
MCI Interrupt Mask Register
Name: MCI_IMR
Access Type: Read-only
31 30 29 28 27 26 25 24
UNRE OVRE
23 22 21 20 19 18 17 16
DTOE TCRCE RTOE RENDE RCRCE RDIRE RINDE
15 14 13 12 11 10 9 8
TXBUFE RXBUFF
7 6 5 4 3 2 1 0
ENDTX ENDRX NOTBUSY DTIP BLKE TXRDY RXRDY CMDRDY
524 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The USB Device Port (UDP) is compliant with the Universal Serial Bus (USB) V2.0 full-speed
device specification. It is designed to be associated with Atmels embedded USB transceiver
and interfaced with an ARM7TDMI and ARM9TDMI core.
The number and size of endpoints is product-dependent. Each endpoint is associated with
one or two banks of a dual-port RAM used to store the current data payload. If two banks are
used, one DPR bank is read or written by the processor, while the other is read or written by
the USB device peripheral. This feature is mandatory for isochronous endpoints. Thus the
device maintains the maximum bandwidth (1M bytes/s) by working with endpoints with two
banks of DPR.
Suspend and resume are automatically detected by the USB device, which notifies the pro-
cessor by raising an interrupt. Depending on the product, an external signal can be used to
send a wake-up to the USB host controller.
The main features of the UDP are:
USB V2.0 Full-speed Compliant, 12 Mbits per second
Embedded USB V2.0 Full-speed Transceiver
Number and Size of Endpoints Fully Parametrizable in RTL
Embedded Dual-port RAM for Endpoints
Suspend/Resume Logic
Ping-pong Mode (2 Memory Banks) for Isochronous and Bulk Endpoints
525
1768BATARM08/03
Block Diagram
Figure 243. USB Device Port Block Diagram
Atmel Bridge
USB Device
APB
to
MCU txoen
Bus eopn
MCK U W W
s r Dual r Serial DP
txd
UDPCK e a Port a Interface Embedded
r p RAM p Engine USB
p p rxdm DM
Transceiver
I e e rxd
n FIFO 12 MHz SIE
r r
t rxdp
e
r
f
udp_int a
c
e Suspend/Resume Logic
Access to the UDP is via the APB bus interface. Read and write to the data FIFO are done by
reading and writing 8-bit values to APB registers.
The UDP peripheral requires two clocks: one peripheral clock used by the MCK domain and a
48 MHz clock used by the 12 MHz domain.
A USB 2.0 full-speed pad is embedded and controlled by the SIE.
The signal external_resume is optional. It allows the UDP peripheral to wake-up once in sys-
tem mode. The host will then be notified that the device asks for a resume. This optional
feature must be also negotiated with the host during the enumeration.
526 AT91RM9200
1768BATARM08/03
AT91RM9200
Product Dependencies
The USB physical transceiver is integrated into the product. The bi-directional differential sig-
nals DP and DM are available from the product boundary.
Two I/O lines may be used by the application:
One to check that VBUS is still available from the host. Self-powered devices may use this
entry to be notified that the host has been powered off. In this case, the board pull-up on
DP must be disabled in order to prevent feeding current to the host.
One to control the board pull-up on DP. Thus, when the device is ready to communicate
with the host, it activates its DP pull-up through this control line.
I/O Lines DP and DM are not controlled by any PIO controllers. The embedded USB physical trans-
ceiver is controlled by the USB device peripheral.
To reserve an I/O line to check VBUS, the programmer must first program the PIO controller to
assign this I/O in input PIO mode.
To reserve an I/O line to control the board pull-up, the programmer must first program the PIO
controller to assign this I/O in output PIO mode.
Power The USB device peripheral requires a 48 MHz clock. This clock must be generated by a PLL
Management with an accuracy of 0.25%.
Thus, the USB device receives two clocks from the Power Management Controller (PMC): the
master clock, MCK, used to drive the peripheral user interface and the UDPCK used to inter-
face with the bus USB signals (recovered 12 MHz domain).
Interrupt The USB device interface has an interrupt line connected to the Advanced Interrupt Controller
(AIC).
Handling the USB device interrupt requires programming the AIC before configuring the UDP.
527
1768BATARM08/03
Typical Connection
Figure 244. Board Schematic to Interface USB Device Peripheral
15K
PAm
22K
3V3
15k
47k
PAn
System
Reset 33pF
27
DDM
100nF
DDP
27 Type B
15pF Connector
15pF
528 AT91RM9200
1768BATARM08/03
AT91RM9200
Functional Description
USB V2.0 Full- The USB V2.0 full-speed provides communication services between host and attached USB
speed Introduction devices. Each device is offered with a collection of communication flows (pipes) associated
with each endpoint. Software on the host communicates with an USB device through a set of
communication flows.
529
1768BATARM08/03
USB Bus Transactions Each transfer results in one or more transactions over the USB bus. There are five kinds of
transactions flowing across the bus in packets:
1. Setup Transaction
2. Data IN Transaction
3. Data OUT Transaction
4. Status IN Transaction
5. Status OUT Transaction
USB Transfer Event As shown in Table 100, transfers are sequential events carried out on the USB bus.
Definitions
Table 100. USB Transfer Events
Control Transfers(1) (3) Setup transaction > Data IN transactions >
Status OUT transaction
Setup transaction > Data OUT transactions >
Status IN transaction
Setup transaction > Status IN transaction
Interrupt IN Transfer Data IN transaction > Data IN transaction
(device toward host)
Interrupt OUT Transfer Data OUT transaction > Data OUT transaction
(host toward device)
Isochronous IN Transfer(2) Data IN transaction > Data IN transaction
(device toward host)
Isochronous OUT Transfer(2) Data OUT transaction > Data OUT transaction
(host toward device)
Bulk IN Transfer Data IN transaction > Data IN transaction
(device toward host)
Bulk OUT Transfer Data OUT transaction > Data OUT transaction
(host toward device)
Notes: 1. Control transfer must use endpoints with no ping-pong attributes.
2. Isochronous transfers must use endpoints with ping-pong attributes.
3. Control transfers can be aborted using a stall handshake.
530 AT91RM9200
1768BATARM08/03
AT91RM9200
Setup Transaction Setup is a special type of host-to-device transaction used during control transfers. Control
transfers must be performed using endpoints with no ping-pong attributes. A setup transaction
needs to be handled as soon as possible by the firmware. It is used to transmit requests from
the host to the device. These requests are then handled by the USB device and may require
more arguments. The arguments are sent to the device by a Data OUT transaction which fol-
lows the setup transaction. These requests may also return data. The data is carried out to the
host by the next Data IN transaction which follows the setup transaction. A status transaction
ends the control transfer.
When a setup transfer is received by the USB endpoint:
The USB device automatically acknowledges the setup packet
RXSETUP is set in the USB_CSRx register
An endpoint interrupt is generated while the RXSETUP is not cleared. This interrupt is
carried out to the microcontroller if interrupts are enabled for this endpoint.
Thus, firmware must detect the RXSETUP polling the USB_CSRx or catching an interrupt,
read the setup packet in the FIFO, then clear the RXSETUP. RXSETUP cannot be cleared
before the setup packet has been read in the FIFO. Otherwise, the USB device would accept
the next Data OUT transfer and overwrite the setup packet in the FIFO.
USB Setup Data Setup ACK Data OUT Data OUT NAK Data OUT Data OUT ACK
Bus Packets PID PID PID PID PID PID
531
1768BATARM08/03
Data IN Transaction Data IN transactions are used in control, isochronous, bulk and interrupt transfers and conduct
the transfer of data from the device to the host. Data IN transactions in isochronous transfer
must be done using endpoints with ping-pong attributes.
Prevous Data IN TX Microcontroller Load Data in FIFO Data is Sent on USB Bus
USB Bus Packets Data IN Data IN 1 ACK Data IN NAK Data IN Data IN 2 ACK
PID PID PID PID PID PID
TXPKTRDY Flag
(USB_CSRx)
Set by the Firmware
Cleared by USB Device Data Payload Written in FIFO
Interrupt Pending Interrupt Pending
TXCOMP Flag Start to Write Data
Cleared by Firmware
(USB_CSRx) Payload in FIFO
532 AT91RM9200
1768BATARM08/03
AT91RM9200
Using Endpoints With The use of an endpoint with ping-pong attributes is necessary during isochronous transfer. To
Ping-pong Attribute be able to guarantee a constant bandwidth, the microcontroller must prepare the next data
payload to be sent while the current one is being sent by the USB device. Thus two banks of
memory are used. While one is available for the microcontroller, the other one is locked by the
USB device.
Write Read
1st Data Payload
Bank 0
Endpoint 1
Read and Write at the Same Time
Data IN Packet
Bank 0
Endpoint 1 3rd Data Payload
When using a ping-pong endpoint, the following procedures are required to perform Data IN
transactions:
1. The microcontroller checks if it is possible to write in the FIFO by polling TXPKTRDY to
be cleared in the endpoints USB_CSRx register.
2. The microcontroller writes the first data payload to be sent in the FIFO (Bank 0), writing
zero or more byte values in the endpoints USB_FDRx register.
3. The microcontroller notifies the USB peripheral it has finished writing in Bank 0 of the
FIFO by setting the TXPKTRDY in the endpoints USB_CSRx register.
4. Without waiting for TXPKTRDY to be cleared, the microcontroller writes the second
data payload to be sent in the FIFO (Bank 1), writing zero or more byte values in the
endpoints USB_FDRx register.
5. The microcontroller is notified that the first Bank has been released by the USB device
when TXCOMP in the endpoints USB_CSRx register is set. An interrupt is pending
while TXCOMP is being set.
6. Once the microcontroller has received TXCOMP for the first Bank, it notifies the USB
device that it has prepared the second Bank to be sent rising TXPKTRDY in the end-
points USB_CSRx register.
7. At this step, Bank 0 is available and the microcontroller can prepare a third data pay-
load to be sent.
533
1768BATARM08/03
Figure 249. Data IN Transfer for Ping-pong Endpoint
TXPKTRDY Flag
(USB_MCSRx) Cleared by USB Device,
Data Payload Fully Transmitted Set by Firmware,
Set by Firmware, Data Payload Written in FIFO Bank 1
Data Payload Written in FIFO Bank 0
Interrupt Pending
Set by USB
TXCOMP Flag Device Set by USB Device
(USB_CSRx)
Interrupt Cleared by Firmware
Warning: There is software critical path due to the fact that once the second bank is filled, the
driver has to wait for TX_COMP to set TX_PKTRDY. If the delay between receiving
TX_COMP is set and TX_PKTRDY is set is too long, some Data IN packets may be NACKed,
reducing the bandwidth.
534 AT91RM9200
1768BATARM08/03
AT91RM9200
Data OUT Transaction Data OUT transactions are used in control, isochronous, bulk and interrupt transfers and con-
duct the transfer of data from the host to the device. Data OUT transactions in isochronous
transfers must be done using endpoints with ping-pong attributes.
Data OUT Transaction To perform a Data OUT transaction, using a non ping-pong endpoint:
Without Ping-pong 1. The host generates a Data OUT packet.
Attributes
2. This packet is received by the USB device endpoint. While the FIFO associated to this
endpoint is being used by the microcontroller, a NAK PID is returned to the host. Once
the FIFO is available, data are written to the FIFO by the USB device and an ACK is
automatically carried out to the host.
3. The microcontroller is notified that the USB device has received a data payload polling
RX_DATA_BK0 in the endpoints USB_CSRx register. An interrupt is pending for this
endpoint while RX_DATA_BK0 is set.
4. The number of bytes available in the FIFO is made available by reading RXBYTECNT
in the endpoints USB_CSRx register.
5. The microcontroller carries out data received from the endpoints memory to its mem-
ory. Data received is available by reading the endpoints USB_FDRx register.
6. The microcontroller notifies the USB device that it has finished the transfer by clearing
RX_DATA_BK0 in the endpoints USB_CSRx register.
7. A new Data OUT packet can be accepted by the USB device.
USB Bus Data OUT ACK Data OUT2 Data OUT2 NAK Data OUT Data OUT2 ACK
PID Data OUT 1 PID PID PID PID PID
Packets
FIFO (DPR)
Data OUT 1 Data OUT 1 Data OUT 2
Content
Written by USB Device Microcontroller Read Written by USB Device
An interrupt is pending while the flag RX_DATA_BK0 is set. Memory transfer between the
USB device, the FIFO and microcontroller memory can not be done after RX_DATA_BK0 has
been cleared. Otherwise, the USB device would accept the next Data OUT transfer and over-
write the current Data OUT packet in the FIFO.
