Académique Documents
Professionnel Documents
Culture Documents
Electrical Characteristics
of CMOS
Jin-Fu Li
Department of Electrical Engineering
National Central University
Jungli, Taiwan
Outline
Resistance Estimation
Resistance
R ( / t )( L / W ) , where ( , t , L, W ) is (resistivity,
thickness, conductor length, conductor width)
Sheet resistance
Rs /
Thus R Rs ( L / W )
1 rectangular block
R Rs ( L / W )
t
L
4 rectangular block
R Rs (2 L / 2W ) Rs ( L / W )
Advanced Reliable Systems (ARES) Lab.
G
Rn
Cs
CD
c
Vds
Drain-Source Resistance
The resistance at point a
The current is approximated by
I ds n (Vgs Vt )Vds
Thus the resistance is
Rn 1 / n (Vgs Vt )
[
2
(
V
V
)
V
V
ds
n
gs
t
ds
ds ]
2
Thus the resistance is
Rn 2 / n [2(Vgs Vt ) Vds ]
Drain-Source Resistance
The resistance at point c
The current is
I ds
1
n (V gs V t ) 2
2
Capacitance Estimation
The switching speed of MOS circuits are
heavily affected by the parasitic capacitances
associated with the MOS device and
interconnection capacitances
The total load capacitance on the output of a
CMOS gate is the sum of
Gate capacitance
Diffusion capacitance
Routing capacitance
MOS-Capacitor Characteristics
The capacitance of an MOS is varied with the
applied voltages
Capacitance can be calculated by
C 0 x A
d
x is dielectric constant
0 is permittivity of free space
Cox
High frequency
Accumulation
Depletion
Vt
Vg
Inversion
Cgb
channel
source
Cgd
drain
depletion layer
Cdb
Csb
Cgs
Cgb Csb
substrate
Cg=Cgb+Cgs+Cgd
Advanced Reliable Systems (ARES) Lab.
10
1 0 SiO2
A
2 tox
gs
t ox
11
Approximation of the Cg
The Cg can be further approximated with
C g C ox A ,
where Cox
o SiO
tox
Cg
3 . 9 8 . 854 10
150 10 8
14
2 25 . 5 2 10 4 pF 0 . 005 pF
Jin-Fu Li, EE, NCU
12
Diffusion Capacitance
Diffusion capacitance Cd is proportional to the
diffusion-to-substrate junction area
Substrate
b
a
Source
Diffusion
Area
Drain
Diffusion
Area
b
Cjp
Xc (a finite depth)
Cja
13
Junction Capacitance
Semiconductor physics reveals that a PN
junction automatically exhibits capacitance
due to the opposite polarity charges involved.
This is called junction or depletion
capacitance and is found at every drain or
source region of a MOS
The junction capacitance is varies with the
junction voltage, it can be estimate as
C j C j 0 (1
Vj
Vb
)m
14
L
T
H
substrate
Insulator (Oxide)
15
C23
Layer 3
C22
C21
Layer 2
Layer 1
C2=C21+C23+C22
16
m2
m2
m2
m2
C
C
C
poly
m1
m2
m2
m1
poly
m1
Thin-oxide/diffusion
Substrate
17
Inductor
For bond wire inductance
L ln( 4h )
2
4h
18
Distributed RC Effects
The propagation delay of a signal along a wire
mainly depends on the distributed resistance
and capacitance of the wire
A long wire can be represented in terms of
several RC sessions, as shown below
Ij-1
R
C
R
C
Vj-1 R Vj
C
Ij
R Vj+1
C
R
C
19
Distributed RC Effects
As the number of sections in the network
becomes large (and the sections become
small), the above expression reduces to the
differential
form
2
dV
d V
kx
rc
x
dt
dx 2
r : resistance per unit length
c : capacitance per unit length
RCn( n 1)
t n 0 .7
2
rcl 2
t1 0.7
2
20
10
1000
10
1000
With buffer p
buf
4ns tbuf 4ns 8ns tbuf
21
Crosstalk
A capacitor does not like to change its voltage
instantaneously.
A wire has high capacitance to its neighbor.
When the neighbor switches from 1-> 0 or 0->1,
the wire tends to switch too.
Called capacitive coupling or crosstalk.
