ASIC-System On Chip-VLSI Design - SRAM Cell Design PDF
ASIC-System On Chip-VLSI Design - SRAM Cell Design PDF
Asic Design
Circuit Design
System Design
Circuit Simulator
Pages
Home
About
Physical Design
Low Power VLSI
STA
Synthesis
DFT
FV
Verilog
Links
Jobs
Training
Layoffs Watch
Ads by Google
1.1 Introduction
Memory SRAM
DRAM SRAM
Starting from the design specification to the generation of mask layout, layout design of an
integrated circuit has several processing steps which have to be carefully exercised. These steps include
Layout Design
design of transistor level schematic, SPICE simulation of the circuit according to the designed W/L
ratios of the individual transistors, drawing of the layout using a layout editor, design rule check,
parasitic extraction and final simulation and verification. These all processing methods are inevitable for
TOP POSTS
What is the difference between
FPGA and ASIC?
Backend (Physical Design)
Interview Questions and Answers
Companywise ASIC/VLSI Interview
Questions
Process-Voltage-Temperature
(PVT) Variations and Static Timing
Analysis
What is the difference between
FPGA and CPLD?
Clock Gating
Embedded System for Automatic
Washing Machine using Microchip
PIC18F Series Microcontroller
Synthesizable and NonSynthesizable Verilog constructs
Power Planning
Transition Delay and Propagation
Delay
the error free operation of chip and similar methodology is followed for the design of 1 KByte SRAM
IC. Basic building block of the SRAM is SRAM cell which stores one bit data. Using common bit lines
data can be read and written to the SRAM cell. SPICE, being an industry standard tool for circuit
simulation and analysis, is used for the simulation and analysis of SRAM cell and subsequently for the
whole design. Precharge circuit, sense amplifier and read-write circuits completes the one SRAM
memory. The memory is arranged in row- column matrix which facilitates easy addressing of memory
bits and also provides design flexibility. Once the functionality of one memory cell array is proved it can
be duplicated several times with minor design change in the I/O control circuitry.
WRITE TO A
To subscribe as
READ MORE
ASIC synthesi
Synthesis
Verilog
Verilog HDL
verilog examp
verilog intervie
questions
verilog tutorial
Verification
verilog tutorial
http://asic-soc.blogspot.in/2007/07/sram-cell-design.html
1/8
27/10/2015
ASIC
DSP
Sense amplifier is used to sense the data present in the memory cell. Input and output data
control block consists of read and write circuitry and related driver circuits. 8 bits of the data D0 to D7
is read or written parelell. Read enable (RE) and write enable (WE) are the two control lines available
for the chip. As the name indicate, before data read operation is performed RE is activated for read
operations. Similarly for write operations WE signal has to be enabled.
HDL
Static Timing A
(STA)
logic synhesis
Low Power
Techniques
logic synthesis
FPGA
MATLAB
Timing Analys
Mentor Graph
Verification IP
Physical Desig
DSP filters
Digital design
CMOS
Verilog Classe
Asynchronous
3-D ICs
Basic gates us
MUX
Digital filters
PIC Microcont
constraints
low power
.lib
Libraries
PIC 16F877A
STA
Synopsys
interview
Leakage Powe
VIP
VLSI
Design For Te
Multi Vdd
Power Plannin
Reconfigurabl
Computing
System on Ch
hold time
operating Con
setup time
ASIC syynthes
Clock Tree Sy
(CTS)
DVFS
Design For Te
(DFT)
EDA
will flow through M1 and M5 charging the bit line capacitance, say CBL. The existing capacitance on
Floorplanning
the line BLbar, say CBLbar discharges through the transistors M6 and M4. This process develops a
Full Custom
voltage difference between node Y and node Ybar which is sensed by the sense amplifier to detect it as
1. Similarly a 0 in the cell is also detected by the sense amplifier.
Write operation:
Let us consider the write operation of zero to the cell which is storing a value of 1. For this,
sense amplifiers and precharge circuits are disabled. The cell is selected by activating the corresponding
WL signal. To write zero to the cell, BL line held low and BLbar line is raised to VDD by the write
http://asic-soc.blogspot.in/2007/07/sram-cell-design.html
Multi Vt
New Devices
OVM
Placement
RTL
SPICE
circuit. Thus the node Ybar is pulled up towards the VDD/2 while node Y is pulled down to VDD/2.
SoC Design
When the voltage crosses this level on two nodes feedback action starts. Parasitic capacitances
SoC Integratio
2/8
27/10/2015
developed by M3, M5 and M4, M6 are charged and discharged respectively. Ultimately node Y
Timing paths
stabilises at the value 1. Since these parasitic capacitances offered by transistors are comparatively
Transition dela
much lesser than the bit line capacitances, write operation is faster than read operation.