535
1768BATARM08/03
Using Endpoints With During isochronous transfer, using an endpoint with ping-pong attributes is necessary. To be
Ping-pong Attributes able to guarantee a constant bandwidth, the microcontroller must read the previous data pay-
load sent by the host, while the current data payload is received by the USB device. Thus two
banks of memory are used. While one is available for the microcontroller, the other one is
locked by the USB device.
Figure 251. Bank Swapping in Data OUT Transfers for Ping-pong Endpoints
Write Read
When using a ping-pong endpoint, the following procedures are required to perform Data OUT
transactions:
1. The host generates a Data OUT packet.
2. This packet is received by the USB device endpoint. It is written in the endpoints FIFO
Bank 0.
3. The USB device sends an ACK PID packet to the host. The host can immediately send
a second Data OUT packet. It is accepted by the device and copied to FIFO Bank 1.
4. The microcontroller is notified that the USB device has received a data payload, polling
RX_DATA_BK0 in the endpoints USB_CSRx register. An interrupt is pending for this
endpoint while RX_DATA_BK0 is set.
5. The number of bytes available in the FIFO is made available by reading RXBYTECNT
in the endpoints USB_CSRx register.
6. The microcontroller transfers out data received from the endpoints memory to the
microcontrollers memory. Data received is made available by reading the endpoints
USB_FDRx register.
7. The microcontroller notifies the USB peripheral device that it has finished the transfer
by clearing RX_DATA_BK0 in the endpoints USB_CSRx register.
8. A third Data OUT packet can be accepted by the USB peripheral device and copied in
the FIFO Bank 0.
9. If a second Data OUT packet has been received, the microcontroller is notified by the
flag RX_DATA_BK1 set in the endpoints USB_CSRx register. An interrupt is pending
for this endpoint while RX_DATA_BK1 is set.
536 AT91RM9200
1768BATARM08/03
AT91RM9200
10. The microcontroller transfers out data received from the endpoints memory to the
microcontrollers memory. Data received is available by reading the endpoints
USB_FDRx register.
11. The microcontroller notifies the USB device it has finished the transfer by clearing
RX_DATA_BK1 in the endpoints USB_CSRx register.
12. A fourth Data OUT packet can be accepted by the USB device and copied in the FIFO
Bank 0.
USB Bus Data OUT ACK Data OUT ACK Data OUT
Data OUT 1 PID Data OUT 2 Data OUT 3
Packets PID PID PID PID
A
P
FIFO (DPR)
Bank 0 Data OUT1 Data OUT 1 Data OUT 3
Write by USB Device Read By Microcontroller Write In Progress
FIFO (DPR)
Bank 1 Data OUT 2 Data OUT 2
Write by USB Device Read By Microcontroller
537
1768BATARM08/03
Status Transaction A status transaction is a special type of host to device transaction used only in a control trans-
fer. The control transfer must be performed using endpoints with no ping-pong attributes.
According to the control sequence (read or write), the USB device sends or receives a status
transaction.
No Data
Control Setup TX Status IN TX
Notes: 1. During the Status IN stage, the host waits for a zero length packet (Data IN transaction with
no data) from the device using DATA1 PID. Please refer to Chapter 8 of the Universal Serial
Bus Specification, Rev. 1.1, to get more information on the protocol layer.
2. During the Status OUT stage, the host emits a zero length packet to the device (Data OUT
transaction with no data).
538 AT91RM9200
1768BATARM08/03
AT91RM9200
Status IN Transfer Once a control request has been processed, the device returns a status to the host. This is a
zero length Data IN transaction.
1. The microcontroller waits for TXPKTRDY in the USB_CSRx endpoints register to be
cleared. (At this step, TXPKTRDY must be cleared because the previous transaction
was a setup transaction or a Data OUT transaction.)
2. Without writing anything to the USB_FDRx endpoints register, the microcontroller sets
TXPKTRDY. The USB device generates a Data IN packet using DATA1 PID.
3. This packet is acknowledged by the host and TXPKTRDY is set in the USB_CSRx end-
points register.
Interrupt Pending
539
1768BATARM08/03
Status OUT Transfer Once a control request has been processed and the requested data returned, the host
acknowledges by sending a zero length packet. This is a zero length Data OUT transaction.
1. The USB device receives a zero length packet. It sets RX_DATA_BK0 flag in the
USB_CSRx register and acknowledges the zero length packet.
2. The microcontroller is notified that the USB device has received a zero length packet
sent by the host polling RX_DATA_BK0 in the USB_CSRx register. An interrupt is
pending while RX_DATA_BK0 is set. The number of bytes received in the endpoints
USB_BCR register is equal to zero.
3. The microcontroller must clear RX_DATA_BK0.
Interrupt Pending
TXCOMP
(USB_CSRx)
540 AT91RM9200
1768BATARM08/03
AT91RM9200
Stall Handshake A stall handshake can be used in one of two distinct occasions. (For more information on the
stall handshake, refer to Chapter 8 of the Universal Serial Bus Specification, Rev 1.1.)
A functional stall is used when the halt feature associated with the endpoint is set. (Refer
to Chapter 9 of the Universal Serial Bus Specification, Rev 1.1, for more information on the
halt feature.)
To abort the current request, a protocol stall is used, but uniquely with control transfer.
The following procedure generates a stall packet:
1. The microcontroller sets the FORCESTALL flag in the USB_CSRx endpoints register.
2. The host receives the stall packet.
3. The microcontroller is notified that the device has sent the stall by polling the
STALLSENT to be set. An endpoint interrupt is pending while STALLSENT is set. The
microcontroller must clear STALLSENT to clear the interrupt.
When a setup transaction is received after a stall handshake, STALLSENT must be cleared in
order to prevent interrupts due to STALLSENT being set.
Cleared by Firmware
FORCESTALL Set by Firmware
Interrupt Pending
Cleared by Firmware
STALLSENT
Set by
USB Device
Interrupt Pending
541
1768BATARM08/03
Controlling Device A USB device has several possible states. Please refer to Chapter 9 of the Universal Serial
States Bus Specification, Rev 1.1.
Attached
Hub Reset
Hub
or
Configured
Deconfigured
Bus Inactive
Powered Suspended
Bus Activity
Power
Interruption Reset
Bus Inactive
Suspended
Default
Bus Activity
Reset
Address
Assigned
Bus Inactive
Address Suspended
Bus Activity
Device Device
Deconfigured Configured
Bus Inactive
Configured
Suspended
Bus Activity
Movement from one state to another depends on the USB bus state or on standard requests
sent through control transactions via the default endpoint (endpoint 0).
After a period of bus inactivity, the UDP device enters Suspend Mode. Accepting Sus-
pend/Resume requests from the USB host is mandatory. Constraints in Suspend Mode are
very strict for bus-powered applications; devices may not consume more than 500 uA on the
USB bus.
While in Suspend Mode, the host may wake up a device by sending a resume signal (bus
activity) or a USB device may send a wake-up request to the host, e.g., waking up a PC by
moving a USB mouse.
The wake-up feature is not mandatory for all devices and must be negotiated with the host.
542 AT91RM9200
1768BATARM08/03
AT91RM9200
From Powered State to After its connection to a USB host, the USB device waits for an end-of-bus reset. The USB
Default State host stops driving a reset state once it has detected the devices pull-up on DP. The unmasked
flag ENDBURST is set in the register UDP_ISR and an interrupt is triggered. The UDP soft-
ware enables the default endpoint, setting the EPEDS flag in the UDP_CSR[0] register and,
optionally, enabling the interrupt for endpoint 0 by writing 1 to the UDP_IER register. The enu-
meration then begins by a control transfer.
From Default State to After a set address standard device request, the USB host peripheral enters the address state.
Address State Before this, it achieves the Status IN transaction of the control transfer, i.e., the UDP device
sets its new address once the TXCOMP flag in the UDP_CSR[0] register has been received
and cleared.
To move to address state, the driver software sets the FADDEN flag in the
UDP_GLB_STATE, sets its new address, and sets the FEN bit in the UDP_FADDR register.
From Address State to Once a valid Set Configuration standard request has been received and acknowledged, the
Configured State device enables endpoints corresponding to the current configuration. This is done by setting
the EPEDS and EPTYPE fields in the UDP_CSRx registers and, optionally, enabling corre-
sponding interrupts in the UDP_IER register.
Enabling Suspend When a Suspend (no bus activity on the USB bus) is detected, the RXSUSP signal in the
UDP_ISR register is set. This triggers an interrupt if the corresponding bit is set in the
UDP_IMR register.
This flag is cleared by writing to the UDP_ICR register. Then the device enters Suspend
Mode. As an example, the microcontroller switches to slow clock, disables the PLL and main
oscillator, and goes into Idle Mode. It may also switch off other devices on the board.
The USB device peripheral clocks may be switched off. However, the transceiver and the USB
peripheral must not be switched off, otherwise the resume is not detected.
Receiving a Host In suspend mode, the USB transceiver and the USB peripheral must be powered to detect the
Resume RESUME. However, the USB device peripheral may not be clocked as the WAKEUP signal is
asynchronous.
Once the resume is detected on the bus, the signal WAKEUP in the UDP_ISR is set. It may
generate an interrupt if the corresponding bit in the UDP_IMR register is set. This interrupt
may be used to wake-up the core, enable PLL and main oscillators and configure clocks. The
WAKEUP bit must be cleared as soon as possible by setting WAKEUP in the UDP_ICR
register.
Sending an External The External Resume is negotiated with the host and enabled by setting the ESR bit in the
Resume USB_GLB_STATE. An asynchronous event on the ext_resume_pin of the peripheral gener-
ates a WAKEUP interrupt. On early versions of the USP peripheral, the K-state on the USB
line is generated immediately. This means that the USB device must be able to answer to the
host very quickly. On recent versions, the software sets the RMWUPE bit in the
UDP_GLB_STATE register once it is ready to communicate with the host. The K-state on the
bus is then generated.
The WAKEUP bit must be cleared as soon as possible by setting WAKEUP in the UDP_ICR
register.
543
1768BATARM08/03
USB Device Port (UDP) User Interface
Table 101. USB Device Port Memory Map
Offset Register Name Access Reset State
0x000 Frame Number Register USB_FRM_NUM Read 0x0000_0000
0x004 Global State Register USB_GLB_STAT Read/write 0x0000_0010
0x008 Function Address Register USB_FADDR Read/write 0x0000_0100
0x00C Reserved
0x010 Interrupt Enable Register USB_IER Write
0x014 Interrupt Disable Register USB_IDR Write
0x018 Interrupt Mask Register USB_IMR Read 0x0000_1200
0x01C Interrupt Status Register USB_ISR Read 0x0000_0000
0x020 Interrupt Clear Register USB_ICR Write
0x024 Reserved
0x028 Reset Endpoint Register USB_RST_EP Read/write
0x02C Reserved
0x030 Endpoint 0 Control and Status Register USB _CSR0 Read/write 0x0000_0000
0x034 Endpoint 1 Control and Status Register USB _CSR1 Read/write 0x0000_0000
0x038 Endpoint 2 Control and Status Register USB _CSR2 Read/write 0x0000_0000
0x03C Endpoint 3 Control and Status Register USB _CSR3 Read/write 0x0000_0000
0x040 Endpoint 4 Control and Status Register USB _CSR4 Read/write 0x0000_0000
0x044 Endpoint 5 Control and Status Register USB _CSR5 Read/write 0x0000_0000
0x048 Endpoint 6 Control and Status Register USB _CSR6 Read/write 0x0000_0000
0x04C Endpoint 7 Control and Status Register USB _CSR7 Read/write 0x0000_0000
0x050 Endpoint 0 FIFO Data Register USB_FDR0 Read/write 0x0000_0000
0x054 Endpoint 1 FIFO Data Register USB_FDR1 Read/write 0x0000_0000
0x058 Endpoint 2 FIFO Data Register USB_FDR2 Read/write 0x0000_0000
0x05C Endpoint 3 FIFO Data Register USB_FDR3 Read/write 0x0000_0000
0x060 Endpoint 4 FIFO Data Register USB_FDR4 Read/write 0x0000_0000
0x064 Endpoint 5 FIFO Data Register USB_FDR5 Read/write 0x0000_0000
0x068 Endpoint 6 FIFO Data Register USB_FDR6 Read/write 0x0000_0000
0x06C Endpoint 7 FIFO Data Register USB_FDR7 Read/write 0x0000_0000
0x070 Reserved
0x074 Reserved
544 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
FRM_OK FRM_ERR
15 14 13 12 11 10 9 8
FRM_NUM
7 6 5 4 3 2 1 0
FRM_NUM
545
1768BATARM08/03
USB Global State Register
Register Name: USB_GLB_STAT
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
RMWUPE RSMINPR ESR CONFG FADDEN
This register is used to get and set the device state as specified in Chapter 9 of the USB Serial Bus Specification, Rev.1.1.