Crosstalk effects
Noise on nonswitching wires
Increased delay on switching wires
22
Crosstalk Delay
Assume layers above and below on average are
quiet
Second terminal of capacitor can be ignored
Model as Cgnd = Ctop + Cbot
Ceff(A)
MCF
Constant
VDD
Cgnd + Cadj
Switching with A
Cgnd
Switching opposite A
2VDD
Cgnd + 2 Cadj
A
Cgnd
B
Cadj
Cgnd
23
Crosstalk Noise
Crosstalk causes noise on nonswitching wires
If victim is floating:
model as capacitive voltage divider
Vvictim
Cadj
Cgnd v Cadj
Vaggressor
Aggressor
Vaggressor
Cadj
Victim
Cgnd-v
Vvictim
24
Driven Victim
Usually victim is driven by a gate that fights
noise
Noise depends on relative resistances
Victim driver is in linear region, agg. in saturation
If sizes are same, Raggressor = 2-4 x Rvictim
Vvictim
Cadj
C gnd v Cadj
1
Vaggressor
1 k
Raggressor
Cgnd-a
Vaggressor
victim
Rvictim C gnd v Cadj
Aggressor
Cadj
Rvictim
Victim
Cgnd-v
Vvictim
25
Simulation Waveforms
Simulated coupling for Cadj = Cvictim
Aggressor
1.8
1.5
1.2
0.9
0.6
0.3
0
0
200
400
600
800
1000
1200
1400
1800
2000
t(ps)
Advanced Reliable Systems (ARES) Lab.
26
DC Response
DC Response: Vout vs. Vin for a gate
Ex: Inverter
When Vin = 0 Vout=VDD
VDD
When Vin = VDD Vout=0
In between, Vout depends on
Idsp
Vout
transistor size and current Vin
Idsn
By KCL, must settle such that
Idsn = |Idsp|
We could solve equations
But graphical solution gives more insight
27
Transistor Operation
Current depends on region of transistor
behavior
For what Vin and Vout are NMOS and PMOS
in
Cutoff?
Linear?
Saturation?
28
NMOS Operation
Cutoff
Vgsn < Vtn
Linear
Vgsn > Vtn
Saturated
Vgsn > Vtn
VDD
Vgsn = Vin
Vin
Vout
Idsn
Vdsn = Vout
Idsp
29
NMOS Operation
Cutoff
Linear
Saturated
VDD
Vgsn = Vin
Vdsn = Vout
Vin
Idsp
Vout
Idsn
Jin-Fu Li, EE, NCU
30
PMOS Operation
Cutoff
Vgsp > Vtp
Linear
Vgsp < Vtp
Saturated
Vgsp < Vtp
VDD
Vgsp = Vin - VDD
Vdsp = Vout - VDD
Vtp < 0
Vin
Idsp
Vout
Idsn
31
PMOS Operation
Cutoff
Linear
Saturated
VDD
Vgsp = Vin - VDD
Vdsp = Vout - VDD
Vtp < 0
Vin
Idsp
Vout
Idsn
32
I-V Characteristics
Make pMOS is wider than nMOS such that n
= p
Vgsn5
Vgsn4
Idsn
Vgsn3
-Vdsp
Vgsp1
Vgsp2
-VDD
0
VDD
Vdsn
Vgsp3
Vgsp4
Vgsn2
Vgsn1
-Idsp
Vgsp5
Advanced Reliable Systems (ARES) Lab.
33
Idsn, |Idsp|
Vin1
Vin5
Vin2
Vin4
Vin3
Vin3
Vin4
Vin5
Vin2
Vin1
Vout
VDD
34
Idsn , |Idsp |
Vin1
Vin5
Vin2
Vin4
Vin3
Vin3
Vin4
Vin5
Vin2
Vin1
Vout
VDD
Vin
Idsp
Vout
Idsn
VDD
35
DC Transfer Curve
Transcribe points onto Vin vs. Vout plot
Vin1
Vin5
Vin2
Vin4
Vin3
Vin3
Vin4
Vin5
Vin2
Vin1
VDD
Vout
VDD
A
Vout
D
0
Vtn
VDD/2
E
VDD+Vtp
VDD
Vin
Advanced Reliable Systems (ARES) Lab.
