Introduction to CDMA engineering

What is CDMA?

CDMA is a Direct Sequence (DS) Spread Spectrum Modulation Technique, where in, a narrow band (of less than 10 kHz) user information (voice or data) is transmitted over a wide RF bandwidth (of 1.23 MHz); with many users occupying the same RF transmission band but are separated by what are called codes.

CDMA uses orthogonal codes called Walsh Codes to distinguish each individual user in the system.

CDMA was developed by Qualcomm in the US. This standard quickly found acceptance in Korea, Japan and many other Asian countries including India.

Comparison of Information and Transmission Bandwidth

Time Domain Frequency Domain

n/t

n/t

Wide (large time duration) pulses in the time domain result in energy being transmitted over a narrow frequency range
Shorter pulses (used in CDMA PN spreading sequences) result in energy being transmitted over a wide range of frequencies

Noise in Narrow Band and Spread Spectrum Communication Systems

-40

-60

Power (in dBm)

-80

RSL _ _ _ Wide band Noise Floor Narrow band Noise Floor

-100

-120

12

16

20

24

28

32

Distance in Miles

Some concepts in CDMA:

* Spread Spectrum Modulation -- Narrow band information is transmitted over a wide band RF channel * N=1 Frequency Reuse -- Multiple users (in adjacent cells) operate on the same frequency. * Code Division Access -- Each user and base station is associated with a unique code rather than a frequency or time slot. * Coherent Reception -- mobile units coherently combine multi-path components and signals from multiple base stations. * Dynamic Power Control -- Forward and reverse link transmit power is controlled to the minimum require to reduce interference and achieve a given capacity. * Variable Rate Speech Encoding -- Voice is encoded at a slower rates based on when the user is not speaking in order to minimize transmitted power and system interference

Advantages of CDMA
Frequency Re-use: Higher Capacity is achieved with CDMA's inherent N=1 frequency reuse pattern. This is different from the typical AMPS N=7 frequency reuse pattern in which only on seventh of the available frequencies are used in a given cell.

Coherent Signal Reception: CDMA has the ability to coherently combine signals from multiple sources. This multiple correlation system employs a RAKE receiver, which combines signals arriving at a given location, with different time delays, thus taking advantage of multi-path. In addition, this feature allows the mobile receiver to use signals from multiple base station transmitters, thus improving cell boundary performance and minimizing dropped calls.

User Privacy: CDMA's spread spectrum modulation technique distributes the user information over an RF bandwidth that is much larger than the information bandwidth. The resulting power spectral density (PSD) of the transmitted signal resembles thermal noise making the signal very difficult to be detected by simple receivers. In addition, a unique address code is required to recover user information.

2nd Generation IS-95 A/B CDMA Standards and now 1X

CDMA as described in this document is based on a standard known as IS (Interim Standard) 95. IS95 is the Mobile Station Base Station Compatibility Standard for DualMode Wideband Spread Spectrum Cellular. IS95 is also known as the CDMA Air Interface specification.

Term cdmaOne has been adopted by CDG as a designator for CDMA technology based on IS95 and accompanying standards. CDMA air interface for PCS applications is described in J-STD-008. The basic CDMA process is the same in all standards

Processing Gain
Processing Gain is a term common to all direct sequence spread spectrum systems. Process gain is defined as the ratio of the Chip Rate (Rc) to the information bit rate (Rb). This provides a measure of ``spreading'' in the system.

Processing Gain = Rc / Rb
Chip Rate (Rc): The Chip Rate is the rate at which the PN sequence is generated. For CDMA, IS95, the chip rate is 1.2288 * 10 ^ 6 cps (chips per second).
Bit Rate (Rb): The bit rate is base band user information (i.e. user voice/data) rate. In CDMA, voice is digitized at different rates depending on the speech activity level. The system parameters presented in this discussion are based on a maximum bite rate of 9.6 kbps and 14.4 kbps per IS95 For CDMA (IS95A/B): Rc = 1.2288 Mcps, Rb = 14.4 kbps (max), resulting in a Processing Gain of 85.33 (19.3 dB).

Some Terminology

Eb/No = Ratio of Transmitted energy per bit (Eb) to Thermal Noise (No), expressed in dB.

Eb/Nt = Ratio of Transmitted energy per bit (Eb) to Total Noise (Nt) including thermal, spurious, and interference from other CDMA users, expressed in dB.

Ec/Nt = Ratio of Transmitted energy per chip (Ec) to Total Noise (Nt), expressed in dB.

