Sram 6T
Sram 6T
Abstract - This paper deals with the design and analysis of availability, and low cost and low static power
1Kb 6-T Static Random Access Memory (SRAM) at 180nm, consumption [5]. However, static power consumption is
focusing on optimizing power and delay. This paper worsening with the scaling of the technology due to
contains two types of architecture to design SRAM, one is significant reduction in threshold voltages. Hence it is
bank partitioning architecture and other is matrix array. In more challenging to design the low-power SRAMs in
memory bank architecture SRAM is divided into 4 blocks the technologies 0.18um and below since the SRAM
with each block having equal capacity of 256b. Memory consumes significant static power due to sub threshold
bank is selected using block selector circuit. The key of low leakage [6]. An SRAM is matrix of static volatile
power operation in the SRAM is to reduce the word line
memory cells, and it addresses decoding functions
capacitance and bit line capacitances. The power dissipation
integrated on-chip to allow access to each cell for the
is reduced to 78% in the circuit containing memory bank
because in memory bank wordline capacitance and bitline read and write functions. The basic architecture of the
capacitance are reduced as only one bank is selected at a SRAM contains one or more rectangular arrays of
time and all the other remain in standby mode at the memory cells with control circuitry to decode addresses
expense of 6.930% more transistor count. Speed is also for few basic and special operations.
improved by 23% in the architecture containing memory
bank. Here sense amplifier, bit line conditioning circuit and II. ARCHITECTURE OF 1 KB 6T SRAM
decoder are also designed and verify various results. All the
simulations are performed using IC flow tools at TSMC
The 1 Kb SRAM organizations is random-access
180nm technology. The proposed memory circuit has architecture which is an Asynchronous design. The
applications in SoC and NoC. name is derived from the fact that memory locations
(addresses) can be accessed in random order at a fixed
rate, independent of physical location, for reading or
Keywords-180 nm, low power, VLSI, enable decoder, SRAM
writing. Here two type of SRAM architecture is
designed. One is simple architecture containing 5x 32
I. INTRODUCTION bit row and column decoder (Fig1) and other is bank
Memory is an important part of computer and partitioning based architecture. In bank partitioning
microprocessor based system design. It is used to store based architecture entire SRAM is be divided into 4
data or information in terms of binary number (0 or 1). blocks as shown in Fig 2, with each block is of 8 x 32
Also data that is used in program as well as for columns, where each word is 32 bits. Block decoder of
executing the program are stored in the memory. 2 x 4 sizes is chosen to select one block at a time
Therefore memory is required for temporary as well as keeping rest of the block in ideal state. Depending on
permanent storage of data in digital system. Generally the address of 2 x 4 decoder one out of four 3 x 8
memories are of two types (1) RAM and (2) ROM. decoder becomes on and one can perform read and
ROM is also called as permanent memory as it is write operation in the bank connected to 3 x 8 decoder.
designed once, another way we can say that it is used
only for reading purpose. While RAM is used for both
read and write. RAM is again classified in two types
SRAM and DRAM. Here in this paper SRAM memory
is designed.
Since memory is an array type of structure, so cost
per bit of the memory decreases with the cell area. For
smaller memory cells, we can achieve larger storage
capacity in the given silicon area. Hence the technology
with the smallest feature size available should be used
for the memory design. The aggressive technology-
scaling trend is driven by the requirement of large
amount of inexpensive memories for most of the
computing and networking applications. CMOS is an
excellent choice due to its superior noise margin,
Fig 1: One Kb-SRAM Core
scaling capability, mature process, worldwide
451
to write data 1, write signal is asserted and the word line
is pulled high, the q node starts to rise to vdd while the
qbar node starts dropping down to gnd. When data is
written in SRAM cell the precharge circuit become on,
which is initially in off state and the voltages on the bit
lines blbar and bl become equalized. Read operation
starts by disabling the pre-charge circuit and enabling
the word line ‘wl’ which causes the bit line blbar to
decrease at a linear rate while the other bit line bl to
remains at high value i.e. constant that is shown in the
waveform.