36
Operation Regions
Revisit transistor operating regions
Region
nMOS
pMOS
Cutoff
Linear
Saturation
Linear
Saturation
Saturation
Linear
Saturation
Linear
Cutoff
VDD
A
Vout
D
0
Vtn
VDD/2
E
VDD+Vtp
VDD
Vin
37
Beta Ratio
If p / n 1, switching point will move from
VDD/2
Called skewed gate
Other gates: collapse into equivalent inverter
VDD
p
10
n
Vout
2
1
0.5
p
0.1
n
0
Vin
Advanced Reliable Systems (ARES) Lab.
VDD
38
Noise Margin
How much noise can a gate input see before it
does not recognize the input?
Output Characteristics
Logical High
Output Range
VDD
Input Characteristics
Logical High
Input Range
VOH
NMH
VIH
VIL
Indeterminate
Region
NML
Logical Low
Output Range
VOL
Logical Low
Input Range
GND
39
Transient Analysis
DC analysis tells us Vout if Vin is constant
Transient analysis tells us Vout(t) if Vin(t)
changes
40
Switching Characteristics
Switching characteristics for CMOS inverter
Vin(t)
Vout(t)
Vds=Vgs-Vt
CL
VDD
Ids
Vin(t)
VDD
tdr
tdf
90%
Vout(t)
Vout(t)
50%
10%
tf
tr
VDD
t
Jin-Fu Li, EE, NCU
41
Switching Characteristics
Rise time (tr)
42
PMOS
Vout(t)
Input rising
NMOS
Idsn
CL
Saturated Vout>=VDD-Vtn
Vout(t)
NMOS
Rcn
CL
Nonsaturated 0<Vout<=VDD-Vtn
43
Timing Calculation
tf1 can be calculated with the current-voltage
equation as shown below, while in saturation
CL
dVout n
(VDD Vtn ) 2 0
dt
2
CL
nVDD
tr k
CL
pVDD
44
Design Challenges
n = p , rise time=fall time
This implies Wp=2-3Wn
Reduce CL
Increase n and p
Increase VDD
45
Gate Delays
Consider a 3-input NAND gate as shown below
P3
P2
P1
out
N3
IN-3
IN-2
N2
IN-1
N1
(1 / n1 ) (1 / n 2 ) (1 / n 3 )
For n1 n 2 n 3 neff
n
3
46
Gate Delays
Graphical illustration of the effect of series
transistors
L
L
3L
L
w
In general, the fall time tf is mtf (tf/m) for m ntransistors in series (parallel). Similarly the rise
time tr for k p-transistors in series (parallel) is ktr
(tr/k)
Advanced Reliable Systems (ARES) Lab.
47
Switch-Level RC Model
RC modeling
Transistors are regarded as a resistance
discharging or charging a capacitance
Simple RC modeling
Lumped RCs
tdf Rpulldown C pulldown path
Elmore RC modeling
Distributed RCs
t d Ri Ci
Rp
Rn
48
Example
Consider a 4-input NAND as shown below
Simple RC model
t df R pulldown C pulldown path
(RN1 RN2 RN3 RN4 ) (Cout Cab Cbc Ccd )
t dr R p 4 Cout
P4
Elmore RC model
N4
Cab
t d Ri Ci
P2
P1
out
Cout
N3
P3
Cbc
N2
Ccd
N1
49
50
tinv-pair
R 3 C eq 2
Icharge
4/1
R
2/1
3Ceq
Idischarge
3Ceq
R
3 C eq
2
3 RC eq 3 RC eq
6 RC eq
Wp=2Wn
Example 2:
tinv-pair
2/1
R 2 C eq 2 R 2 C eq
2R
2/1
Icharge
2Ceq
Idischarge
2Ceq
6 RC eq
Wp=Wn
Advanced Reliable Systems (ARES) Lab.
51
Stage Ratio
To drive large capacitances such as long buses,
I/O buffers, etc.
Using a chain of inverters where each successive
inverter is made larger than the previous one until
the last inverter in the chain can drive the large
load in the time required
The ratio by which each stage is increased in size
is called stage ratio
a2
a3
CL
n(4) stages
Advanced Reliable Systems (ARES) Lab.