Ec/Io = Ratio of Transmitted energy per chip (Ec) to Total Noise including selfinterference (Io) usually expressed in dB

BER (Bit Error Rate) = Probability that a transmitted bit will be received incorrectly (i.e. 1 received as a 0 or a 0 received as a 1)

FER (Frame Error Rate) = Probability that a transmitted frame will be received incorrectly

Summary of Codes
PN Long Code
The Long Code is a PN sequence that is 2^42 1 bits (chips) long. It is generated at a rate of 1.2288 Mbps (or Mcps) giving it a period (time before the sequence repeats) of approximately 41.4 days. The long code is used to encrypt user information. Both the base station and the mobile unit have knowledge of this sequence at any given instant in time based on a specified private ``long code mask'' that is exchanged. PN Short Code The Short Code is a PN sequence that is 2 ^ 15 bits (chips) in length. This code is generated at 1.2288 Mbps (or Mcps) giving a period of 26.67 ms. This code is used for final spreading of the signal and is transmitted as a reference known as the ``Pilot Sequence'' by the base station. All base stations use the same short code. Base stations are differentiated from one another by transmitting the PN short code at different ``offsets'' in absolute. Walsh Codes CDMA defines a group of 64 orthogonal sequences, each 64 bits long, known as Walsh Codes. These sequences are also referred to as Wash Functions. These codes are generated at 1.2288 Mbps (Mcps) with a period of approximately 52 s. These are used to identify users on the forward link. For this reason they are also referred to as either Walsh Channels or TCH. All base stations and mobile users have knowledge of all Walsh codes.

FER to Eb/Nt
10^0 10^-1 10^-2
Good Call Quality

Reverse Link FER Performance

Probab -ility of Frame Error

10^-3 10^-4 10^-5 10^-6

Forward Link FER Performance

-3

-1

11

Average Eb / Nt

CDMA Physical and Logical Channels

Physical Channel Physical channels are described in terms of a wideband RF channel and code sequence. As defined in IS95, each RF channel is 1.2288 MHz wide. For each RF channel, there are 64 Walsh sequences (W0 through W63) available for use on the forward link. These Walsh sequences are commonly referred to as CDMA code channels.

Logical Channel The physical channel that carry specific types of information are known as logical channels. Logical channels in CDMA are divided into two categories: Traffic Channels and Control Channels. For the forward link there are three types of Control/Signaling channels and one Traffic Channel (per user). For the Reverse Link there is one type Signaling Channel and one Traffic Channel per user.

It is important to note that signals on the forward link are identified by Walsh codes, however, signals on the reverse link are identified by Long Codes.

Forward Link (Downlink)

The logical channels for the Forward Link must provide identification of the Base station, timing and synchronization of the transmissions between the base station and mobile station, paging of mobile units in the area, and the voice/data transmission from the base station to the mobile unit. The forward link is comprised of: The Pilot Channel, Up to one Sync Channel, Up to seven Paging Channels, and Up to 55 Traffic Channels

Power Control SubChannel

A Power Control SubChannel is continuously transmitted on the forward traffic channel as part of the traffic frame. Information on this channel commands the mobile unit to adjust its transmitted power + 1 dB every 1/16 of a speech frame (800 times per second).

Pilot and Sync and Paging Channels

Pilot Channel
The Pilot Channel allows a mobile station to acquire the timing of the Forward Traffic Channel user information. It provides a phase reference for coherent demodulation and provides a means for signal strength comparisons between base stations, which is used to determine when to handoff. It consists of the un-modulated spreading sequences (PN short codes). The Pilot signal is transmitted continuously on Walsh 0 by each CDMA base station at the transmitter (cell/sector) level.

Sync Channel
The Synchronization Channel is an encoded, interleaved and modulated spread spectrum signal that is used with the Pilot Channel to acquire initial system time and synchronization. The sync channel is always transmitted on Walsh 32.

Paging Channel
The Paging Channel is used for transmission of control information to the mobile. When a mobile is to receive a call it will receive a ``page'' from the base station. Up to seven (7) channels may be configured for paging depending on the expected demand. Page channel messaging to each user takes place in an 80 ms ``slot''. The 80 ms slots are grouped into cycles of 2048 slots (cycle duration 163.84 s) referred to as maximum slot cycles. The base station can limit the maximum slot cycle used by the mobile.

Reverse Link (Uplink)

The logical channel requirements of the reverse link must provide for the identification and access request by the mobile unit to the base stations in the area and the voice/data transmission from the mobile unit to the base station. The reverse link is composed of: Access Channels and Traffic Channels. These channels share the same CDMA center frequency on the reverse link (a different frequency is used for forward link transmissions). The total number of channels( max 55) is determined by base station activity. .. The reverse link capability of a given base station is limited by the number of traffic channels assigned (up to 55) and up to seven (7) access channels (correlating to a maximum of 7 paging channels). Note that a mobile does not ``tie up'' an access channel, it only borrows it for a short amount of time.

Access and Traffic Channels

Access Channel
The Access Channel is used for the transmission of control information to the base station. When a mobile is to place a call it uses the ``access'' channel to inform the base station. This channel is also used when responding to a ``page''. Each Access Channel is identified by a distinct ``Access Channel Long PN Code ''. An Access Channel is selected randomly by the mobile unit from the total number of access channels available from the serving cell/sector.

Traffic Channel
The Traffic Channel carries all the calls (voice or data signal) from a given base station to all the mobile units active in the coverage area or vice versa. Each user has a dedicated TCH, and corresponding Walsh code, on the down link. The forward traffic channel message consists of user voice (or data), power control data, and error correction bits. The message is transmitted as a series of traffic frames. The traffic channel may also carry signaling information with or in place of user voice (or data). A Walsh code is assigned by the base station for each Traffic Channel in use. The Traffic Channel for the reverse link is identical to the forward link Traffic Channel in function and structure. Each traffic channel is identified by a ``User Long PN Code'' which is unique to each CDMA user.