452
(Vdr). There are two nodes (q and qbar) of the SRAM Write margin increases with the increases value of
cell for storing value either 0 or 1. Then decrease the the pull up ratio. So be carefully you have to design
power supply voltage until the flip the state of SRAM SRAM cell inverters before calculating the write margin
cell or content of the SRAM cell remain constant. When
Vdd scales down to DRV, the Voltage Transfer Curves
(VTC) of the internal inverters degrade to such a level
Based on the VTCs, we define the read margin to S. Vd Without memory bank With memory bank
characterize the SRAM cell's read stability. We No. d Power(mW) Delay(nS) Power(m Delay(nS)
calculate the read margin based on the transistor's W)
current model. Experimental results show that the read 1 1.8 9.0475 21.981 1.0909 13.298
margin accurately captures the SRAM's read stability as 2 2.0 14.6818 17.265 2.2124 10.986
3 3.0 62.9057 12.351 4.6871 9.010
a function of the transistor's threshold voltage and the
4 4.0 143.657 10.174 9.4112 7.935
power supply voltage variations. The static read 5 5.0 258.156 8.869 16.3612 5.470
margin is preferably greater than the static write margin Table 3: Comparison of Power and Delay between containing 1Kb
SRAM memory
bank and without memory bank with varying Supply Voltage
453
time window of transition region are large so device [5] Sung-Mo Kang and Yusuf Leblebici, “CMOS Digital Integrated
found more time to short between Vdd and ground. It Circuits Analysis and Design”, Tata McGraw-Hill third
edition.
[6] Meng-Fan Chang, Shu-Meng Yang, Kuang-Ting Chen, Hung-Jen
Liao and Robin Lee, “Improving the Speed and Power of
Compilable SRAM using Dual-Mode Self-Timed Technique”,
IEEE Journal of Solid-State Circuits, 2007.
[7] Ali Fazli Yeknami, “Design and Evaluation of a Low-Voltage,
Process-Variation-Tolerant SRAM Cache in 90 nm MOS
Technology”, March 2008.
[8] Behzad Razavi (2001), Design of Analog CMOS Integrated
Circuits, McGraw Hill, New York.
[9] Bharadwaj S. Amrutur and Mark A. Horowitz, “Speed and Power
Scaling of SRAM’s”, IEEE Transactions on Solid-State
Circuits, vol. 35,pp. 175-185, no. 2, February 2000.
[11] Meng-Fan Chang, Shu-Meng Yang, Kuang-Ting Chen, Hung Jen
Liao and Robin Lee, “Improving the Speed and Power of
Compliable SRAM using Dual-Mode Self-Timed Technique”,
Fig 13: Increases of Speed with increases of CMMR IEEE Journal of Solid-State Circuits, 2007.
V. CONCLUSION
For SRAM cell we have calculate SNM, DRV RM
and WM. Architecture containing banks have less
power dissipation and high speed at the expense of
6.930% more transistor count as comparison to
architecture without memory bank. In memory bank
architecture power consumption is less. The reason
behind that, at a time only one memory bank is enabled
and other banks remain in standby mode Use of sense
amplifier also have good impact on power consumption.
Until recently, the 6T cell architecture was reserved
for niche markets such as military or space that needed
high immunity and low power components. However,
with commercial applications needing faster SRAMs,
the 6T cell may be implemented into more widespread
applications in the future. Much process development
has been done to reduce the size of the 6T cell.
ACKNOWLEDGEMENT
Authors would like to thank to Dr. Arti Noor and all
CDAC Noida staff for their support during the course
of this work.
REFERENCES
[1] Sreerama Reddy G M and P Chandrasekhara Reddy,“Design and
Implementation of 8K-bits Low Power SRAM in 180nm
Technology”, Proceedings of the International Multi Conference
of Engineers and Compute Scientists 2009 Vol. II IMECS
2009, March -20, 2009, Hong Kong.
[2] Sreerama Reddy G M and Dr. P. Chandrasekhara Reddy, “Design
and VLSI Implementation of 8Mb Low Power SRAM in 90nm”,
European Journal of Scientific search, Vol. 26, No.2, Feb 2009.
[3] Andrei Pavlov and Manoj Sachdev, “CMOS SRAM Circuit
Design and Parametric Test in Nano- Scale Technologies”, 2008
Springer Science Business Media B.V.ISBN 978-1-4020-8362-4
e- ISBN 978-1-4020-8363-1.
[4] Jan M. Rabaey, Anantha Chandrakasan and Borivoje Nikolic,
“Digital Integrated Circuit”,Pearson Education Electronics and
VLSI series, second edition.
454