52
Stage Ratio
The delay through each stage is atd, where td is
the average delay of a minimum-sized inverter
driving another minimum-sized inverter
Hence the delay through n stages is natd
If the ratio of the load capacitance to the
capacitance of a minimum inverter, CL/Cg, is R,
then an=R
Hence ln(R)=nln(a)
Thus the total delay is ln(R)(a/ln(a))td
The optimal stage ratio may be determined from
k a opt
a opt e
a opt
where k is
C drain
C gate
53
Power Dissipation
Instantaneous power
Peak power
Average power
54
55
Static Power
Static dissipation is major contributed by
Reverse bias leakage between diffusion regions
and the substrate
Subthreshold conduction
Vin
Gnd
VDD
Vout
leakage current
p+
n+
n+
p+
p+
n+
i0 is (e qV / KT 1)
n-well
p-substrate
n=number of devices
Advanced Reliable Systems (ARES) Lab.
56
VDD
dvo
dt
dv
in io C L o
dt
i p io C L
ip
Vout
Vin
in
i (t )v (t )dt
CL
io
0
1 VDD
Psw [ CL vo dvo CL vo dvo ]
VDD
T 0
2
CLVDD
2
Psw
fCLVDD
T
Jin-Fu Li, EE, NCU
57
VDD
1
2
Cvo dvo C LVDD
2
58
59
VDD
VDD-|Vtp|
isc
Vin
Vout
tr
tf
Vtn
CL
Imax
Imean
t1 t2 t3
I mean
I mean
t3
1 t2
2 [ i (t )dt i (t )dt ]
t2
T t1
4 t2
[ i (t )dt ]
T t1
60
VT
tr
VDD
t2
tr
2
Psc
12
(VDD 2VT ) 3f ( tr t f )
61
VDD
B
C
C1
out
A
C
B
C2
62
ABC
A
B
100
111
D
Z
Z
Unit delay
Spurious transition
63
64
C V V
i 1
DD
65
66
Electromigration is affected by
Current density
Temperature
Crystal structure
67
Vout
Time
Current
Time
VDD Pad
Vout
Vin
I
VSS Pad
VL
L
VL=L(di/dt)
Time
Ground bounce
Advanced Reliable Systems (ARES) Lab.
68
69
Contact Replication
Current tends to concentrate around the
perimeter in a contact hole
70
Charge Sharing
Charge Q=CV
A bus example is illustrated to explain the
charge sharing phenomenon
A bus can be modeled as a capacitor Cb
An element attached to the bus can be modeled as
a capacitor Cs
Bus
Vb
Cb
Vs
(Qb CbVb )
QT CbVb CsVs
CT Cb Cs
Advanced Reliable Systems (ARES) Lab.
VR
Cs
(Qs CsVs )
QT
(CbVb C sVs ) /(Cb C s )
CT
71
Design Margining
The operating condition of a chip is influenced by
three major factors
Operating temperature
Supply voltage
Process variation
Fast
Slow
Advanced Reliable Systems (ARES) Lab.
SF
SS
FF
TT
FS
NMOS
Fast
72
Package Issues
Packaging requirements
Bonding techniques
Wire Bonding
Substrate
Die
Pad
Lead Frame
73
Yield Estimation
No. of good chips per wafer
Y
100%
Total number of chips per wafer
Wafer cost
Die cost
Dies per wafer Die yield
wafer diameter/22 wafer diameter
Dies per wafer
die area
2 die area
74
Die Cost
Single die
Wafer
Going up to 12 (30cm)
75
Scaling Theory
Consider a transistor that has a channel width
W and a channel length L
We wish to find out how the main electrical
characteristics change when both dimensions
are reduced by a scaling factor S>1 such that
the new transistor has sizes
~ L
~ W
L
S
S
Gate area of the scaled transistor
A~ SA
The aspect ratio of the scaled transistor
~
W W
L L~
2
76
Scaling Theory
The oxide capacitance is given by
Cox t ox
ox
If the new transistor has a thinner oxide that is
t
decreased as ~tox ox , then the scaled device has
S
~
Cox SCox
S (VDD VT )
77
Scaling Theory
On the other hand, if we can scale the voltages in
the scaled device to the new values of
V~DD VDD V~T VT
S
S
The resistance of the scaled device would be
~
unchanged with R R
~ V
VDS DS
S
~ V
VGS GS
S
S VGS VT VDS
I
~
[(
)
] D
ID
2
S
S S
S
78
Summary
We have presented models that allow us to
estimate circuit timing performance, and
power dissipation
Guidelines for low-power design have also
been presented
The concepts of design margining were also
introduced
The scaling theory has also introduced
79