Types of Spread Spectrum Modulation

The types of spread spectrum modulation commonly used in communication systems are classified as: Direct Sequence Frequency Hopping CDMA is a direct sequence system.
Direct Sequence

In direct sequence modulation the carrier frequency is fixed and the bandwidth of the transmitted signal is larger and independent of the bandwidth of the information signal.

Frequency Hopping

The carrier frequency is varied and the bandwidth of the transmitted signal is comparable to the bandwidth of the information signal. Information is modulated on top of a rapidly changing carrier frequency.

Transmit and Receive Processes of Spread Spectrum

Transmitting a spread spectrum signal involves:

1. Modulating the information signal with the spreading PN sequence 2. Modulating the resulting signal with the desired carrier wave 3. Band Pass Filtering the output 4. Transmitting the resulting RF signal.

Receiving a spread spectrum signal involves: 1. Demodulating the signal with the RF carrier, 2. Low Pass Filtering the resulting wide band signal, 3. Demodulating with the signal with the known spreading sequence, and 4. Integrating the despread signal over a bit time to recover the information signal

Multipath
Multipath, is the result of reflections and scattering of radio waves off of buildings, mountains and other objects, etc. Multipath will exist anywhere the incident wave and one or more reflected and/or defracted waves can reach the receiver Multipath, in effect, creates ``multiple versions'' of the transmitted signal which arrive at the receiver at different times When the components arrive perfectly in phase, the overall Received Signal Level (RSL) will be stronger than any of the individual components. When they arrive out of phase, as a result of the reflective / defractive process, the overall received level is less than the strongest individual component. Destructive (and constructive) interference due to the arrival of equal amplitude and random phase multipath components is referred to as Rayleigh Fading. The significance or degree that Rayleigh Fading affects system operation is determined by the surrounding environment. In general we would expect to see the greatest effects of Rayleigh fading in the Dense Urban environment and the least in a Rural Environment. This is due to the greater concentration of scattering structures in a Dense Urban Environment than in rural areas.

Destructive Interference due to Multipath

Relative Amplitude

Time

The RAKE Receiver

The RAKE receiver is an optimum demodulator structure for channels encountering multipath propagation paths in a land mobile telephone environment. Essentially the receiver has the capability of ``looking'' at a given window in time, picking out multipath components of a given signal and lining them up so that they are in phase again. This process is referred to as coherent addition and results in a greater probability of making or maintaining the forward link received level at the mobile. The RAKE receiver is also applied to reverse link, however, because of a lack of a coherent reference (pilot signal) the reverse link uses a noncoherent RAKE demodulator.

The RAKE receiver provides for the coherent combination of multipath components from a single base station and multiple cells/sectors jointly in a CDMA Handoff scenario

The RAKE receiver performs the following functions:

Identifies the components which are the strongest (maximum of three), Performs time alignment of the select components, and Sums the components.

Coherent Combination of Three Strongest Multipath Components from a Single Transmitter

Magnitude of Coherently Combined Multipath

A4

h(t)
A2

Received Signals

A2

A4 A1

A5
A3 A1

Time

Coherent Combination of Three Strongest Components of a Multiple Transmitter BandLimited Channel Impulse Response with Discrete Multipath Components
Magnitude of Coherently combined Multipaths

A3

A3 Relative Power

B3

C2 B3

C2

Time

Effects of Multipath on FDMA, TDMA, and CDMA

The quality of service provided by a FDMA System is a function of the received signal level and proper frequency planning. Assuming no frequency reuse or the assignment of adjacent channels within the system, the problem becomes one dimensional as a function of signal strength. In FDMA, the carrier wave is subjected to the multipath fading (Rayleigh fading) as discussed above and Multipath may greatly impact voice quality. Multipath in a TDMA adversely effects the performance in two ways that must be compensated for in the design and implementation of the system. First, multipath fading of the carrier results in reduced signal strength. The reduction in signal strength results in increased bit error rate as E b /N t falls below what is required for acceptable call quality. The second effect of multipath, the time delay in arrival over which multipath components arrive (delay spread), can be large enough to create Inter Symbol Interference (ISI). This effect is known as time dispersion. ISI may result in false 1's when a zero is sent or visa versa.

CDMA is less susceptible to the adverse effects of multipath. Specifically, degradation of system performance due to frequency selective Rayleigh fading is reduced because the signal energy is spread over 1.25 MHz as opposed to 200 kHz for GSM TDMA or 30kHz channels for AMPS TDMA. In effect spread spectrum is also a form of frequency diversity. Also, by coherently combining multipath components, the effects of time dispersion and destructive interference are reduced by aligning the three strongest multipath components in time and combining for a net positive sum greater than any individual component.

RAKE Receiver Example: Improvement in Call Quality (Eb/Nt)

Let's suppose the minimum acceptable Eb/Nt for a given system is 7 dB and the three strongest multipath components have a corresponding Eb/Nt of 5, 3, and 2 dB. It is clear that the individual components will not provide for service. However, the RAKE receiver will coherently combine these components such that the resultant Eb/Nt is sufficient to provide the quality of service we desire. This can be accomplished as follows: Linearize the respective Eb/Nt measurements so as to allow for coherent combination by using the equation given below: Linear Value = 10 ^ [ 0.1 x (Eb/Nt) in dB] The linearized values for each of the multipath components are 3.16, 2.00, 1.58 respectively. Assuming perfect phase alignment and zero processing losses, the combined value for all of the components is 6.74 which corresponds to a calculated E b /N t of 8.29 dB which provides higher level of call quality.