FADDEN: Function Address Enable
Read:
0 = Device is not in address state.
1 = Device is in address state.
Write:
0 = No effect, only a reset can bring back a device to the default state.
1 = Set device in address state. This occurs after a successful Set Address request. Beforehand, the USB_FADDR register
must have been initialized with Set Address parameters. Set Address must complete the Status Stage before setting FAD-
DEN. Please refer to chapter 9 of the Universal Serial Bus Specification, Rev. 1.1 to get more details.
CONFG: Configured
Read:
0 = Device is not in configured state.
1 = Device is in configured state.
Write:
0 = Set device in a nonconfigured state
1 = Set device in configured state.
The device is set in configured state when it is in address state and receives a successful Set Configuration request.
Please refer to Chapter 9 of the Universal Serial Bus Specification, Rev. 1.1 to get more details.
ESR: Enable Send Resume
0 = Disable the Remote Wake Up sequence.
1 = Remote Wake Up can be processed and the pin send_resume is enabled.
RSMINPR: A Resume Has Been Sent to the Host
Read:
0 = No effect.
1 = A Resume has been received from the host during Remote Wake Up feature.
RMWUPE: Remote Wake Up Enable
0 = Must be cleared after receiving any HOST packet or SOF interrupt.
1 = Enables the K-state on the USB cable if ESR is enabled.
546 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
FEN
7 6 5 4 3 2 1 0
FADD
547
1768BATARM08/03
USB Interrupt Enable Register
Register Name: USB_IER
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
WAKEUP SOFINT EXTRSM RXRSM RXSUSP
7 6 5 4 3 2 1 0
EP7INT EP6INT EP5INT EP4INT EP3INT EP2INT EP1INT EP0INT
548 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
WAKEUP SOFINT EXTRSM RXRSM RXSUSP
7 6 5 4 3 2 1 0
EP7INT EP6INT EP5INT EP4INT EP3INT EP2INT EP1INT EP0INT
549
1768BATARM08/03
USB Interrupt Mask Register
Register Name: USB_IMR
Access Type: Read-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
WAKEUP SOFINT EXTRSM RXRSM RXSUSP
7 6 5 4 3 2 1 0
EP7INT EP6INT EP5INT EP4INT EP3INT EP2INT EP1INT EP0INT
550 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
WAKEUP ENDBUSRES SOFINT EXTRSM RXRSM RXSUSP
7 6 5 4 3 2 1 0
EP7INT EP6INT EP5INT EP4INT EP3INT EP2INT EP1INT EP0INT
551
1768BATARM08/03
EP3INT: Endpoint 3 Interrupt Status
0 = No Endpoint3 Interrupt pending.
1 = Endpoint3 Interrupt has been raised.
Several signals can generate this interrupt. The reason can be found by reading USB_CSR3:
RXSETUP set to 1
RX_DATA_BK0 set to 1
RX_DATA_BK1 set to 1
TXCOMP set to 1
STALLSENT set to 1
EP3INT is a sticky bit. Interrupt remains valid until EP3INT is cleared by writing in the corresponding USB_CSR3 bit.
EP4INT: Endpoint 4 Interrupt Status
0 = No Endpoint4 Interrupt pending.
1 = Endpoint4 Interrupt has been raised.
Several signals can generate this interrupt. The reason can be found by reading USB_CSR4:
RXSETUP set to 1
RX_DATA_BK0 set to 1
RX_DATA_BK1 set to 1
TXCOMP set to 1
STALLSENT set to 1
EP4INT is a sticky bit. Interrupt remains valid until EP4INT is cleared by writing in the corresponding USB_CSR4 bit.
EP5INT: Endpoint 5 Interrupt Status
0 = No Endpoint5 Interrupt pending.
1 = Endpoint5 Interrupt has been raised.
Several signals can generate this interrupt. The reason can be found by reading USB_CSR5:
RXSETUP set to 1
RX_DATA_BK0 set to 1
RX_DATA_BK1 set to 1
TXCOMP set to 1
STALLSENT set to 1
EP5INT is a sticky bit. Interrupt remains valid until EP5INT is cleared by writing in the corresponding USB_CSR5 bit.
EP6INT: Endpoint 6 Interrupt Status
0 = No Endpoint6 Interrupt pending.
1 = Endpoint6 Interrupt has been raised.
Several signals can generate this interrupt. The reason can be found by reading USB_CSR6:
RXSETUP set to 1
RX_DATA_BK0 set to 1
RX_DATA_BK1 set to 1
TXCOMP set to 1
STALLSENT set to 1
EP6INT is a sticky bit. Interrupt remains valid until EP6INT is cleared by writing in the corresponding USB_CSR6 bit.
552 AT91RM9200
1768BATARM08/03
AT91RM9200
553
1768BATARM08/03
USB Interrupt Clear Register
Register Name: USB_ICR
Access Type: Write-only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
WAKEUP ENDBURST SOFINT EXTRSM RXRSM RXSUSP
7 6 5 4 3 2 1 0
554 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
EP7 EP6 EP5 EP4 EP3 EP2 EP1 EP0
555
1768BATARM08/03
USB Endpoint Control and Status Register
Register Name: USB_CSRx [x = 0. 7]
Access Type: Read/Write
31 30 29 28 27 26 25 24
RXBYTECNT
23 22 21 20 19 18 17 16
RXBYTECNT
15 14 13 12 11 10 9 8
EPEDS DTGLE EPTYPE
7 6 5 4 3 2 1 0
DIR RX_DATA_ FORCE TXPKTRDY STALLSENT RXSETUP RX_DATA_ TXCOMP
BK1 STALL ISOERROR BK0
556 AT91RM9200
1768BATARM08/03
AT91RM9200
1 = No effect.
This flag is used to notify the USB device firmware that a valid Setup data packet has been sent by the host and success-
fully received by the USB device. The USB device firmware may transfer Setup data from the FIFO by reading the
USB_FDRx register to the microcontroller memory. Once a transfer has been done, RXSETUP must be cleared by the
device firmware.
Ensuing Data OUT transactions is not accepted while RXSETUP is set.
STALLSENT: Stall sent (Control, Bulk Interrupt endpoints)/ ISOERROR (Isochronous endpoints)
This flag generates an interrupt while it is set to one.
557
1768BATARM08/03
Please refer to chapters 8.4.4 and 9.4.5 of the Universal Serial Bus Specification, Rev. 1.1 to get more information on the
STALL handshake.
Control endpoints: during the data stage and status stage, this indicates that the microcontroller can not complete the
request.
Bulk and interrupt endpoints: notify the host that the endpoint is halted.
The host acknowledges the STALL, device firmware is notified by the STALLSENT flag.
RX_DATA_BK1: Receive Data Bank 1 (only used by endpoints with ping-pong attributes)
This flag generates an interrupt while it is set to one.
Write (Cleared by the firmware)
0 = Notify USB device that data have been read in the FIFOs Bank 1.
1 = No effect.
Read (Set by the USB peripheral)
0 = No data packet has been received in the FIFO's Bank 1.
1 = A data packet has been received, it has been stored in FIFO's Bank 1.
When the device firmware has polled this bit or has been interrupted by this signal, it must transfer data from the FIFO to
microcontroller memory. The number of bytes received is available in RXBYTECNT field. Bank 1 FIFO values are read
through USB_FDRx register. Once a transfer is done, the device firmware must release Bank 1 to the USB device by clear-
ing RX_DATA_BK1.
DIR: Transfer Direction (only available for control endpoints)
Read/Write
0 = Allow Data OUT transactions in the control data stage.
1 = Enable Data IN transactions in the control data stage.
Please refer to Chapter 8.5.2 of the Universal Serial Bus Specification, Rev. 1.1 to get more information on the control data
stage.
This bit must be set before USB_CSRx/RXSETUP is cleared at the end of the setup stage. According to the request sent in
the setup data packet, the data stage is either a device to host (DIR = 1) or host to device (DIR = 0) data transfer. It is not
necessary to check this bit to reverse direction for the status stage.
EPTYPE[2:0]: Endpoint Type
Read/Write
000 Control
001 Isochronous OUT
101 Isochronous IN
010 Bulk OUT
110 Bulk IN
011 Interrupt OUT
111 Interrupt IN
558 AT91RM9200
1768BATARM08/03
AT91RM9200
559
1768BATARM08/03
USB FIFO Data Register
Register Name: USB_FDRx [x = 0. 7]
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
FIFO_DATA
560 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The USB Host Port interfaces the USB with the host application. It handles Open HCI protocol
(Open Host Controller Interface) as well as USB v2.0 Full-speed and Low-speed protocols. It
also provides a simple read/write protocol on the ASB.
The USB Host Port integrates a root hub and transceivers on downstream ports. It provides
several high-speed half-duplex serial communication ports at a baud rate of 12 Mbit/s. Up to
127 USB devices (printer, camera, mouse, keyboard, disk, etc.) and the USB hub can be con-
nected to the USB host in the USB tiered star topology.
The USB Host Port controller is fully compliant with the Open HCI specification. The standard
OHCI USB stack driver can be easily ported to ATMELs architecture in the same way all exist-
ing class drivers run without hardware specialization.
This means that all standard class devices are automatically detected and available to the
user application. As an example, integrating an HID (Human Interface Device) class driver
provides a plug & play feature for all USB keyboards and mouses.
Key features of the USB Host Port are:
Compliance with Open HCI Rev 1.0 Specification
Compliance with USB V2.0 Full Speed and Low Speed Specification
Supports Both Low-speed 1.5 Mbps and Full-speed 12 Mbps USB devices
Root Hub Integrated with Two Downstream USB Ports
Embedded USB Transceivers (Number of Transceivers is Product Dependant)
Supports Power Management
Operates as a Master on the ASB Bus
Block Diagram
Figure 259. USB Host Port Block Diagram
ASB
HCI List Processor OHCI Root
Slave Block Block Hub Registers Embedded USB
v2.0 Full-speed Transceiver
OHCI Control ED & TD
Registers PORT S/M USB transceiver DP
Regsisters
Root Hub DM
and
DP
Host SIE PORT S/M USB transceiver
DM
uhp_int
MCK
UDPCK
561
1768BATARM08/03
Access to the USB host operational registers is achieved through the ASB bus interface. The
Open HCI host controller initializes master DMA transfers with the ASB bus as follows:
Fetches endpoint descriptors and transfer descriptors
Access to endpoint data from system memory
Access to the HC communication area
Write status and retire transfer Descriptor
All of the ASB memory map is accessible to the USB host master DMA. Thus there is no need
to define a dedicated physical memory area to the USB host.
The USB root hub is integrated in the USB host. Several USB downstream ports are available.
The number of downstream ports can be determined by the software driver reading the root
hubs operational registers. Device connection is automatically detected by the USB host port
logic.
Warning: a pull-down must be connected to DP on the board. Otherwise The USB host will
permanently detect a device connection on this port.
USB physical transceivers are integrated in the product and driven by the root hubs ports.
Over current protection on ports can be activated by the USB host controller. Atmels standard
product does not dedicate pads to external over current protection.
Product
Dependencies
I/O Lines DPs and DMs are not controlled by any PIO controllers. The embedded USB physical trans-
ceivers are controlled by the USB host controller.
Power The USB host controller requires a 48 MHz clock. This clock must be generated by a PLL with
Management a correct accuracy of 0.25%.
Thus the USB device peripheral receives two clocks from the Power Management Controller
(PMC): the master clock MCK used to drive the peripheral user interface (MCK domain) and
the UHPCLK 48 MHz clock used to interface with the bus USB signals (Recovered 12 MHz
domain).
Interrupt The USB host interface has an interrupt line connected to the Advanced Interrupt Controller
(AIC).
Handling USB host interrupts requires programming the AIC before configuring the UHP.
Functional Please refer to the Open Host Controller Interface Specification for USB Release 1.0.a.
