Transcript Chapter 3

Chapter 3
Layout design rules
1
Introduction
 Layout rules is also referred as design rules.
 It is considered as a prescription for preparing photomasks.
 Provides a link between circuit designer and processor engineer
during manufacturing phase.
 Design rules specify geometric constraints on the layout artwork.
• Objective:




To obtain a circuit with optimum yield.
To minimize the area of the circuit.
To provide long term reliability of the circuit.
Design rules represent the best compromise between performance and
yield:
• More conservative rules increase yield.
• More aggressive rules increase performance.
• Design rules represent a tolerance that ensures high probability of
correct fabrication - rather than a hard boundary between correct
and incorrect fabrication.
2
Layout or Design Rules
• Two approaches to describing design rules:
• Lambda-based rules: Allow first order scaling by linearizing the
resolution of the complete wafer implementation.
 To move a design from 4 micron to 2 micron, simply reduce the value of
lambda.
 Worked well for 4 micron processes down to 1.2 micron processes.
 However, in general, processes rarely shrink uniformly.
 Probably not sufficient for submicron processes.
• Micron rules: List of minimum feature sizes and spacings for all
masks, e.g., 3.25 microns for contact-poly-contact (transistor pitch)
and 2.75 micron metal 1 contact-to-contact pitch.
 Stated at some micron resolution, alpha() and beta ( ) rules. Basic
feature size is defined in terms of  while minimum grid size is
described by .  and  may related by a constant factor.
 Normal style for industry.
3
Layout or Design Rules
4
Lambda-based p-well rules
 A version of p-well rules loosely based on the JPL rules.
 These rules are only representative and are the result of averaging a
large number of processes.
 The rules are defined in terms of:
 Feature sizes
 Separations and overlaps.
5
Mask No. 1:Thinox
6
Mask No. 2: P-well
7
Mask No. 3: Poly
8
Mask No 4: P-plus
9
Mask No 5: Contact
10
Mask No 5: Contact
11
Contact No 6: Metal
12
CMOS Design Rules
 A layout resembles the top view of the IC
 The layout design is 2-dimentional from the viewpoint of an IC
designer.
 in fact, designers normally do not control the depth dimension of an
IC. The depth of a transistor source or the thickness of a metal wire
is determined by the fabrication process.
 The layout designer only decides the dimensions and locations of
the transistors and interconnects them into the target circuit.
 The minimum feature size of a technology is typically denoted by the
narrowest width of a polysilicon wire that it can produce.
 For instance, if the narrowest polysilicon wire in a technology is 1m
wide, it is called 1 m technology.
 Design rules capture the physical limitations of a fabrication process.
 Design rules release designers from the details of fabrication so they
can concentrate on the design instead.
13
CMOS Design Rules
 In -based design rules (developed by Mead and Conway) the
minimum feature size is 2.
 For a 2m technology, =1m
 Different fabrication technologies apparently require different rules.
 Drawbacks of scalable design rules: the layout produced according
to scalable design rules are often larger than necessary.
 For example, a minimum spacing of 4m may be ordered although
in reality 3.2m is sufficient.
14
MOSIS Design rules
 Scalable -based design rules are used in MOSIS projects.
 http//:www.mosis.org
 A subset of MOSIS scalable CMOS (SCMOS) design rules is used
to provide an overview of their use.
15
Minimum width of
diffusion region is
3λ
Two unrelated
diffusion regions of
the same type
have to be
separated by at
least 3λ
The largest
spacing is in
the -based
design rules is
the minimum
distance
between
diffusion
regions of
different
types. (to
avoid latch-up
problem.
16
Polysilicon is used both as
transistor gates and as short
distance interconnection.
Minimum width of polysilicon
wire is 2λ
And two unrelated polysilicon
wires must be separated by as
least 2λ
17
Legend
18
Legend
19
Legend
20
Legend
21
Layout Examples
Example 1: symbolic layout for and inverter
According to -based design rules, the smallest transistor channel is 2 long
and 3 wide (the minimum width of diffusion region). However, in the
following figure the width of transistor has been increased to 4 so that a
diffusion contact (4 X 4 required) can be readily made. This 2 X 4
transistor is referred as the minimum size transistor.
An inverter formed of
minimum size
transistors is called a
minimum size inverter.
Area: 42 X 15= 6302
Stick diagram
22
Layout Examples
Example 3.1 :Alternative layouts for inverter
Area of both: 40 X 18= 7202
23
Layout Examples
Use 2 X 4 transistors to create a symbolic layout for a two NAND gate,
F  AB
24
layout Examples
Example: Minimize the area of the NAND gate discussed in previous
Example
Ans: area reduction by 12%
25
layout Examples
Example Design a layout for F  A  BC
The design objective in this example is to form two rows of tranaiators.
All pMOS transistors should be in one row and all nMOS transistors should be
in another row.
This layout structure eliminates the need to cross input signals.
26
layout Examples
Example: design the symbolic layout for the function F  A  BD  C
with 2  X 8  transistors.
27
Figure 3.28 (p. 97)
Three series-connected nFETs.
28
layout Examples
Figure 3.29 (p. 98)
A parallel-connected FET patterning.
29
layout Examples
Figure 3.30 (p. 98)
Alternate layout strategy for parallel FETs.
30
Figure 3.31 (p. 99)
Translating a NOT gate circuit to silicon.
31
Layout Examples
Figure 3.32 (p. 100)
Alternate layout for a NOT gate.
32
Layout Examples
Figure 3.34 (p. 101)
Non-inverting buffer.
33
Layout Examples
Figure 3.35 (p. 101)
Layout of a transmission gate with a driver.
34
Layout Examples
Figure 3.37 (p. 102)
NOR2 gate design.
35
Layout Examples
Figure 3.39 (p. 103)
Layout for 3-input gates.
36
Layout Examples
Figure 3.40 (p. 104)
Extension of layout technique to a complex logic gate.
37
Layout Examples
Figure 3.41 (p. 105)Creation of the dual network.
38
Layout Examples
Figure 3.42 (p. 105)A general 4-input AOI gate.
39
Stick Diagrams
poly
ndiff
pdiff
m1
m2
In the early days of MOS integrated
circuits it was noticed that when a
chip was illuminated with a white
light source, each conducting layer
had a distinct coloring associated
with t when viewed under a
microscope. This observation
provided the basis for developing
the technique:
contact
pFET
nFET
40
An example : An Inverter
Vdd
in
out
Vdd
in
out
GND
41
Vdd
An example : NOR gate
A
B
Z
Vdd
A
B
out
GND
42
Example
Z  (A B )  C
a'
c
b
Z
a'
illegal
magic
name
b
c
Convention : use a_ for a'
Vdd
a_
b
out
c
GND
43