UVM
Transistor sizing:
layout
both the directions equal to that of the basic inverter. From the basic inverter design (W/L)n is usually
5.2.8. Blocking
non-blocking-r
condition
1.5 to 2 and for a matched design, (W/L)p=(n/p)(W/L)n. The SRAM cell must be designed such a
7 Segment Dis
way that, during read operation, the changes in Y and Ybar are small enough to prevent the cell from
AMBA AHB
The W/L ratio of the transistor is selected to provide the gate with current driving capability in
changing its state. Generally two back to back coupled inverters of the SRAM cell is designed so that
Kn and Kp are matched. This design places the inverter threshold at VDD/2. The size of the access
transistors are usually made 2 to 3 times wider than Kn of the inverters.
To achieve optimum operation of the cell following (W/L) ratio is choosed for different
transistors. A minimum ratio of 2 is required for NMOS transistors of inverters and 4 is necessary for
PMOS transistors. Access transistors must be made double wider or more by providing a W/L ratio of
more than 4. But these set of ratios does not match with the design rule of Cadence Virtuoso layout
editor for 0.18 micron technology. For 0.18 technology minimum width for an NMOS transistor
comes out to be 0.6 . Thus (W/L) ratio is 3.33. For PMOS transistor the ratio becomes 6.66. This
implies a width of 1.2 . Based on the SPICE simulation results and its analysis, W/L ratio for access
transistor is kept at 9.99. This refers to a gate width of 1.8 .
1.4 Simulation
Mainly two types of simulation analysis is reported here. The first one focusses on normal
AMBA APB
AMBA AXI
AMBA Bus
ASIC Jobs
Backend train
Basic
Microelectroni
Blocking Vs
Nonblocking
Broadcom
CMOS Design
Clock Gating
Clock Logic
Clock definitio
Congestion
CoreConnect
operation of SRAM cell with approapriate W/L ratio. The second simulation studies the affect of
DFT
DTMF
Simulation 1:
Figure (1.3) shows the SPICE simulation waveform of the SRAM cell. For 0.18 technology, the gate
width choosed for NMOS transistors are 0.6 ; for PMOS it is 1.2 V and for access transistors gate
width is 1.8 . When WL is disabled(i.e. low) SRAM cell is disconnected from the BL and BLbar
lines. Hence voltage at node Y and Ybar is complement to each other and remains in a stable state.
The stable state value is dependant on the previous value present at BL and BLbar lines.
BL signal is forced with pulse waveform of period 8ns and pulse width of 4ns. Upon activation of WL
signal SRAM cell gets connected with BL and BLbar signal.
Design For
Manufacture-D
Dynamic Powe
Embedded Job
Embedded Sy
Embedded etc
Embedded tra
FFT
FIR Filter
FSM
Finite State M
Flash memory
Frontend train
Gate Delay
History of VLS
IC Fabrication
ICV
IP Cores
IPs
Indsustry watc
Intel
Internal Power
Intrinsic Delay
Now the expected waveform at the BLbar is the inverted signal of BL. The same result can be
seen in simulation waveform 1. Voltage at node Y follow the pulse voltage of BL; and node Ybar and
BLbar are same and complement to pulse signal at BL.
Removal of WL again disconnects the SRAM cell from the BL and BLbar line. SRAM cell holds the
value whatever it had while removing the WL signal.
Simulation 2:
If W/L ratio of access transistors is reduced to the ratio of PMOS transistors then the cell fails to
operate as expected due to the inefficient current driving capability of access transistor. Related
http://asic-soc.blogspot.in/2007/07/sram-cell-design.html
Jobs
Lay-off
Lynx Design S
Magma
Memory Desig
Microprocesso
Monitors
Nangate 45nm
libraries
Net delay
NoC
OpenSPARC
processor
3/8
27/10/2015
waveform is shown in Figure (1.4). Here for access transistors W/L ratio is set same as that of PMOS
Others ..DSP
transistor i.e. 6.66 which implies a gate width of 1.2 . Observe the waveform shape at area marked by
PVT vs STA
circles. The waveform at BLbar should be complement to waveform at BL. But it is remaining at high
Physical Desig
training
state.
Power Gating
Propagation d
Protocols
Qualcomm
RTL Coding
Reset Logic
Routing
SDC
SNUG
SRAM Chip
Sandisk
Short Circuit P
Static memory
Sub Threshold
Leakage
Systolic Array
Texas Instrum
(TI)
Training
VLSI Jobs
VLSI fabricatio
This is due to the poor switching of access transistor. Similar waveforms can be observed at node Y
VLSI training
Voltage Scalin
WLM
Washing Mach
Western Digita
particular manufacturing process has to be followed for the physical mask layout generation.
Geometries are determined by the electrical properties of the devices and design rules pertaining to the
free download
processor
jitter
latency
optical lithogra
optimization
process variat
skew
method. Stick diagram representation is drawn which shows the location of the transistors, local
transition fixing
intersections between transistors and location of the contacts. Mask layers are formed using a layout
transition viola
editor tool. After several iterations of editing and design rule check (DRC) and layout versus schematic
uncerainty
(LVS) check the layout is subjected to extraction procedure. Extraction procedure extracts parasitic
capacitance values and actual sizes of the transistors.