Description
Host Controller There are two communication channels between the Host Controller and the Host Controller
Interface Driver. The first channel uses a set of operational registers located on the USB Host Control-
ler. The Host Controller is the target for all communications on this channel. The operational
registers contain control, status and list pointer registers. They are mapped in the ASB mem-
ory mapped area. Within the operational register set there is a pointer to a location in the
processor address space named the Host Controller Communication Area (HCCA). The
HCCA is the second communication channel. The host controller is the master for all commu-
nication on this channel. The HCCA contains the head pointers to the interrupt Endpoint
562 AT91RM9200
1768BATARM08/03
AT91RM9200
Descriptor lists, the head pointer to the done queue and status information associated with
start-of-frame processing.
The basic building blocks for communication across the interface are Endpoint Descriptors
(ED, 4 double words) and Transfer Descriptors (TD, 4 or 8 double words). The host controller
assigns an Endpoint Descriptor to each endpoint in the system. A queue of Transfer Descrip-
tors is linked to the Endpoint Descriptor for the specific endpoint.
Open HCI
Status Interrupt 2
...
Event
Interrupt 31
Frame Int
...
Ratio
Control
Bulk
...
Done
Device Register
in Memory Space Shared RAM
HUB Driver
USBD Driver
Hardware
Host Controller Hardware
563
1768BATARM08/03
USB Handling is done through several layers as follows:
Host controller hardware and serial engine: Transmit and receive USB data on the bus.
Host controller driver: Drives the Host controller hardware and handle the USB protocol
USB Bus driver and hub driver: Handles USB commands and enumeration. Offers a
hardware independent interface.
Mini driver: Handles device specific commands.
Class driver: handles standard devices. This acts as a generic driver for a class of devices,
for example the HID driver.
Type A Connector
10F 100nF 10nF
HDMA
27
or
HDMB
HDPA
or
27
HDPB 47pF
15k 15k
47pF
As device connection is automatically detected by the USB host port logic, a pull-down must
be connected on DP and DM on the board. Otherwise the USB host will permanently detect a
device connection on this port.
564 AT91RM9200
1768BATARM08/03
AT91RM9200
Overview The Ethernet MAC is the hardware implementation of the MAC sub-layer OSI reference model
between the physical layer (PHY) and the logical link layer (LLC). It controls the data
exchange between a host and a PHY layer according to Ethernet IEEE 802.3u data frame for-
mat. The Ethernet MAC contains the required logic and transmit and receive FIFOs for DMA
management. In addition, it is interfaced through MDIO/MDC pins for PHY layer management.
The Ethernet MAC can transfer data in media-independent interface (MII) or reduced media-
independent interface (RMII) modes depending on the pinout configuration.
The aim of the reduced interface is to lower the pin count for a switch product that can be con-
nected to multiple PHY interfaces. The characteristics specific to RMII mode are:
Single clock at 50 MHz frequency
Reduction of required control pins
Reduction of data paths to di-bit (2-bit wide) by doubling clock frequency
10 Mbits/sec. and 100 Mbits/sec. data capability
The major features of the EMAC are:
Compatibility with IEEE Standard 802.3
10 and 100 Mbits per second data throughput capability
Full- and half-duplex operation
MII or RMII interface to the physical layer
Register interface to address, status and control registers
DMA interface
Interrupt generation to signal receive and transmit completion
28-byte transmit and 28-byte receive FIFOs
Automatic pad and CRC generation on transmitted frames
Address checking logic to recognize four 48-bit addresses
Supports promiscuous mode where all valid frames are copied to memory
Supports physical layer management through MDIO interface control of alarm and update
time/calendar data in
565
1768BATARM08/03
Block Diagram
Figure 263. Block Diagram
EXTEN-EXTER
APB Bridge
ECRS-ECOL
ERXER-ERXDV
APB
Ethernet MAC PIO
ERX0-ERX3
ETX0-ETX3
EMDC
PMC MCK
EMDIO
Interrupt Control
EF100
EMAC IRQ
UDP TCP
IP ARP/RARP
ETHERNET Driver
NETWORK
566 AT91RM9200
1768BATARM08/03
AT91RM9200
Product Dependencies
I/O Lines The pins used for interfacing the compliant external devices may be multiplexed with PIO
lines. The programmer must first program the PIO controllers to assign the EMAC pins to their
peripheral functions. In RMII mode, unused pins (see Table 102: MII/RMII Signal Mapping)
can be used as general I/O lines.
Power The EMAC may be clocked through the Power Management Controller (PMC), so the pro-
Management grammer must first configure the PMC to enable the EMAC clock.
Interrupt The EMAC has an interrupt line connected to the Advanced Interrupt Controller (AIC). Han-
dling the EMAC interrupt requires programming the AIC before configuring the EMAC.
567
1768BATARM08/03
Functional Description
The Ethernet Media Access Control (EMAC) engine is fully compatible with the IEEE 802.3
Ethernet standard. It manages frame transmission and reception including collision detection,
preamble generation and detection, CRC control and generation and transmitted frame
padding.
The MAC functions are:
Frame encapsulation and decapsulation
Error detection
Media access management (MII, RMII)
Address Checker
568 AT91RM9200
1768BATARM08/03
AT91RM9200
General The Ethernet MAC is capable of interfacing to both RMII and MII Interfaces. The RMII bit in the
ETH_CFG register controls the interface that is selected. When this bit is set, the RMII inter-
face is selected, else the MII interface is selected.
The MII and RMII interface are capable of both 10Mb/s and 100Mb/s data rates as described
in the IEEE 802.3u standard. The signals used by the MII and RMII interfaces are described in
the Table 102.
The intent of the RMII is to provide a reduced pin count alternative to the IEEE 802.3u MII. It
uses 2 bits for transmit (ETX0 and ETX1) and two bits for receive (ERX0 and ERX1). There is
a Transmit Enable (ETXEN), a Receive Error (ERXER), a Carrier Sense (ECRS_DV), and a
50 MHz Reference Clock (ETXCK_REFCK) for 100Mb/s data rate.
RMII Transmit and The same signals are used internally for both the RMII and the MII operations. The RMII maps
Receive Operation these signals in a more pin-efficient manner. The transmit and receive bits are converted from
a 4-bit parallel format to a 2-bit parallel scheme that is clocked at twice the rate. The carrier
sense and data valid signals are combined into the ECRS_ECRSDV signal. This signal con-
tains information on carrier sense, FIFO status, and validity of the data. Transmit error bit
(ETXER) and collision detect (ECOL) are not used in RMII mode.
569
1768BATARM08/03
Transmit/Receive A standard IEEE 802.3 packet consists of the following fields: preamble, start of frame delim-
Operation iter (SFD), destination address (DA), source address (SA), length, data (Logical Link Control
Data) and frame check sequence CRC32 (FCS).
Preamble and Start of The preamble field is used to acquire bit synchronization with an incoming packet. When
Frame Delimiter (SFD) transmitted, each packet contains 62 bits of alternating 1,0 preamble. Some of this preamble
is lost as the packet travels through the network. Byte alignment is performed with the Start of
Frame Delimiter (SFD) pattern that consists of two consecutive 1's.
Destination Address The destination address (DA) indicates the destination of the packet on the network and is
used to filter unwanted packets. There are three types of address formats: physical, multicast
and broadcast. The physical address is a unique address that corresponds only to a single
node. All physical addresses have an MSB of 0.
Multicast addresses begin with an MSB of 1. The MAC filters multicast addresses using a
standard hashing algorithm that maps all multicast addresses into a 6-bit value. This 6-bit
value indexes a 64-bit array that filters the value. If the address consists of all ones, it is a
broadcast address, indicating that the packet is intended for all nodes.
Source Address The source address (SA) is the physical address of the node that sent the packet. Source
addresses cannot be multicast or broadcast addresses. This field is passed to buffer memory.
Length/Type If the value of this field is less than or equal to 1500, then the Length/Type field indicates the
number of bytes in the subsequent LLC Data field. If the value of this field is greater than or
equal to 1536, then the Length/Type field indicates the nature of the MAC client protocol (pro-
tocol type).
LLC Data The data field consists of anywhere from 46 to 1500 bytes. Messages longer than 1500 bytes
need to be broken into multiple packets. Messages shorter than 46 bytes require appending a
pad to bring the data field to the minimum length of 46 bytes. If the data field is padded, the
number of valid data bytes is indicated in the length field.
FCS Field The Frame Check Sequence (FCS) is a 32-bit CRC field, calculated and appended to a packet
during transmission to allow detection of errors when a packet is received. During reception,
error free packets result in a specific pattern in the CRC generator. Packets with improper
CRC will be rejected.
570 AT91RM9200
1768BATARM08/03
AT91RM9200
Frame Format The original Ethernet standards defined the minimum frame size as 64 bytes and the maxi-
Extensions mum as 1518 bytes. These numbers include all bytes from the Destination MAC Address field
through the Frame Check Sequence field. The Preamble and Start Frame Delimiter fields are
not included when quoting the size of a frame. The IEEE 802.3ac standard extended the max-
imum allowable frame size to 1522 bytes to allow a VLAN tag to be inserted into the Ethernet
frame format. The bit BIG defined in the ETH_CFG register aims to process packet with VLAN
tag.
The VLAN protocol permits insertion of an identifier, or tag, into the Ethernet frame format to
identify the VLAN to which the frame belongs. It allows frames from stations to be assigned to
logical groups. This provides various benefits, such as easing network administration, allowing
formation of work groups, enhancing network security, and providing a means of limiting
broadcast domains (refer to IEEE standard 802.1Q for definition of the VLAN protocol). The
802.3ac standard defines only the implementation details of the VLAN protocol that are spe-
cific to Ethernet.
If present, the 4-byte VLAN tag is inserted into the Ethernet frame between the Source MAC
Address field and the Length field. The first 2-bytes of the VLAN tag consist of the 802.1Q
Tag Type and are always set to a value of 0x8100. The 0x8100 value is a reserved
Length/Type field assignment that indicates the presence of the VLAN tag, and signals that
the traditional Length/Type field can be found at an offset of four bytes further into the frame.
The last two bytes of the VLAN tag contain the following information:
The first three bits are a User Priority Field that may be used to assign a priority level to
the Ethernet frame.
The following one bit is a Canonical Format Indicator (CFI) used in Ethernet frames to
indicate the presence of a Routing Information Field (RIF).
The last twelve bits are the VLAN Identifier (VID) that uniquely identifies the VLAN to
which the Ethernet frame belongs.
With the addition of VLAN tagging, the 802.3ac standard permits the maximum length of an
Ethernet frame to be extended from 1518 bytes to 1522 bytes. Table 104 illustrates the format
of an Ethernet frame that has been tagged with a VLAN identifier according to the IEEE
802.3ac standard.
571
1768BATARM08/03
DMA Operations Frame data is transferred to and from the Ethernet MAC via the DMA interface. All transfers
are 32-bit words and may be single accesses or bursts of two, three or four words. Burst
accesses do not cross 16-byte boundaries.
The DMA controller performs four types of operations on the ASB bus. In order of priority,
these operations are receive buffer manager read, receive buffer manager write, transmit data
DMA read and receive data DMA write.
Transmitter Mode Transmit frame data needs to be stored in contiguous memory locations. It does not need to
be word-aligned.
The transmit address register is written with the address of the first byte to be transmitted.
Transmit is initiated by writing the number of bytes to transfer (length) to the transmit control
register.
The transmit channel then reads data from memory 32 bits at a time and places them in the
transmit FIFO.
The transmit block starts frame transmission when three words have been loaded into the
FIFO.
The transmit address register must be written before the transmit control register. While a
frame is being transmitted, it is possible to set up one other frame for transmission by writing
new values to the transmit address and control registers. Reading the transmit address regis-
ter returns the address of the buffer currently being accessed by the transmit FIFO.
Reading the transmit control register returns the total number of bytes to be transmitted. The
BNQ bit in the Transmit Status Register indicates whether another buffer can be safely
queued. An interrupt is generated whenever this bit is set.
Frame assembly starts by adding preamble and the start frame delimiter. Data is taken from
the transmit FIFO word-by-word. If necessary, padding is added to make the frame length 60
bytes. The CRC is calculated as a 32-bit polynomial. This is inverted and appended to the end
of the frame, making the frame length a minimum of 64 bytes. The CRC is not appended if the
NCRC bit is set in the transmit control register.
In full-duplex mode, frames are transmitted immediately. Back-to-back frames are transmitted
at least 96 bit times apart to guarantee the inter-frame gap.