The Near-Far Problem and Dynamic Power Control

The ``nearfar'' problem in spreadspectrum systems relates to the problem of very strong signals at a receiver swamping out the effects of weaker signals located on the edge of the coverage area in a CDMA system resulting in heavy interference.

Overcoming ``nearfar'' in the mobile environment requires that the reverse link transmit power for all subscribers be continuously adjusted. The rate and degree of adjustment should depend on the maximum rate and magnitude of change in required power to maintain a constant received level at the base station. This is accomplished through the implementation of dynamic power control.

Dynamic Power Control (DPC) is used on the forward and reverse links to avoid the near-far problem. Using DPC the power of all mobile units is controlled so their transmitted signals arrive at the base station at an equal and minimum received power level. In addition, the traffic channel power on the forward link is varied as a function of voice coding rate. In this way, the interference generated from one mobile unit to another is kept to a minimum resulting in increased system capacity.

Reverse Link Power Control Mechanism

Open Loop Power Control:
Dynamic Power Control (DPC) is used on the forward and reverse links to avoid the near-far problem. Using DPC the power of all mobile units is controlled so their transmitted signals arrive at the base station at an equal and minimum received power level. In addition, the traffic channel power on the forward link is varied as a function of voice coding rate. In this way, the interference generated from one mobile unit to another is kept to a minimum resulting in increased system capacity.

Closed Loop Power Control:

Closedloop power control is used to allow the power from the mobile unit to deviate from the nominal as set by openloop control. The base station monitors the power received from each mobile station and commands each mobile unit to raise or lower its power by a fixed step (nominally 1 dB) to keep the received signal at the minimum acceptable level. Acceptable signal is defined by < 1% to 2% FER. This process is repeated 800 times per second, or every 1.25 ms. This is accomplished by dividing each 20 ms traffic frame into 16 power control groups. Each power control group is preceded by a power control bit. Mobile units support a dynamic range of about 80 dB and can be controlled to transmit as low as 60 dBm.

Forward Link Power Control

Forward traffic channel (TCH) power is attenuated (for each TCH) based on voice coding rate that is being used. As the data rate is lowered, the output signal is attenuated. This provides a constant Eb for the output signal.

Vocoder Rate 1 1/2 1/4 1/8

Data Rate (Rb) kbps (per IS-95) 9.6 4.8 2.4 1.2

Attenuation Level (dB) 0 3 6 9

In addition, the available base station transmit power is divided among the pilot, sync, paging, and traffic channels in use.

Base Station Nominal Channel Power Allocation

Logical Channel Relative Power Allocation Nominal Allocation 20 % 80 % 3%

Pilot Sync + Paging + Traffic Sync

0.2 of total power (linear) Remainder (0.8) of total power (linear) 3 dB less than one Traffic Channel; always 1/8 rate 3 dB greater than one Traffic Channel; full rate only Equal power in each Traffic Channel: full rate only (or specified maximum per TCH)

Paging

2%

Traffic

75 %

These allocations are not dynamic with time but may be adjusted on a per transmitter basis as necessary by the operator.

CDMA Forward Link

Variable Low Bit Rate Speech Coding Variable Low Bit Rate Speech Decoding

Transmit Path in Base Station

Channel Coding

Channel Decoding

Receive path in Mobile

Bit Interleaving

Bit Deinterleaving

Encryption: Long Code Scrambling

Decryption: Long Code Descrambling

Walsh Function Modulation

Walsh Function Demodulation

Quadrature Spreading and Multiplexing

Quadrature Despreading

Quadrature Carrier Modulation RF Channel

Quadrature Carrier Demodulation

Variable Rate Speech Coding

When voice is transmitted over the commercial telephone system (land line) it is assumed to be band limited to the frequency range of 300 to 3300 Hz.. In CDMA speech is sampled at 8 kHz and uniformly quantized to 13 bits. This data is divided into 20 ms frames and transmitted at 104 kbps. The first step in the Speech Coding process is to transcode and rate adapt (modify the quantization and data rate) to the 104 kbps bit stream. Note that transcoding is not required by mobile unit for the reverse link voice transmissions. The transcoded data is then fed to the CodeExcited Linear Predictive (CELP) coder The CELP speech encoder produces a variable output data rate based on speech activity. The encoder generates one frame (a.k.a. packet, a.k.a. block) every 20 ms. The coded data frame is at one of the following data rates: 1. Rate 1: 171 bits / packet (8.55 kbps) 2. Rate 1/2: 80 bits / packet (4.0 kbps) 3. Rate 1/4: 40 bits / packet (2.0 kbps) 4. Rate 1/8: 16 bits / packet (0.8 kbps) The advantage of using lower bit rates when there is little or no speech activity is that it allows the transmit power to be decreased while maintaining a constant Eb/Nt. Reduction in transmit power decreases the level of interference imposed on other users of the system.