4/8
27/10/2015
allowable separations between two different features. Mainly there are two ways of design rules. They
are known as micron rules and lambda rules. In micron rule all layout constraints are defined in
micrometers while in lambda rule it is defined in terms of a single parameter lambda (). For the present
assignment micron rules are used.
Design specifications of the SRAM cell and its corresponding SPICE simulation results have
been studied in the previous sections. For the SRAM cell layout shown in the Figure (1.5) first we need
to design the individual transistors according to the design rules. PMOS transistors are placed in an nwell region whereas NMOS transistors are placed directly above the substrate. Transistors M1 to M6
are placed. Polysilicon gate of both NMOS and PMOS transistors of cross coupled inverters (i.e.
latch) are aligned so that polysilicon length is minimized to reduce parasitic resistance and capacitance.
Metal 1 and metal 2 layer is used for the interconnections between transistors. Metal 1 is used for
direct interconnections and wherever the connections crisscross metal 2 layer is used. Bit line (BL) and
bit linebar(BLbar) are vertically drawn with metal 1 layer while word line(WL), vdd and gnd are drawn
horizontally with metal 2 layer. This layout method helps to extend SRAM memory by adding more
SRAM cells.
From the layout Figure (1.5) it can be noted that total area of the SRAM cell can be further
reduced by using optimization methods (like Euler method). Since basic motto of this SRAM cell
design is to understand the layout methodology no optimization effort is put.
5/8
27/10/2015
Read operation:
At startup both decoders are inactive. As soon as decoders are enabled either by separate
address enable signal or chip enable signal, they are precharged first. This process makes all output
high for a small amount of time. This address is invalid. Then address settles down according to the
input of the decoder and one particular SRAM cell is activated.
To begin the read operation, precharge circuit is enabled by activating PE for a small amount of time
and then disabled. This process precharges BL and BLbar lines to either VDD or VDD/2 depending on
the precharge circuit used. To sense the voltage difference established at BL and BLbar, sense
amplifier is enabled. Sense amplifier reinforces the state of the BL and BLbar lines. Activation of read
enable (RE) signal enables the read buffer. Since BL and BLbar lines are commonly connected DL and
DLbar lines and these two signals are input to read buffer. The read SRAM cell data traverses towards
read buffer. The read buffer reads both DL and DLbar lines and outputs the data available in DL line.
Thus the data bit is read from memory cell. To continue the read operation address bits are changed to
address the next memory cell. Precharge is activated and then deactivated. Since sense amplifier and
read buffer is already activated read data is immediately available at the output of the buffer.
For simulation PE, SE and RE are separately forced. But in practical cases taking all these enable
signals to chip I/O may not be efficient design strategy. A single read enable signal is provided for the
chip. One more circuitry has to be added which provides sufficient delay between PE, SE and RE so
that all circuits are enabled one by one.
Write operation:
For the write operation PE, SE and RE signal is disabled which disables all read related circuits
from interacting with SRAM cell. The address is selected and data is given to write circuit as input.
Upon the activation of write enable (WE) signal, write buffer output change according to the input. The
outputs are connected to DL and DLbar lines and hence BL and BLbar lines, both signals are forced to
change to a new value. The feedback action in SRAM cell then stabilizes the data of the memory. WE
signal is then disabled for safe write operation and to avoid further writing of spurious data. To continue
the write operation to other cells address bits are changed and same procedure is repeated.
1.7 Conclusion
Single bit SRAM cell is designed and simulated. From the simulation and related analysis it is found that
access transistor size plays vital role in the memory bit design. The latch transistors should be matched
and access transistors must be twice of the PMOS transistor size. Layout is drawn and area required is
approximately 1 x 0.7 . Applying optimization methods will reduce the area required for the cell.
Further study and analysis of extracted parasitic values and its affect on circuit functionality will prove
beneficial to include the cell in a standard cell library.
References
[1] Sung Mo Kang and Yusuf Leblebici, CMOS digital integrated circuits-analysis and design, Tata
McGraw hill, third edition, 2003
[2] Jan M Rabaey & Anantha Chandrakasan & Borivoje Nikolic, Digital integrated circuits-a design
perspective, Pearson education, third edition, 2005
[3] Sedra & smith, Microelectronic circuits, oxford university press, fifth edition, 2004
[4] One KByte SRAM chip Specifications and Design Plan
Tags: SPICE, SRAM cell design
11 comments:
vani September 9, 2008 at 11:15 AM
Murali
Sometime back i deleted those diagrams due to some errors. I will upload. Sorry for the
incomplete article.
Reply
Reply
6/8
27/10/2015
Murali
Can you tell me why the shape of memory is rectangular instead of square? I mean, why 128x8
instead of 32x32?
Reply
good artical
pls modify the ckt and add mux/demux and bidirectional driver
Reply
good artical.
Reply
Comment as:
Publish
Google Account
Preview
Home
Older Post
http://asic-soc.blogspot.in/2007/07/sram-cell-design.html
7/8
27/10/2015
About Me
Murali
Thank You !
26,589
Like
Share
109
Copyright/Disclaimer
http://asic-soc.blogspot.in/2007/07/sram-cell-design.html
8/8