In half-duplex mode, the transmitter checks carrier sense. If asserted, it waits for it to de-assert
and then starts transmission after the inter-frame gap of 96 bit-times.
If the collision signal is asserted during transmission, the transmitter transmits a jam sequence
of 32 bits taken from the data register and then retries transmission after the backoff time has
elapsed. An error is indicated and any further attempts aborted if 16 attempts cause collisions.
If transmit DMA underruns, bad CRC is automatically appended using the same mechanism
as jam insertion. Underrun also causes TXER to be asserted.
Receiver Mode When a packet is received, it is checked for valid preamble, CRC, alignment, length and
address. If all these criteria are met, the packet is stored successfully in a receive buffer. If at
the end of reception the CRC is bad, then the received buffer is recovered. Each received
frame including CRC is written to a single receive buffer.
Receive buffers are word-aligned and are capable of containing 1518 or 1522 bytes (BIG = 1
in ETH_CFG) of data (the maximum length of an Ethernet frame).
The start location for each received frame is stored in memory in a list of receive buffer
descriptors at a location pointed to by the receive buffer queue pointer register. Each entry in
572 AT91RM9200
1768BATARM08/03
AT91RM9200
the list consists of two words. The first word is the address of the received buffer; the second is
the receive status. Table 105 defines an entry in the received buffer descriptor list.
To receive frames, the buffer queue must be initialized by writing an appropriate address to
bits [31:2] in the first word of each list entry. Bit zero of word zero must be written with zero.
After a frame is received, bit zero becomes set and the second word indicates what caused
the frame to be copied to memory. The start location of the received buffer descriptor list
should be written to the received buffer queue pointer register before receive is enabled (by
setting the receive enable bit in the network control register). As soon as the received block
starts writing received frame data to the receive FIFO, the received buffer manager reads the
first receive buffer location pointed to by the received buffer queue pointer register. If the filter
block is active, the frame should be copied to memory; the receive data DMA operation starts
writing data into the receive buffer. If an error occurs, the buffer is recovered. If the frame is
received without error, the queue entry is updated. The buffer pointer is rewritten to memory
with its low-order bit set to indicate successful frame reception and a used buffer. The next
word is written with the length of the frame and how the destination address was recognized.
The next receive buffer location is then read from the following word or, if the current buffer
pointer had its wrap bit set, the beginning of the table. The maximum number of buffer pointers
before a wrap bit is seen is 1024. If a wrap bit is not seen by then, a wrap bit is assumed in that
entry. The received buffer queue pointer register must be written with zero in its lower-order bit
positions to enable the wrap function to work correctly.
If bit zero is set when the receive buffer manager reads the location of the receive buffer, then
the buffer has already been used and cannot be used again until software has processed the
frame and cleared bit zero. In this case, the DMA block sets the buffer unavailable bit in the
received status register and triggers an interrupt. The frame is discarded and the queue entry
is reread on reception of the next frame to see if the buffer is now available. Each discarded
frame increments a statistics register that is cleared on being read. When there is network
congestion, it is possible for the MAC to be programmed to apply backpressure.
This is when half-duplex mode collisions are forced on all received frames by transmitting 64
bits of data (a default pattern).
Reading the received buffer queue register returns the location of the queue entry currently
being accessed. The queue wraps around to the start after either 1024 entries (i.e., 2048
words) or when the wrap bit is found to be set in bit 1 of the first word of an entry.
573
1768BATARM08/03
Table 105. Received Buffer Descriptor List
Bit Function
28 External address (optional)
27 Unknown source address (reserved for future use)
26 Local address match (Specific address 1 match)
25 Local address match (Specific address 2 match)
24 Local address match (Specific address 3 match)
23 Local address match (Specific address 4 match)
22:11 Reserved; written to 0
10:0 Length of frame including FCS
Address Checking Whether or not a frame is stored depends on what is enabled in the network configuration reg-
ister, the contents of the specific address and hash registers and the frame destination
address. In this implementation of the MAC the frame source address is not checked.
A frame is not copied to memory if the MAC is transmitting in half-duplex mode at the time a
destination address is received.
The hash register is 64 bits long and takes up two locations in the memory map.
There are four 48-bit specific address registers, each taking up two memory locations. The
first location contains the first four bytes of the address; the second location contains the last
two bytes of the address stored in its least significant byte positions. The addresses stored
can be specific, group, local or universal.
Ethernet frames are transmitted a byte at a time, LSB first. The first bit (i.e., the LSB of the first
byte) of the destination address is the group/individual bit and is set one for multicast
addresses and zero for unicast. This bit corresponds to bit 24 of the first word of the specific
address register. The MSB of the first byte of the destination address corresponds to bit 31 of
the specific address register.
The specific address registers are compared to the destination address of received frames
once they have been activated. Addresses are deactivated at reset or when the first byte
[47:40] is written and activated or when the last byte [7:0] is written. If a receive frame address
matches an active address, the local match signal is set and the store frame pulse signal is
sent to the DMA block via the HCLK synchronization block.
A frame can also be copied if a unicast or multicast hash match occurs, it has the broadcast
address of all ones, or the copy all frames bit in the network configuration register is set.
The broadcast address of 0xFFFFFFFF is recognized if the no broadcast bit in the network
configuration register is zero. This sets the broadcast match signal and triggers the store
frame signal.
The unicast hash enable and the multicast hash enable bits in the network configuration regis-
ter enable the reception of hash matched frames. So all multicast frames can be received by
setting all bits in the hash register.
The CRC algorithm reduces the destination address to a 6-bit index into a 64-bit hash regis-
ter.If the equivalent bit in the register is set, the frame is matched depending on whether the
frame is multicast or unicast and the appropriate match signals are sent to the DMA block. If
the copy all frames bit is set in the network configuration register, the store frame pulse is
always sent to the DMA block as soon as any destination address is received.
574 AT91RM9200
1768BATARM08/03
AT91RM9200
575
1768BATARM08/03
Table 106. EMAC Register Mapping
Offset Register Register Name Read/Write Reset
0x94 EMAC Hash Address Low [31:0] ETH_HSL Read/write 0x0
EMAC Specific Address 1 Low, First 4
0x98 ETH_SA1L Read/write 0x0
Bytes
EMAC Specific Address 1 High, Last 2
0x9C ETH_SA1H Read/write 0x0
Bytes
EMAC Specific Address 2 Low, First 4
0xA0 ETH_SA2L Read/write 0x0
Bytes
EMAC Specific Address 2 High, Last 2
0xA4 ETH_SA2H Read/write 0x0
Bytes
EMAC Specific Address 3 Low, First 4
0xA8 ETH_SA3L Read/write 0x0
Bytes
EMAC Specific Address 3 High, Last 2
0xAC ETH_SA3H Read/write 0x0
Bytes
EMAC Specific Address 4 Low, First 4
0xB0 ETH_SA4L Read/write 0x0
Bytes
EMAC Specific Address 4 High, Last 2
0xB4 ETH_SA4H Read/write 0x0
Bytes
Note: For further details on the statistics registers, see Table 107 on page 593.
576 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
BP
7 6 5 4 3 2 1 0
WES ISR CSR MPE TE RE LBL LB
LB: Loopback
Optional. When set, loopback signal is at high level.
LBL: Loopback Local
When set, connects ETX[3:0] to ERX[3:0], ETXEN to ERXDV, forces full duplex and drives ERXCK and ETXCK_REFCK
with MCK divided by 4.
RE: Receive Enable
When set, enables the Ethernet MAC to receive data.
TE: Transmit Enable
When set, enables the Ethernet transmitter to send data.
MPE: Management Port Enable
Set to one to enable the management port. When zero, forces MDIO to high impedance state.
CSR: Clear Statistics Registers
This bit is write-only. Writing a one clears the statistics registers.
ISR: Increment Statistics Registers
This bit is write-only. Writing a one increments all the statistics registers by one for test purposes.
WES: Write Enable for Statistics Registers
Setting this bit to one makes the statistics registers writable for functional test purposes.
BP: Back Pressure
If this field is set, then in half-duplex mode collisions are forced on all received frames by transmitting 64 bits of data
(default pattern).
577
1768BATARM08/03
EMAC Configuration Register
Name: ETH_CFG
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RMII RTY CLK EAE BIG
7 6 5 4 3 2 1 0
UNI MTI NBC CAF BR FD SPD
SPD: Speed
Set to 1 to indicate 100 Mbit/sec, 0 for 10 Mbit/sec. Has no other functional effect.
FD: Full Duplex
If set to 1, the transmit block ignores the state of collision and carrier sense and allows receive while transmitting.
BR: Bit Rate
Optional.
CAF: Copy All Frames
When set to 1, all valid frames are received.
NBC: No Broadcast
When set to 1, frames addressed to the broadcast address of all ones are not received.
MTI: Multicast Hash Enable
When set multicast frames are received when six bits of the CRC of the destination address point to a bit that is set in the
hash register.
UNI: Unicast Hash Enable
When set, unicast frames are received when six bits of the CRC of the destination address point to a bit that is set in the
hash register.
BIG: Receive 1522 Bytes
When set, the MAC receives up to 1522 bytes. Normally the MAC receives frames up to 1518 bytes in length.
This bit allows to receive extended Ethernet frame with VLAN tag (IEEE 802.3ac)
EAE: External Address Match Enable
Optional.
CLK
The system clock (MCK) is divided down to generate MDC (the clock for the MDIO). To conform with IEEE standard 802.3
MDC must not exceed 2.5 MHz. At reset this field is set to 10 so that MCK is divided by 32.
CLK MDC
00 MCK divided by 8
01 MCK divided by 16
10 MCK divided by 32
11 MCK divided by 64
578 AT91RM9200
1768BATARM08/03
AT91RM9200
579
1768BATARM08/03
EMAC Status Register
Name: ETH_SR
Access Type: Read only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
IDLE MDIO LINK
LINK
Reserved.
MDIO
0 = MDIO pin not set.
1 = MDIO pin set.
IDLE
0 = PHY logic is idle.
1 = PHY logic is running.
580 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
ADDRESS
15 14 13 12 11 10 9 8
ADDRESS
7 6 5 4 3 2 1 0
ADDRESS
581
1768BATARM08/03
EMAC Transmit Control Register
Name: ETH_TCR
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
NCRC LEN
7 6 5 4 3 2 1 0
LEN
582 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
UND COMP BNQ IDLE RLE COL OVR
583
1768BATARM08/03
EMAC Receive Buffer Queue Pointer Register
Name: ETH_RBQP
Access Type: Read/Write
31 30 29 28 27 26 25 24
ADDRESS
23 22 21 20 19 18 17 16
ADDRESS
15 14 13 12 11 10 9 8
ADDRESS
7 6 5 4 3 2 1 0
ADDRESS
584 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
7 6 5 4 3 2 1 0
OVR REC BNA
585
1768BATARM08/03
EMAC Interrupt Status Register
Name: ETH_ISR
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
ABT ROVR LINK TIDLE
7 6 5 4 3 2 1 0
TCOM TBRE RTRY TUND TOVR RBNA RCOM DONE
586 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
ABT ROVR LINK TIDLE
7 6 5 4 3 2 1 0
TCOM TBRE RTRY TUND TOVR RBNA RCOM DONE
587
1768BATARM08/03
EMAC Interrupt Disable Register
Name: ETH_IDR
Access Type: Write only
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
ABT ROVR LINK TIDLE
7 6 5 4 3 2 1 0
TCOM TBRE RTRY TUND TOVR RBNA RCOM DONE
588 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
ABT ROVR LINK TIDLE
7 6 5 4 3 2 1 0
TCOM TBRE RTRY TUND TOVR RBNA RCOM DONE
Important Note: The interrupt is disabled when the corresponding bit is set. This is non-standard for AT91 products as
generally a mask bit set enables the interrupt.
589
1768BATARM08/03
EMAC PHY Maintenance Register
Name: ETH_MAN
Access Type: Read/Write
31 30 29 28 27 26 25 24
LOW HIGH RW PHYA
23 22 21 20 19 18 17 16
PHYA REGA CODE
15 14 13 12 11 10 9 8
DATA
7 6 5 4 3 2 1 0
DATA
Writing to this register starts the shift register that controls the serial connection to the PHY. On each shift cycle the MDIO
pin becomes equal to the MSB of the shift register and LSB of the shift register becomes equal to the value of the MDIO
pin. When the shifting is complete an interrupt is generated and the IDLE field is set in the Network Status register.