Walsh Function Modulation

The data transmitted on the forward link is spread over a wide bandwidth by modulating it with a Walsh function at a fixed rate of 1.2288 Mbps. There are 64 orthogonal Walsh functions Standard assignments are: Pilot channel: Walsh 0 Sync Channel: Walsh 32 Page Channel: Walsh i, i = 1 to 7, Traffic Channel: Walsh i, i = 8 up to 63 (except 32) The specific Walsh function on to which the data is modulated defines the forward link channelization.

Forward Link Base Station Transmit Power Control

Voice data is coded at varying rates based on the level of speech activity. The base station seeks to transmit signals at a constant Energy per Bit (Eb). Since the bit rate is varying, data coded at a high rate (Rate 1) must be transmitted at a higher power than data coded at a lower rate (e.g. Rate 1/8) in order to maintain a constant Eb. Forward link transmit power control accomplished using a Variable Attenuator which is implemented immediately following Walsh function modulation.

Voice Coding Rate 1 1/2

Data Rate (Rb) kbps (per IS-95) 9.6 4.8

Base Station Transmit Power Attenuation Level (dB) 0 3

1/4
1/8

2.4
1.2

6
9

CDMA Reverse Link

Variable Low Bit Rate Speech Coding Variable Low Bit Rate Speech Decoding

Channel Coding

Channel Decoding

Transmit Path in Mobile

Bit Interleaving Bit Deinterleaving

Receive path in Base Station

64ary Orthogonal Walsh Symbol Modulation

64ary Orthogonal Walsh Symbol Demodulation

Encryption: Long Code Spreading

Decryption: Long Code Despreading

Quadrature Spreading

Demultiplexing and Quadrature Despreading

Quadrature Carrier Modulation RF Channel

Quadrature Carrier Demodulation

Variable Low Bit Rate Speech Coding

The speech coding process at the mobile unit is the same as that used at the base stations except that transcoding is not required. At the mobile unit, filtered analog speech is sampled at 8000 times per second and linearly quantized at 13 bit resolution. This 104 kbps data stream is then fed to the CodeExcited Linear Predictive (CELP) Speech Coder. The CELP speech encoder produces a variable output data rate based on speech activity. The coder generates one frame, or packet, every 20 ms

LPF

A/D

CELP Speech Encoder

The available output rates are: Rate 1: 171 bits / packet (8.55 kbps), Rate 1/2: 80 bits / packet (4.0 kbps), Rate 1/4: 40 bits / packet (2.0 kbps) and Rate 1/8: 16 bits / packet (0.8 kbps)

CDMA Capacity
CDMA technology offers a significant capacity advantage over other multiple access systems. The capacity of FDMA and TDMA systems is limited by the finite amount of spectrum allocated. CDMA is different in that many users operate on a single wideband RF carrier. This carrier frequency may be reused by the adjacent cell (N=1 reuse). CDMA capacity is only interference limited, therefore any reduction in interference converts directly and linearly into an increase in capacity Interference is introduced from several sources including: Same cell mobile users, Adjacent cell mobile users, Adjacent cell base stations, as well as Thermal and spurious noise.

CDMA employs several techniques to reduce these interference sources including: Suppressing or squelching transmissions during quiet periods of each speaker. Using sectored base station antennas. Dynamic power control to keep transmit levels to the minimum required to close the link.

CDMA Capacity equation

For the simplest case of the single CDMA cell site, the approximate capacity in terms of number of users can be written as: N = 1 + { (W / R) / (Eb / No)}- ( / S) Where: W = Spread Spectrum bandwidth (Hz) R = Information bit rate (Hz) Eb = Energy per bit (J) No = System (thermal) noise energy (J) N = Number of users S = Received power of user signals at the base station (Watts) (not including serving signal) = Received background noise level at the base station (Watts)

Some Definitions
Sectorization Gain

A common technique for reducing interference is sectorization at the base station. Sectorization refers to using directional antennas at the cell site for both receiving and transmitting. For a three sectored cell site, the number of interferers seen by any antenna is, theoretically, a third of the number seen by an omni directional antenna.
Voice Activity Factor

Studies have shown that either speaker is active only 35% to 40% of the time when in conversation. The percentage of time that a user is actually speaking is called Voice Activity Factor (VAF). CDMA uses a variable rate voice encoder that monitors the voice activity level and suppresses transmission when no speech is taking place. This process reduces the level of interference introduced into the system
Frequency Reuse Efficiency (IADJ)

The Frequency Reuse Efficiency, (IADJ) accounts for the interference caused by other mobile units as well as base stations in the surrounding cells. Dynamic power control is used on the forward and reverse links to minimize adjacent (as well as co) cell interference. IADJ is statistically dependent on the loading of adjacent cells as well as the location of users within those cells. A typical value for IADJ is given as 0.66. This value implies that a cell located in the center of a seven cell cluster is subject to a noise floor that is 160% of that which would be observed if a cell is operating in total isolation.