When read, gives current shifted value.
DATA
For a write operation this is written with the data to be written to the PHY. After a read operation this contains the data read
from the PHY.
CODE
Must be written to 10 in accordance with IEEE standard 802.3. Reads as written.
REGA
Register address. Specifies the register in the PHY to access.
PHYA
PHY address. Normally is 0.
RW
Read/write Operation. 10 is read. 01 is write. Any other value is an invalid PHY management frame.
HIGH
Must be written with 1 to make a valid PHY management frame. Conforms with IEEE standard 802.3.
LOW
Must be written with 0 to make a valid PHY management frame. Conforms with IEEE standard 802.3.
590 AT91RM9200
1768BATARM08/03
AT91RM9200
23 22 21 20 19 18 17 16
ADDR
15 14 13 12 11 10 9 8
ADDR
7 6 5 4 3 2 1 0
ADDR
ADDR
Hash address bits 63 to 32.
23 22 21 20 19 18 17 16
ADDR
15 14 13 12 11 10 9 8
ADDR
7 6 5 4 3 2 1 0
ADDR
ADDR
Hash address bits 31 to 0.
591
1768BATARM08/03
EMAC Specific Address (1, 2, 3 and 4) High Register
Name: ETH_SA1H,...ETH_SA4H
Access Type: Read/Write
31 30 29 28 27 26 25 24
23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
ADDR
7 6 5 4 3 2 1 0
ADDR
ADDR
Unicast addresses (1, 2, 3 and 4), Bits 47:32.
23 22 21 20 19 18 17 16
ADDR
15 14 13 12 11 10 9 8
ADDR
7 6 5 4 3 2 1 0
ADDR
ADDR
Unicast addresses (1, 2, 3 and 4), Bits 31:0.
592 AT91RM9200
1768BATARM08/03
AT91RM9200
593
1768BATARM08/03
Table 107. Statistics Register Block (Continued)
Register Register Name Description
Undersize Frame Register ETH_USF An 8-bit register counting the number of frames received less that are than 64
bytes in length but that do not have either a CRC error, an alignment error or a
code error.
SQE Test Error Register ETH_SQEE An 8-bit register counting the number of frames where pin ECOL was not
asserted within a slot time of pin ETXEN being deasserted.
Discarded RX Frame Register ETH_DRFC This 16-bit counter is incremented every time an address-recognized frame is
received but cannot be copied to memory because the receive buffer is
available.
594 AT91RM9200
1768BATARM08/03
AT91RM9200
DC Output Current
(SDA10, SDCKE, SDWE, RAS, CAS) .................. 16 mA
DC Output Current
(Any other pin) ........................................................ 8 mA
595
1768BATARM08/03
DC Characteristics
The following characteristics are applicable to the operating temperature range: TA = -40C to 85C, unless otherwise spec-
ified and are certified for a junction temperature up to TJ = 100C.
596 AT91RM9200
1768BATARM08/03
AT91RM9200
Clocks Characteristics
These parameters are given in the following conditions:
VDDCORE = 1.8V
Ambient Temperature = 25C
The Temperature Derating Factor described in the section Temperature Derating Factor on page 604 and VDDCORE Volt-
age Derating Factor described in the section VDDCORE Voltage Derating Factor on page 604 are both applicable to
these characteristics.
597
1768BATARM08/03
Power Consumption
The values in Table 113 and Table 114 are measured values on the AT91RM9200DK Evalua-
tion Board with operating conditions as follows:
VDDIO = 3.3V
VDDCORE = VDDPLL = VDDOSC = 1.8V
TA = 25C
MCK = 60 MHz
PCK = 180 Mhz
SLCK = 32.768 kHz
These figures represent the power consumption measured on the VDDCORE power supply.
598 AT91RM9200
1768BATARM08/03
AT91RM9200
PLL Characteristics
Table 117. Phase Lock Loop Characteristics
Symbol Parameter Conditions Min Typ Max Unit
FOUT Output Frequency 80 240 MHz
FIN Input Frequency 1 32 MHz
KO VCO Gain 120 190 300 MHz/V
IP Pump Current 36 44 60 A
599
1768BATARM08/03
Transceiver Characteristics
Electrical Characteristics
Table 118. Electrical Parameters
Symbol Parameter Conditions Min Typ Max Unit
Input Levels
VIL Low Level 0.8 V
VIH High Level 2.0 V
VDI Differential Input Sensivity |(D+) - (D-)| 0.2 V
VCM Differential Input Common 0.8 2.5 V
Mode Range
CIN Transceiver capacitance Capacitance to ground on each line 20 pF
I Hi-Z State Data Line Leakage 0V < VIN < 3.3V -5 +5 A
REXT Recommended External USB In series with each USB pin with 5% 27
Series Resistor
Output Levels
VOL Low Level Output Measured with RL of 1.425 kOhm tied 0.3 V
to 3.6V
VOH High Level Output Measured with RL of 14.25 kOhm tied 2.8 V
to GND
VCRS Output Signal Crossover Measure conditions described in 1.3 2.0 V
Volatge Figure 266
600 AT91RM9200
1768BATARM08/03
AT91RM9200
Switching Characteristics
Table 119. In Slow Mode
Symbol Parameter Conditions Min Typ Max Unit
tFR Transition Rise Time CLOAD = 400 pF 75 300 ns
tFE Transition Fall Time CLOAD = 400 pF 75 300 ns
tFRFM Rise/Fall time Matching CLOAD = 400 pF 80 120 %
(b)
601
1768BATARM08/03
602 AT91RM9200
1768BATARM08/03
AT91RM9200
AT91RM9200 AC Characteristics
Conditions and The delays are given as typical values in the following conditions:
Timings VDDIOM = 3.3V
Computation VDDCORE = 1.8V
Ambient Temperature = 25C
Load Capacitance = 0 pF
The output level change detection is (0.5 x VDDIOM).
The input level is (0.3 x VDDIOM) for a low-level detection and is (0.7 x VDDIOM) for a high-
level detection.
The minimum and maximum values given in the AC characteristics tables of this datasheet
take into account process variation and design. In order to obtain the timing for other condi-
tions, the following equation should be used:
where:
T is the derating factor in temperature given in Figure 267 on page 604.
VDDCORE is the derating factor for the Core Power Supply given in Figure 268 on page
604.
tDATASHEET is the minimum or maximum timing value given in this datasheet for a load
capacitance of 0 pF.
VDDIOM is the derating factor for the IOM Power Supply given in Figure 269 on page 605.
CSIGNAL is the capacitance load on the considered output pin(1).
CSIGNAL is the load derating factor depending on the capacitance load on the related
output pins given in Min and Max in this datasheet.
The input delays are given as typical values.
Note: 1. The user must take into account the package capacitance load contribution (CIN) described
in Table 109, DC Characteristics, on page 596.
603
1768BATARM08/03
Temperature Derating Factor
1.2
1.1
Derating Factor
0.9
0.8
-60 -40 -20 0 20 40 60 80 100 120 140 160
Operating Temperature (C)
2.5
Derating Factor
1.5
0.5
1 1.05 1.1 1.15 1.2 1.25 1.3 1.35 1.4 1.45 1.5 1.55 1.6 1.65 1.7 1.75 1.8 1.85 1.9 1.95
Core Supply Voltage (V)
604 AT91RM9200
1768BATARM08/03
AT91RM9200
3.5
2.5
Derating Factor
1.5
0.5
1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3 3.1 3.2 3.3 3.4 3.5 3.6
I/O Supply Voltage (V)
Note: The derating factor in this example is applicable only to timings related to output pins.
605
1768BATARM08/03
EBI Timings
MCK Falling to Chip Select Active CNCS = 0 pF (nacss x tCPMCK) + 4.3 (2) (nacss x tCPMCK) + 6.5 (2) ns
SMC5
(Address to Chip Select Setup) (1) CNCS derating 0.028 0.045 ns/pF
(2) (2)
Chip Select Inactive to MCK Falling CNCS = 0 pF (nacss x tCPMCK) + 4.4 (nacss x tCPMCK) + 6.5 ns
SMC6
(Address to Chip Select Setup) (1) CNCS derating 0.028 0.045 ns/pF
NCS Minimum Pulse Width (((n + 2) - (2 x nacss))
SMC7 CNCS = 0 pF ns
(Address to Chip Select Setup) (1) x tCPMCK)
(2) (3)
606 AT91RM9200
1768BATARM08/03
AT91RM9200
.
Table 122. SMC Write Signals
Symbol Parameter Conditions Min Max Units
607
1768BATARM08/03
Table 122. SMC Write Signals (Continued)
Symbol Parameter Conditions Min Max Units
C = 0 pF (((n + 1) - nacss) x tCPMCK) + ns
tCHMCK - 1.4 (2) (3)
Data Out Valid before NCS High
SMC24
(Address to Chip Select Setup Cycles) (1) CDATA derating - 0.044 ns/pF
CNCS derating 0.045 ns/pF
C = 0 pF nacss x tCPMCK - 0.4 (3) ns
Data Out Valid after NCS High
SMC25 CDATA derating - 0.044 ns/pF
(Address to Chip Select Setup Cycles) (1)
CNCS derating 0.045 ns/pF
608 AT91RM9200
1768BATARM08/03
AT91RM9200
609
1768BATARM08/03
Table 123. SMC Read Signals (Continued)
Symbol Parameter Conditions Min Max Units
CNRD = 0 pF n x tCHMCK ns
SMC45 NRD Minimum Pulse Width (2) (3) (7) + tCHMCK - 0.2 (4)
CNRD derating 0.002 ns/pF
CNRD = 0 pF ((n + 1) x tCHMCK) + tCHMCK ns
SMC46 NRD Minimum Pulse Width (2) (3) - 0.2 (4)
CNRD derating 0.002 ns/pF
Notes: 1. Early Read Protocol.
2. Standard Read Protocol.
3. The derating factor is not to be applied to tCHMCK or tCPMCK.
4. n = Number of standard Wait States inserted.
5. nacss = Number of Address to Chip Select Setup Cycles inserted.
6. h = Number of Hold Cycles inserted.
7. Not applicable when Address to Chip Select Setup Cycles are inserted.
8. Only one of these two timings needs to be met.
9. Only one of these two timings needs to be met.
610 AT91RM9200
1768BATARM08/03
Notes:
1768BATARM08/03
MCK
internal
signal
SMC4 SMC4 SMC4 SMC4 SMC4 SMC4 SMC4 SMC4 SMC4 SMC4 SMC4 SMC4
A1 - A25
SMC37
SMC8
SMC37 SMC17 SMC37
NWAIT
SMC1 SMC35 SMC1 SMC1 SMC35 SMC1 SMC1 SMC1
SMC35
SMC2 SMC36 SMC2 SMC2 SMC36 SMC2 SMC2 SMC2
SMC36
NUB/NLB/A0
SMC29
SMC31
SMC29
SMC31
SMC15
SMC16
SMC44 SMC44
NRD(1)
SMC30
SMC32
SMC30
SMC32
SMC30
SMC32
SMC45 SMC45
SMC45
NRD(2)
D0 - D15
Read
SMC10
SMC12
SMC11
SMC13
SMC11
SMC13
NWR
SMC14
SMC19
SMC21
SMC14
SMC22
SMC14
D0 - D15
to Write
AT91RM9200
611
612
Notes:
MCK
internal
signal
AT91RM9200
NCS
A1 - A25
NWAIT
SMC1 SMC35 SMC1
SMC2 SMC36 SMC2
NUB/NLB/A0
SMC30
Figure 271. SMC Signals Relative to MCK in LCD Interface Mode
SMC32
SMC46
NRD(1)
SMC33 SMC34
SMC42 SMC43
D0 - D15
Read
SMC11 SMC13
SMC28
NWR(2)
SMC14
SMC24 SMC25
D0 - D15
to Write
1768BATARM08/03
AT91RM9200
613
1768BATARM08/03
Table 125. SDRAMC Signals (Continued)
Symbol Parameter Conditions Min Max Units
CBA = 0 pF tCLMCK + 0.8 ns
SDRAMC13 Bank Change before SDCK Rising Edge (1)
CBA derating 0.028 ns/pF
CBA = 0 pF tCHMCK - 1.6 ns
SDRAMC14 Bank Change after SDCK Rising Edge (1)
CBA derating - 0.045 ns/pF
CCAS = 0 pF tCLMCK + 0.9 ns
SDRAMC15 CAS Low before SDCK Rising Edge (1)
CCAS derating 0.015 ns/pF
CCAS = 0 pF tCHMCK - 1.5 ns
SDRAMC16 CAS High after SDCK Rising Edge (1)
CCAS derating - 0.023 ns/pF
CDQM = 0 pF tCLMCK + 0.7 ns
SDRAMC17 DQM Change before SDCK Rising Edge (1)
CDQM derating 0.028 ns/pF
CDQM = 0 pF tCHMCK - 1.4 ns
SDRAMC18 DQM Change after SDCK Rising Edge (1)
CDQM derating - 0.045 ns/pF
SDRAMC19 D0-D15 in Setup before SDCK Rising Edge 1.3 ns
SDRAMC20 D0-D15 in Hold after SDCK Rising Edge 0.03 ns
SDRAMC21 D16-D31 in Setup before SDCK Rising Edge 2.0 ns
SDRAMC22 D16-D31 in Hold after SDCK Rising Edge -0.2 ns
CSDWE = 0 pF tCLMCK + 1.0 ns
SDRAMC23 SDWE Low before SDCK Rising Edge
CSDWE derating 0.015 ns/pF
CSDWE = 0 pF tCHMCK - 1.8 ns
SDRAMC24 SDWE High after SDCK Rising Edge
CSDWE derating -0.023 ns/pF
C = 0 pF tCLMCK - 2.7 ns
SDRAMC25 D0-D15 Out Valid before SDCK Rising Edge
CDATA derating -0.044 ns/pF
C = 0 pF tCHMCK - 2.4 ns
SDRAMC26 D0-D15 Out Valid after SDCK Rising Edge
CDATA derating -0.044 ns/pF
C = 0 pF tCLMCK - 3.2 ns
SDRAMC27 D16-D31 Out Valid before SDCK Rising Edge
CDATA derating -0.044 ns/pF
C = 0 pF tCHMCK - 2.4 ns
SDRAMC28 D16-D31 Out Valid after SDCK Rising Edge
CDATA derating -0.044 ns/pF
614 AT91RM9200
1768BATARM08/03
AT91RM9200
SDCK
SDRAMC1 SDRAMC2 SDRAMC3 SDRAMC4
SDCKE
SDRAMC5 SDRAMC6 SDRAMC5 SDRAMC6 SDRAMC5 SDRAMC6
SDCS
SDRAMC7 SDRAMC8
RAS
SDRAMC15 SDRAMC16 SDRAMC15 SDRAMC16
CAS
SDRAMC23 SDRAMC24
SDWE
SDRAMC9 SDRAMC10 SDRAMC9 SDRAMC10 SDRAMC9 SDRAMC10
SDA10
SDRAMC11 SDRAMC12 SDRAMC11 SDRAMC12 SDRAMC11 SDRAMC12
A0 - A9,
A11 - A13
SDRAMC13 SDRAMC14 SDRAMC13 SDRAMC14 SDRAMC13 SDRAMC14
BA0/BA1
SDRAMC17 SDRAMC18 SDRAMC17 SDRAMC18
DQM0 -
DQM3
SDRAMC19 SDRAMC20
D0 - D15
Read
SDRAMC21 SDRAMC22
D16 - D31
Read
SDRAMC25 SDRAMC26
D0 - D15
to Write
SDRAMC27 SDRAMC28
D16 - D31
to Write
615
1768BATARM08/03
BFC Signals Relative to BFCK
Table 126, Table 127 and Table 128 show timings relative to operating condition limits defined in the section Conditions
and Timings Computation on page 603.
Notes: 1. Field BFCC = 1 in Register BFC_MR, see Burst Flash Controller Mode Register on page 221.
2. Field BFCC = 2 in Register BFC_MR, see Burst Flash Controller Mode Register on page 221.
3. Field BFCC = 3 in Register BFC_MR, see Burst Flash Controller Mode Register on page 221.
616 AT91RM9200
1768BATARM08/03
AT91RM9200
617
1768BATARM08/03
618
Note:
BFCK
internal
signal
AT91RM9200
BFCS(1)
BFC1 BFC1
A1 - A25
BFC3 BFC4 BFC3 BFC4
BFAVD
BFC5 BFC6 BFC7 BFC5
Figure 273. BFC Signals Relative to BFCK in Asynchronous Mode
BFOE
BFC8
BFC9 BFC10
BFC11 BFC12
D0 - D15
Read
BFC21
BFC21 BFC22
BFWE
BFC15 BFC16 BFC15 BFC16
BFC17 BFC18
BFC19 BFC20
D0 - D15
to Write
1768BATARM08/03
AT91RM9200
619
1768BATARM08/03
620
Note:
BFCK
if signal controlled
address advance
BFC23 BFC24 BFC23 BFC24
BFBAA
if signal controlled
address advance
AT91RM9200
BFCK
if clock controlled
address advance
BFCS(1)
BFC1
Figure 274. BFC Signals Relative to BFCK in Burst Mode
A1 - A25
BFC3 BFC4
BFAVD
BFC5
BFC25 BFC26 BFC25 BFC26
BFRDY
BFC6 BFC7 BFC6 BFC7
BFC9
BFC10
BFC9
BFC10
BFC9
BFC10
BFC9
BFC9
BFC10
BFC9
BFC10
BFC9
BFC10
BFC9
BFC10
BFC9
BFC10
BFC9
BFC10
BFC9
BFC10
D0 - D15
Read
BFC21 BFC22
BFC15 BFC16
D0 - D15
to Write if
multiplexed
bus only
1768BATARM08/03
AT91RM9200
JTAG/ICE Timings
ICE0
NTRST
ICE1 ICE2
ICE5
TCK
ICE3 ICE4
TMS/TDI
ICE6 ICE7
TDO
ICE8
ICE9
621
1768BATARM08/03
JTAG Interface Signals
Table 130 shows timings relative to operating condition limits defined in the section Condi-
tions and Timings Computation on page 603
622 AT91RM9200
1768BATARM08/03
AT91RM9200
JTAG0
NTRST
JTAG1 JTAG2
JTAG5
TCK
JTAG3 JTAG4
TMS/TDI
JTAG6 JTAG7
TDO
JTAG8
JTAG9
Device
Inputs
JTAG10 JTAG11
Device
Outputs
JTAG12
JTAG13
623
1768BATARM08/03
ETM Timings
Timings Data Table 131 shows timings relative to operating condition limits defined in the section Condi-
tions and Timings Computation on page 603.
TCLK
tCLTCLK
tCPTCLK
TSYNC
TPS[2:0]
TPK[15:0]
ETM0
ETM1
ETM2
ETM3
Design When designing a PCB, it is important to keep the differences between trace length of ETM
Considerations signals as small as possible to minimize skew between them. In addition, crosstalk on the
trace port must be kept to a minimum as it can cause erroneous trace results. Stubs on these
traces can cause unpredictable responses, thus it is recommended to avoid stubs on the trace
lines.
The TCLK line should be series-terminated as close as possible to the microcontroller pins.
The maximum capacitance presented by the trace connector, cabling and interfacing logic
must be less than 15 pF.
624 AT91RM9200
1768BATARM08/03
AT91RM9200
Thermal Data In Table 132, the device lifetime is estimated using the MIL-217 standard in the moderately
controlled environmental model (this model is described as corresponding to an installation in
a permanent rack with adequate cooling air), depending on the device Junction Temperature.
(For details see the section Junction Temperature on page 626.)
Note that the user must be extremely cautious with this MTBF calculation. It should be noted
that the MIL-217 model is pessimistic with respect to observed values due to the way the
data/models are obtained (test under severe conditions). The life test results that have been
measured are always better than the predicted ones.
Table 133 summarizes the thermal resistance data depending on the package.
Reliability Data The number of gates and the device die size are provided Table 134 so that the user can cal-
culate reliability data for another standard and/or in another environmental model.
625
1768BATARM08/03
Junction The average chip-junction temperature, TJ, in C can be obtained from the following:
Temperature 1. T J = T A + ( P D JA )
2. T J = T A + ( P D ( HEATSINK + JC ) )
where:
JA = package thermal resistance, Junction-to-ambient (C/W), provided in Table 133 on
page 625.
JC = package thermal resistance, Junction-to-case thermal resistance (C/W), provided in
Table 133 on page 625.
HEAT SINK = cooling device thermal resistance (C/W), provided in the device datasheet.
PD = device power consumption (W) estimated from data provided in the section Power
Consumption on page 598.
TA = ambient temperature (C).
From the first equation, the user can derive the estimated lifetime of the chip and decide if a
cooling device is necessary or not. If a cooling device is to be fitted on the chip, the second
equation should be used to compute the resulting average chip-junction temperature TJ in C.
626 AT91RM9200
1768BATARM08/03
AT91RM9200
Package Drawings
Figure 278. 208-lead PQFP Package Drawing
C C1
627
1768BATARM08/03
Figure 279. 256-ball BGA Package Drawing
628 AT91RM9200
1768BATARM08/03
AT91RM9200
629
1768BATARM08/03
630 AT91RM9200
1768BATARM08/03
AT91RM9200
Document Details
Title AT91RM9200 Datasheet
Revision History
Page 62 Updated Figure 9, AT91RM9200 Debug and Test Block Diagram with corrected DTXD
and DRXD signal names and transfer direction of signals TST0 - TST1 and NRST.
Page 87 Corrected BMS state to high during reset. Corrected address for internal ROM mapping.
Page 96 Updated Table 24 with new pins used and table note.
Page 109 Code change in Table 29: Xmodem Service, first table cell.
Page 111 Code change in Table 30: DataFlash Service Methods, first table cell.
Page 116 Code change in Steps 1 and 2 in section Using the Service.
Page 245 In AIC Source Mode Register, corrected descriptions of bits PRIOR and SRCTYPE.
Page 255 Change number of programmable clocks to four. Correct oscillator speed to read 32.768
kHz.
Page 256 Updated section I/O Lines with new information on clocks.
631
1768BATARM08/03
Page 258 Updated Processor Clock and Programmable Clock Outputs descriptions. Updated
Clock Generator description.
Page 259 New Clock Generator Block Diagram, Figure 118. Section Slow Clock Oscillator Startup
Time updated.
Page 264 In section Master Clock Controller, changed references to PLLB Output to PLLB Clock.
New Figure 124: Master Clock Controller. In section Processor Clock Source, specified
differences between ARM7-based and ARM9-based systems.
Page 265 Section Programmable Clock Output Controller updated to show change in number of
programmable clocks.
Page 267 In Table 60: Clock Switching Timings (Worst Case), changed PLLA Output to PLLA
Clock and PLLB Output to PLLB Clock.
Page 268 In Figure 125: Switch Master Clock from Slow Clock to PLLA Clock and in Figure 126:
Switch Master Clock from Main Clock to Slow Clock, changed signal names and wave-
form labels.
Page 269 In Figure 127: Change PLLA Programming, changed signal names and labels. New Fig-
ure 128: Programmable Clock Output Programming.
Page 270 Changed register names in Table 61: PMC Register Mapping: PMC_MOR to
CKGR_MOR, PMC_MCFR to CKGR_MCFR, PMC_PLLAR to CKGR_PLLAR and
PCM_PLLBR to CKGR_PLLBR. Remove registers PMC_PCK4, PMC_PCK5,
PMC_PCK6 and PMC_PCK7 (addresses 0x0050 to 0x005C).
Page 271 In register PMC_SCER, deleted bits PCK7 to PCK4, fields 15 to 12. All bit names
updated to include Enable. In UHP bit description, deleted reference to 12 MHz clock.
Page 272 In register PMC_SCDR, deleted bits PCK7 to PCK4, fields 15 to 12. All bit names
updated to include Disable. In UHP bit description, deleted reference to 12 MHz clock.
Page 273 In register PMC_SCSR, deleted bits PCK7 to PCK4, fields 15 to 12. All bit names
updated to include Status. In UHP bit description, corrected to read USB Host Port.
Page 276 Changed register name to PMC Clock Generator Main Oscillator Register. MOSCEN bit
description changed to include information on Main Clock signal and crystal connection.
OSCOUNT bit description changed to remove multiplication factor for Slow Clock
cycles.
Page 277 Changed register name to PMC Clock Generator Main Clock Frequency Register. Cor-
rected in MAINRDY field description reference to MAINF.
Page 278 Changed register name to PMC Clock Generator PLL A Register. In OUTA and MULA
bits, changed references to PLLA Output to PLL A Clock.