Pole Point Equation

The noise resulting from Sync, Paging, Pilot, and the other active calls overwhelm the processing gain, and the desired call can no longer be decoded, at any range, regardless of how high the transmit power is raised. In other words, the cell jams itself with its own cochannel (i.e. co frequency) transmissions. This number of users at which this condition occurs is known as the Pole Point. The pole point equation estimates the maximum number of traffic channels that may be assigned to a single CDMA base station (or sector) on a single carrier frequency. This equation includes the effects of sectorization, voice activity, and adjacent cell interference. # of TCHs at Pole Point = 1 + [ (W/Rb) / {(1+IADJ)(VAF)(Eb/Nt)(Gs)}] .(8.2) where: W = The Spread Bandwidth in Chips/sec = 1.2288 x 10 ^ 6 for IS95 derivatives, Rb = The Information Bit Rate = 14.4 x 10^3 bps (IS95A), (9.6 kbps for IS95) I ADJ = The additional interference contributed by adjacent cells = 0.6 VAF = Voice Activity Factor = 0.5 Eb /Nt= Minimum E b /N t required (after despreading) to provide specified voice quality, G s = Sectorization Gain = 1 for omni cells = 1.18 (that is 3/2.55) for 3sector cells

Some more Definitions

Spread Bandwidth (W) is the actual number of chips transmitted on the RF channel after the data signal is spread by a direct sequence technique, as it is in IS95 and PCS derivatives. Information Bit rate (Rb) is the channel information bit rate, including both the voice channel and system overhead bits to support a single voice channel IADJ is an adjustment factor that has been stated by equipment vendors to be the nominal amount of extra systemgenerated noise contributed by adjacent cells. This is stated as a fraction of the noise generated by the cell under consideration. In other words, 0.6 means that Nt (neglecting the No component) is 160% of that which would be observed if a cell is operating in total isolation. This factor is a function of cell loading, propagation characteristics, and voice activity factor. Voice Activity Factor (VAF) is the fraction of the time that a person is actually speaking (and transmitting fullrate data) during an average conversation. If a person spends 50% of the time talking, VAF=0.5. Minimum Eb/Nt is the Eb/Nt required to maintain a 1% frame error rate, that which has been specified as the min. acceptable to maintain call quality. This is normally expressed as linear energy ratio, not in dB. Sectorization Gain (Gs) is somewhat similar to the additional interference contributed by adjacent cells except that it is a factor to describe the noise introduced by adjacent sectors within the same cell. In other words, it is intended to adjust for the fact that sectorizing a cell does not quite increase the available number of TCHs available at a 3 sector cell by a factor of 3. For the assumptions stated previously, the # of TCHs at Pole Point = 19.09, or 19.

Handoff Terminology
Handoffs are initiated and terminated as a result of the pilot signal strength as measured by the mobile unit in terms of Ec/Nt (Energy per chip to Total Noise). The parameters associated with CDMA Handoffs are TADD, TDROP and TCOMP. TADD is the value of the Pilot signal strength, Ec/Nt, in dB received by the mobile unit at which the mobile will recognize the cell/sector as a possible contributor to the call processing activities. Values provided by vendors are typically on the order of 13 dB. TDROP is the value of the Pilot signal strength, Ec/Nt, in dB received by the mobile unit at which the mobile will drop the cell/sector as a possible contributor to the call processing activities. Values provided by vendors are typically on the order of 17 dB. Note that the received pilot strength must fall below TDROP for some specified length of time before the cell/sector is dropped in order to keep from ``toggling'' the cell on and off. This length of time (T_TDROP) is an addressable parameter with values ranging from 0.1 to 319 seconds. Note that both TADD and TDROP are assigned on a per transmitter (i.e. per cell or sector) basis. These terms need not be the same for every cell in the system. TCOMP is the Active Set versus Candidate Set comparison threshold. Mobile Stations transmit a Pilot Strength Measurement Message when the strength of a pilot in the Candidate Set exceeds that of a pilot in the Active Set by this margin. The base station shall set this field to the threshold Candidate Set pilot to Active Set pilot ratio, in units of 0.5 dB.

Pilot Search Parameters

Classification Active Set Description The pilots associated with the Forward Traffic Channels assigned to the mobile station The pilots that are not currently in the active set but have been received by the mobile station with sufficient strength to indicate that the associated Forward Traffic Channels could be successfully demodulated.

Candidate Set

Neighbor Set

The pilots that are not currently in the Active Set or the Candidate Set and are likely candidates for Handoff.
The set of all possible pilots in the current on the current CDMA frequency assignment, excluding the pilots in the Neighbor Set, the Candidate Set, and the Active Set.

Remaining Set

Types of Handoffs -1
Soft Handoff The condition where two cells are in simultaneous communication with the mobile is called Soft Handoff. Soft Handoff will continue until the pilot signal from one of the contributing cells drops below a predefined threshold (TDROP ). At that time the call will be transferred to the remaining cell. Soft Soft Handoff SoftSoft Handoffs are identical in function and process to that of the soft Handoff, however, SoftSoft Handoffs entail the simultaneous serving of a mobile unit by three cell sites. Three is the maximum number of serving signals due to mobile (RAKE) receiver specification.

Softer Handoff Softer Handoffs are identical in function and process to that of the soft Handoff, however Softer Handoffs entail the simultaneous serving of a mobile unit by two sectors of the same cell.

Types of Handoffs -2
Soft Softer Handoff Soft Softer Handoffs are identical in function and process to that of the Soft Handoff, however, SoftSofter Handoffs are the simultaneous serving of a mobile station by the original sector, an adjacent sector, and an adjacent or neighboring cell. Hard Handoff A hard Handoff occurs when a CDMA call is transferred from one base station to another base station transmitting on a different carrier frequency. Hard Handoff is analogous to the Handoff procedure that takes place in standard AMPS Cellular. The mobile unit will initially seek to perform a soft Handoff. If the cellular network cannot perform a soft Handoff, a hard Handoff is necessary CDMA to Analog Handoff

If there are no CDMA channels to Handoff to, then the call would be handed off to an available analog channel at the serving base station and switched to the analog mode of processing. From that point on, the call will be handled as any other analog call at that base station.

Soft Handoff details

The mobile station typically initiates soft Handoffs. The mobile station continuously searches for pilots to detect the presence of other CDMA signals that have the same carrier frequency and measures the strength of the pilots. When the mobile station detects a pilot of sufficient strength that is not associated with the serving cell, it sends a message to the serving base station. The cellular network decides which neighbor base stations can be involved in a Handoff and selects an idle Walsh function associated with the selected BTS, effectively selecting a traffic channel. The selected BTS is given the mobile's long code mask. The serving base station is directed to send the mobile a message to initiate Soft Handoff. In Forward link, the mobile receives information from the two forward links and uses the RAKE receiver to coherently combine the signals using the pilot sequence transmitted by each cell/sector as its reference. This combination of multiple forward link signals improves overall link performance. In the case of the Reverse link, both base stations are receiving the transmitted speech frames from the mobile. The highest quality traffic frame received from among the two base stations is selected on a framebyframe basis. During a Soft Handoff, closedloop power control of the mobile user is handled ``jointly between the serving base stations. The base stations send identical traffic frames with the exception of the power control bits. If all of the serving base stations request the mobile increase its power, the mobile will increase its output power by 1 dB (nominal). However if any one of the serving stations request a decrease in power, the mobile will drop its output by 1 dB. As with the normal closedloop power control process, adjustments are made for each power control group (1/16 of a 20 ms frame or 800 time per second).

CDMA Call Processing

The following is an example of a cellular call: set-up, progress and termination in a CDMA system. The example describes:

Initial system access Call initiation and setup, Soft Handoff, and Call termination.

Initial System Access

When the mobile is first turns on, it must find the best base station with which to communicate. The mobile unit tunes its receiver to a specified ``primary'' CDMA carrier frequency The mobile then scans for available pilot signals, which are all on different time offsets of the same PN short (2 ^ 15 chips) code The mobile selects the strongest pilot sequence and establishes frequency and time reference with this signal. If the mobile does not detect any pilot signals of adequate strength, the unit tunes its receive to another specified CDMA carrier frequency The mobile then demodulates the sync channel which is always transmitted on Walsh 32. The Sync Channel provides master clock information by sending the state of the 42 bit shift register, which generates the long (2 ^ 42 chips) code, 320 ms in the future. The long code, generated in conjunction with a private user mask, is used for encryption and decryption The mobile then starts listening to the paging channel and waits for a page directed to its phone number.

Call Initiation, Setup and Termination

The mobile user when decides to make a call, enters the desired phone number. This initiates an access probe. The mobile uses the access channel and attempts to contact the serving base station Since no traffic channel has been established, the mobile uses open loop power control. Multiple tries are allowed at random times to avoid collisions that can occur on the access channel. Each successive attempt is made at a higher power level. After each attempt, the mobile listens to the paging channel for a response from the base stations.

Once the access request has been received by the base station, the base station responds with an assignment to a traffic channel (Walsh code). The base station initiates the land link, and conversation takes place.

Call termination can be initiated either from the mobile or the land side. In either case the transmissions are stopped, the Walsh code is freed, and the land line connection is broken. The mobile unit resumes monitoring the page channel of the current serving cell.

Concepts to be kept in mind for CDMA system Engineering

The coverage provided by CDMA system is not static. As the loading on a given base station changes, the coverage provided by that base station changes inversely. Otherwise stated; just because you have great RF coverage doesn't guarantee good signal. 2. Holes in coverage may result when there is either insufficient or abundant levels of RF. System coverage is measured as the ratio of desired signal to all other signals and that the ratio can be unacceptable regardless of the absolute quantity. 3. CDMA systems allow for the nonsymmetrical simultaneous processing of a call by multiple base stations. The energy in the forward link is summed to a greater strength than the individual components. The reverse link employs the shotgun effect in that multiple base stations will receive the transmitted signal and the probability that the signal will be acceptable for at least one of them is greatly increased. 4. Traffic engineering in a CDMA system requires that in addition to all of the factors associated with engineering a FDMA or TDMA system, the element of time also be introduced. PN Offset Planning for a CDMA system requires the careful assignment of 512 available time offsets to the cells/sectors in a system

Popular Propagation Models

Most popular macroscopic propagation models are: 1. 2. 3. Lee's Propagation Model Hata Okumura Propagation Model Cost 231 Propagation Model Lee's Model is valid for 1900 MHz band. The Hatamodel is developed for frequencies from 150 and 1500 MHz band. The COST231 model is for 1500 to 2000 MHz band.