632 AT91RM9200
1768BATARM08/03
AT91RM9200
Page 279 Changed register name to PMC Clock Generator PLL B Register. In OUTB and MULB
bits, changed references to PLLB Output to PLL B Clock. Changed bit description for
USB_96M.
Page 280 In PMC_MCKR, new clock source selections specified for CSS. MDIV bit condition
added.
Page 281 In PMC_PCK0 to PMC_PCK3, new clock source selections specified for CSS.
Page 282 In PMC_IER and PMC_IDR, bits PCK7RDY, PCK6RDY, PCK5RDY and PCK4RDY
removed.
Page 283 In PMC_SR, bits PCK7RDY, PCK6RDY, PCK5RDY and PCK4RDY removed.
Page 284 In PMC_IMR, bits PCK7RDY, PCK6RDY, PCK5RDY and PCK4RDY removed.
Page 331 In DBGU Chip ID Register, corrected NVPTYP field to 000 for ROM.
Page 343 In Table 67: PIO Register Mapping, PIO_OWSR access changed to read-only.
Page 368 Changed all references from CPHA to NCPHA. Updated Figures 159 and 160 for clarity.
Page 391 In CHDIV and CLDIV bit descriptions in register TWI_CWGR, corrected equations for
calculation of SCL high and low periods. In CHDIV, CLDIV and CKDIV bit descriptions in
register TWI_CWGR, SCL replaced by TWCK.
Page 452 Updated Figure 214, Transmit Frame Format in Continuous Mode. Updated Figure 215,
Receive Frame Format in Continuous Mode.
Page 596 In Table 109, DC Characteristics, changed conditions for Static Current.
Page 598 New consumption figures in Table 113 and Table 114.
Page 599 In Table 115: 32 kHz Oscillator Characteristics, VDDOSC defined in Startup Time condi-
tions. In Table 116: Main Oscillator Characteristics, V DDPLL defined in Startup Time
conditions. In Table 117: Phase Lock Loop Characteristics, corrected errors in Pump
current max/min values.
Page 601 In Table 120: Switching Characteristics in Full Speed, min/max values for Rise/Fall Time
Matching added.
Page 614 In Table 125: SDRAMC Signals, changed min values for SDRAMC23 to SDRAMC28.
633
1768BATARM08/03
634 AT91RM9200
1768BATARM08/03
AT91RM9200
i
1768BATARM08/03
Peripheral Identifiers ........................................................................................ 28
System Interrupt............................................................................................. 29
External Interrupts.......................................................................................... 29
Product Memory Mapping................................................................................ 30
External Memory Mapping ............................................................................. 30
Internal Memory Mapping .............................................................................. 31
Peripheral Mapping ........................................................................................ 32
Peripheral Implementation............................................................................... 34
USART ........................................................................................................... 34
Timer Counter ................................................................................................ 34
ii AT91RM9200
1768BATARM08/03
AT91RM9200
Boot Program................................................................................................. 85
Overview............................................................................................................ 85
Flow Diagram .................................................................................................... 86
Bootloader......................................................................................................... 87
Valid Image Detection .................................................................................... 88
Structure of ARM Vector 6 ............................................................................. 89
Bootloader Sequence..................................................................................... 90
Boot Uploader ................................................................................................... 94
External Communication Channels ................................................................ 94
Hardware and Software Constraints............................................................... 96
iii
1768BATARM08/03
Required Features for the Reset Controller ................................................. 121
iv AT91RM9200
1768BATARM08/03
AT91RM9200
v
1768BATARM08/03
Burst Flash Interface .................................................................................... 211
Product Dependencies................................................................................... 212
Supported Burst Flash Devices.................................................................... 212
I/O Lines....................................................................................................... 212
Functional Description................................................................................... 212
Burst Flash Controller Reset State............................................................... 212
Burst Flash Controller Clock Selection......................................................... 212
Burst Flash Controller Asynchronous Mode................................................. 213
Burst Flash Controller Synchronous Mode .................................................. 215
Burst Flash Controller (BFC) User Interface ................................................ 221
Burst Flash Controller Mode Register .......................................................... 221
vi AT91RM9200
1768BATARM08/03
AT91RM9200
vii
1768BATARM08/03
Power Management Controller (PMC) User Interface ................................ 270
PMC System Clock Enable Register............................................................ 271
PMC System Clock Disable Register ........................................................... 272
PMC System Clock Status Register............................................................. 273
PMC Peripheral Clock Enable Register ....................................................... 274
PMC Peripheral Clock Disable Register ...................................................... 274
PMC Peripheral Clock Status Register ........................................................ 275
PMC Clock Generator Main Oscillator Register ........................................... 276
PMC Clock Generator Main Clock Frequency Register ............................... 277
PMC Clock Generator PLL A Register ......................................................... 278
PMC Clock Generator PLL B Register ......................................................... 279
PMC Master Clock Register ......................................................................... 280
PMC Programmable Clock Register 0 to 3 .................................................. 281
PMC Interrupt Enable Register .................................................................... 282
PMC Interrupt Disable Register ................................................................... 282
PMC Status Register.................................................................................... 283
PMC Interrupt Mask Register ....................................................................... 284
viii AT91RM9200
1768BATARM08/03
AT91RM9200
ix
1768BATARM08/03
Debug Unit Status Register.......................................................................... 326
Debug Unit Receiver Holding Register ........................................................ 328
Debug Unit Baud Rate Generator Register.................................................. 329
Debug Unit Chip ID Register ........................................................................ 330
Debug Unit Chip ID Extension Register ....................................................... 332
Debug Unit Force NTRST Register.............................................................. 332
x AT91RM9200
1768BATARM08/03
AT91RM9200
xi
1768BATARM08/03
Modes of Operation...................................................................................... 383
Transmitting Data......................................................................................... 383
Read/Write Flowcharts................................................................................. 385
Two-wire Interface (TWI) User Interface ...................................................... 388
TWI Control Register.................................................................................... 389
TWI Master Mode Register .......................................................................... 390
TWI Internal Address Register ..................................................................... 391
TWI Clock Waveform Generator Register.................................................... 391
TWI Status Register ..................................................................................... 392
TWI Interrupt Enable Register...................................................................... 393
TWI Interrupt Disable Register ..................................................................... 394
TWI Interrupt Mask Register ........................................................................ 395
TWI Receive Holding Register ..................................................................... 396
TWI Transmit Holding Register .................................................................... 396
xii AT91RM9200
1768BATARM08/03
AT91RM9200
xiii
1768BATARM08/03
Interrupt........................................................................................................ 475
Functional Description................................................................................... 475
TC Description ............................................................................................. 475
Capture Operating Mode.............................................................................. 478
Waveform Operating Mode ............................................................................ 480
Timer Counter (TC) User Interface ................................................................ 487
TC Block Control Register............................................................................ 488
TC Block Mode Register .............................................................................. 488
TC Channel Control Register ....................................................................... 489
TC Channel Mode Register: Capture Mode................................................. 490
TC Channel Mode Register: Waveform Mode ............................................. 492
TC Counter Value Register .......................................................................... 495
TC Register A............................................................................................... 495
TC Register B............................................................................................... 495
TC Register C .............................................................................................. 496
TC Status Register....................................................................................... 496
TC Interrupt Enable Register ....................................................................... 498
TC Interrupt Disable Register....................................................................... 499
TC Interrupt Mask Register .......................................................................... 500
xiv AT91RM9200
1768BATARM08/03
AT91RM9200
xv
1768BATARM08/03
Product Dependencies................................................................................... 567
I/O Lines....................................................................................................... 567
Power Management ..................................................................................... 567
Interrupt........................................................................................................ 567
Functional Description................................................................................... 568
Media Independent Interface ....................................................................... 569
Transmit/Receive Operation ........................................................................ 570
Frame Format Extensions ............................................................................ 571
DMA Operations........................................................................................... 572
Address Checking ........................................................................................ 574
Ethernet MAC (EMAC) User Interface .......................................................... 575
EMAC Control Register................................................................................ 577
EMAC Configuration Register ...................................................................... 578
EMAC Status Register ................................................................................. 580
EMAC Transmit Address Register ............................................................... 581
EMAC Transmit Control Register ................................................................. 582
EMAC Transmit Status Register .................................................................. 583
EMAC Receive Buffer Queue Pointer Register............................................ 584
EMAC Receive Status Register ................................................................... 585
EMAC Interrupt Status Register................................................................... 586
EMAC Interrupt Enable Register .................................................................. 587
EMAC Interrupt Disable Register ................................................................. 588
EMAC Interrupt Mask Register .................................................................... 589
EMAC PHY Maintenance Register .............................................................. 590
EMAC Hash Address High Register ............................................................ 591
EMAC Hash Address Low Register ............................................................. 591
EMAC Specific Address (1, 2, 3 and 4) High Register ................................. 592
EMAC Specific Address (1, 2, 3 and 4) Low Register.................................. 592
EMAC Statistics Register Block Registers ................................................... 593
xvi AT91RM9200
1768BATARM08/03
AT91RM9200
xvii
1768BATARM08/03
Atmel Corporation Atmel Operations
2325 Orchard Parkway Memory RF/Automotive
San Jose, CA 95131, USA 2325 Orchard Parkway Theresienstrasse 2
Tel: 1(408) 441-0311 San Jose, CA 95131, USA Postfach 3535
Fax: 1(408) 487-2600 Tel: 1(408) 441-0311 74025 Heilbronn, Germany
Fax: 1(408) 436-4314 Tel: (49) 71-31-67-0
Fax: (49) 71-31-67-2340
Regional Headquarters Microcontrollers
Europe 2325 Orchard Parkway 1150 East Cheyenne Mtn. Blvd.
Atmel Sarl San Jose, CA 95131, USA Colorado Springs, CO 80906, USA
Route des Arsenaux 41 Tel: 1(408) 441-0311 Tel: 1(719) 576-3300
Case Postale 80 Fax: 1(408) 436-4314 Fax: 1(719) 540-1759
CH-1705 Fribourg
Switzerland La Chantrerie Biometrics/Imaging/Hi-Rel MPU/
Tel: (41) 26-426-5555 BP 70602 High Speed Converters/RF Datacom
Fax: (41) 26-426-5500 44306 Nantes Cedex 3, France Avenue de Rochepleine
Tel: (33) 2-40-18-18-18 BP 123
Asia Fax: (33) 2-40-18-19-60 38521 Saint-Egreve Cedex, France
Room 1219 Tel: (33) 4-76-58-30-00
Chinachem Golden Plaza ASIC/ASSP/Smart Cards Fax: (33) 4-76-58-34-80
77 Mody Road Tsimshatsui Zone Industrielle
East Kowloon 13106 Rousset Cedex, France
Hong Kong Tel: (33) 4-42-53-60-00
Tel: (852) 2721-9778 Fax: (33) 4-42-53-60-01
Fax: (852) 2722-1369
1150 East Cheyenne Mtn. Blvd.
Japan Colorado Springs, CO 80906, USA
9F, Tonetsu Shinkawa Bldg. Tel: 1(719) 576-3300
1-24-8 Shinkawa Fax: 1(719) 540-1759
Chuo-ku, Tokyo 104-0033
Japan Scottish Enterprise Technology Park
Tel: (81) 3-3523-3551 Maxwell Building
Fax: (81) 3-3523-7581 East Kilbride G75 0QR, Scotland
Tel: (44) 1355-803-000
Fax: (44) 1355-242-743
Literature Requests
www.atmel.com/literature
Disclaimer: Atmel Corporation makes no warranty for the use of its products, other than those expressly contained in the Companys standard
warranty which is detailed in Atmels Terms and Conditions located on the Companys web site. The Company assumes no responsibility for any
errors which may appear in this document, reserves the right to change devices or specifications detailed herein at any time without notice, and
does not make any commitment to update the information contained herein. No licenses to patents or other intellectual property of Atmel are
granted by the Company in connection with the sale of Atmel products, expressly or by implication. Atmels products are not authorized for use
as critical components in life support devices or systems.
Atmel Corporation 2003. All rights reserved. ATMEL and combinations thereof and DataFlash are the
registered trademarks of Atmel Corporation or its subsidiaries.
ARM , ARM7TDMI and Thumb are the registered trademarks and ARM9TDMI , ARM920T and AMBA
are the trademarks of ARM Ltd.; CompactFlash is a registered trademark of the CompactFlash Association;
SmartMedia is a trademark of the Solid State Floppy Disk Card Forum.
Printed on recycled paper.
Other terms and product names may be the trademarks of others.
1768BATARM08/03