Receiver Sensitivity for Different CDMA Channel Types

Channel Type Bit Rate (Kbps) PG (dB) Quality Requirement (dB) Noise Figure (dB) Rx Sens. (dBm)

Pilot Paging

N/A 7.2 4.8

0 22.3 24.1 25.3 27.1 19.3 21.1 19.3 21.07

-14 8 8 8 8 6.5 5 7 6

8 8 8 8 8 8 8 5 5

-119 -119 -121 -122 -124 -118 -121 -120 -123

Sync

1.8 1.2

TrafficForward

14.4 9.6

Traffic- Reverse

14.4 9.6

IS95 CDMA Link Budget for InVehicle Coverage

Reverse Link
Pilot

Forward Link
Paging Sync Traffic

Reverse Link Tx Power in dBm

Forward Link Tx Power in dBm MS Antenna Gain (dBd) Human/Head Loss Cable Loss

23.01 (0.2 watts)

34.7 (2.9 watts) -2.16 -3.0 -3.0 -2.16 -3.0 -3.0 28.7 (0.74 watts) -2.16 -3.0 -3.0 25.7 (0.37 watts) -2.16 -3.0 -3.0 25.0 (0.3 watts) -2.16 -3.0 -3.0

BS Antenna Gain (dBd) In Vehicle Loss

Soft Hand-off gain Fade Margin Interference Margin (60% load) Rx Sensitivity Maximum Path Loss (dB)

15.0
-8.0 3.7 -4.3 -4.0 -123 140.24

15.0
-8.0 0 -4.3 -8.0 -119 140.24

15.0
-8.0 0 -4.3 -4.0 -121 140.24

15.0
-8.0 0 -4.3 -4.0 -124 140.24

15.0
-8.0 3.7 -4.3 -4.0 -121 140.24

Typical CDMA System Parameters

Coverage and Capacity Relationship

There is an inverse relationship between the coverage area of a given cell and the loading on that cell due to a rising of the noise floor induced by the users on that site. This phenomena results in a ``breathing'' and a ``self regulating'' communication system. The average required capacity of a given base station has to be estimated at the time of design so as to predict both the coverage it will provide and the interference it will introduce at a given average loading. On the system level, the design of a network that will provide seamless coverage at 70% theoretical loading will require a greater number of cells spaced closer together than a network designed to operate at 50% theoretical loading. A sensitivity analysis will provide the design engineer with an idea of the extent system performance will change for increases and decreases in instantaneous traffic loading. This analysis is essentially an overlay of the coverage provided at the maximum anticipated operating level placed atop the coverage provided at an average anticipated operating levels. By performing several iterations at various levels, an engineer will be able to determine the maximum average loading the system can sustain and still meet the design coverage objectives.

Comparison of Coverage due to change in traffic (5% to 80% of capacity)

PN Offset Planning
In general, PN offset planning for a CDMA system is analogous to frequency planning in an FDMA or TDMA system. For a given CDMA system, PN offset planning is a function of the same basic parameter as an AMPS channel plan such as: 1. Base Station Locations 2. Propagation Characteristics 3. Topography of the area As discussed earlier, each base station transmits a pilot signal used for acquisition, system synchronization, cell selection, and coherent demodulation of the traffic channels. All base stations transmit a unique pilot signal using the same Pseudo Random Noise (PN or PRN) spreading code (Short Code) but with different time offsets. There are a total of 512 phase offsets that are used to uniquely describe a base station. PN offsets can be reused if there is sufficient separation between cells using the same offset.

PN offsets are selected based upon the relative time delay (signal travel time at the speed of light) between sites and exact served areas of those sites. The development of a PN offset plan depends upon exact information on final site locations. There are 512 PN offsets available to allocate to cells / sectors. Each PN offset is 64 chips.

Basic PN Offset Planning Strategy

In the mobile radio environment the signal transmitted from a BS and arriving at a mobile unit will be from different paths as a result of the multipath reflection phenomenon. Since each path has a different path length, the time of arrival for each path is different.

This means that, for an impulse transmitted from the BS, by the time the impulse is received at the MS it is no longer an impulse but rather a pulse with a spread width which is referred to as the delay spread. Measured data indicates that the mean delay spread value is different for different kinds of environments. This fact is intuitive because of the increasing amount of multipath reflectors that are present in different environments

Environment Heavy Mountains Dense Urban Urban Suburban Rural

Expected Range of Delay Spread (microseconds) 1 100 6 10 4 6 2 4 .2 2

PN Interference

Since all pilot signals in a system are timeshifted versions of the same bitsequence (short code), a pilot from any sector can appear to belong to any other sector. When receiver can not distinguish pilots from different sectors, demodulation is erroneous and it is known as PN interference. There are three types of PN interference:

1. CoPN interference -- if there is no enough space separation (signal attenuation) between cells that reuse PN offset. 2. Adjacent PN offset interference -- if there is no enough separation (signal attenuation) between cells that have adjacent PN offsets (i.e. serving site has PN offset 100 and interferer has a PN offset 101) 3. Handoff confusion -- interference to a neighbor set pilot (i.e. due to time delay, strong pilot appear to be a strong neighbor list pilot: unnecessary handoff occurs)

References

[1] Rappaport, T.S., Wireless Communications, Principles and Practices, Prentice Hall, 1996. [2] Lee, W. C. Y., Overview of the Cellular CDMA, IEEE Transactions on Vehicular Technology, Vol. 40, No.2, May 1991. [3] David J. Goodman, Wireless Personal Communications Systems, Addison-Wesley Wireless Communication series, 1997.