© Digital Integrated Circuits 2nd Inverter Digital Integrated Circuits A Design Perspective The Inverter
© Digital Integrated Circuits2nd Inverter
Digital Integrated
CircuitsA Design Perspective
The Inverter
© Digital Integrated Circuits2nd Inverter
The CMOS Inverter: Switch models
Vin Vout
CL
VDD VDD VDD
0
VoutVout
Rn
Rp
VOL = 0
VOH = VDD
VM = f(Rn, Rp)
© Digital Integrated Circuits2nd Inverter
Important Properties of CMOS
Invertor is ratioless, because output voltage
level is not dependent on the device size.
Low output impedance: there always a path
between output and VDD or GND.
Input gate resistance is extremely high. A
single invertor can (theoretically) drive infinite
invertors.
In steady state, no direct path exists between
VDD and GND.
© Digital Integrated Circuits2nd Inverter
Voltage Transfer
Characteristic
© Digital Integrated Circuits2nd Inverter
PMOS Load Lines
VDSp
IDp
VGSp=-2.5
VGSp=-1VDSp
IDn
Vin=0
Vin=1.5
Vout
IDn
Vin=0
Vin=1.5
Vin = VDD+VGSpIDn = - IDp
Vout = VDD+VDSp
Vout
IDn
Vin = VDD+VGSp
IDn = - IDp
Vout = VDD+VDSp
© Digital Integrated Circuits2nd Inverter
CMOS Inverter Load Line Characteristics
IDn
Vout
Vin = 2.5
Vin = 2
Vin = 1.5
Vin = 0
Vin = 0.5
Vin = 1
NMOS
Vin = 0
Vin = 0.5
Vin = 1Vin = 1.5
Vin = 2
Vin = 2.5
Vin = 1Vin = 1.5
PMOS
© Digital Integrated Circuits2nd Inverter
CMOS Inverter VTC
Vout
Vin0.5 1 1.5 2 2.5
0.5
11.5
22.5
NMOS resPMOS off
NMOS satPMOS sat
NMOS offPMOS res
NMOS satPMOS res
NMOS resPMOS sat
VTC exhibits very narrow
transition zone. This results
from high gain when both
NMOS and PMOS are in
Saturation mode.
High gain means a small
change in the input results
in large output variation.
© Digital Integrated Circuits2nd Inverter
Switching Threshold as a function of
Transistor Ratio
The switching threshold (VM ) is defined
as the point where Vin = Vout
Both PMOS and NMOS are in
Saturation (and velocity saturated)
© Digital Integrated Circuits2nd Inverter
Switching Threshold as a function
of Transistor Ratio For identical oxide thickness for NMOS and MOS
VM r VDD / (1+r)
This means VM is set by r, which compares the driving
strength of PMOS and NMOS
It is desirable ( for noise margin) to have VM in the
middle of VDD, where r=1
The ratio PMOS to NMOS sizes for desired VM:
)2/('
)2/('
)/(
)/(
DSATpTpMDDDSATpp
DSATnTnMDSATnn
n
p
VVVVVk
VVVVk
LW
LW
© Digital Integrated Circuits2nd Inverter
Switching Threshold as a function of Transistor Ratio
If devices are long-channel or VDD is
low, then velocity saturation does not
occur. Under these conditions VM can
be expressed as:
n
p
TpDDTn
M
k
kr
r
VVrVV
1
)(
© Digital Integrated Circuits2nd Inverter
Switching Threshold as a function of
Transistor Ratio
100
101
0.8
0.9
1
1.1
1.2
1.3
1.4
1.5
1.6
1.7
1.8
MV
(V)
Wp/Wn
Calculated VM is at sizes
of NMOS and PMOS:
(W/L)p/(W/L)n = 3.5
Simulated VM is at ration
of 3.4
Note:
-VM is relatively insensitive
to device variations.
-Effect of changing the
device sizes is to shift the
VTC.
© Digital Integrated Circuits2nd Inverter
Determining VIH and VIL
VOH
VOL
Vin
Vout
VM
VIL VIH
A simpler approach for calculating
noise margins is to use piece-wise
linear approximation approach.
g: gain at the switching threshold
VM . g is negative.
High gain increases
noise margins, and
therefore desirable.
© Digital Integrated Circuits2nd Inverter
Inverter Gain
0 0.5 1 1.5 2 2.5-18
-16
-14
-12
-10
-8
-6
-4
-2
0
Vin
(V)
gain
© Digital Integrated Circuits2nd Inverter
Robustness: Impact of Process Variations
0 0.5 1 1.5 2 2.50
0.5
1
1.5
2
2.5
Vin
(V)
Vo
ut(V
)
Good PMOSBad NMOS
Good NMOSBad PMOS
Nominal
Manufacturing causes
devices to deviate from
ideal behavior.
The variation causes shift
in the switching threshold.
But it does not affect the
device operation.
© Digital Integrated Circuits2nd Inverter
Robustness: Gain as a function of VDD
0 0.05 0.1 0.15 0.20
0.05
0.1
0.15
0.2
Vin
(V)
Vout (
V)
0 0.5 1 1.5 2 2.50
0.5
1
1.5
2
2.5
Vin
(V)
Vout(V
)
Reducing power supply improves gain, but it deteriorates
for very low supply voltage. The minimum voltage is:
VDDmin = 2..4 kT/q = 50-100mV
However, reducing power supply increases gate delay.
© Digital Integrated Circuits2nd Inverter
Propagation Delay
16
© Digital Integrated Circuits2nd Inverter
CMOS Inverter: Transient Response
tpHL = f(Ron.CL)
= 0.69 RonCL
VoutVout
Rn
Rp
VDDVDD
Vin VDDVin 0
(a) Low-to-high (b) High-to-low
CLCL
© Digital Integrated Circuits2nd Inverter
CMOS Inverter: Transient Response (2)
CL consists of:
Overlap gate-drain capacitance of driving gate
(~0.31 fF/m)
– Causes glitch in the output before switching
Diffusion capacitance of driving gate
– Bottom junction capacitance: (2 fF/m2)
– Sidewall junction capacitance: (~0.25 fF/m)
Gate capacitance of fanout gates (6 fF/m2)
Wire capacitance
© Digital Integrated Circuits2nd Inverter
Inverter Propagation Delay Propagation delay is proportional to the time-constant of the network
formed by the pull-down resistor and the load capacitance
tpHL = ln(2) Reqn CL = 0.69 Reqn CL
tpLH = ln(2) Reqp CL = 0.69 Reqp CL
tp = (tpHL + tpLH)/2 = 0.69 CL(Reqn + Reqp)/2
To equalize rise and fall times make the on-resistance of the NMOS
and PMOS approximately equal.
VDD
Rn
Vout = 0
Vin = V DD
CL
tpHL = f(Rn, CL)
19
Inverter Transient Response
-0.5
0
0.5
1
1.5
2
2.5
3
0 0.5 1 1.5 2 2.5
Vin
t (sec) x 10-10
VDD=2.5V
0.25m
W/Ln = 1.5
W/Lp = 4.5
Reqn= 13 k ( 1.5)
Reqp= 31 k ( 4.5)
tpHL = 36 psec
tpLH = 29 psec
so
tp = 32.5 psec
tf trtpHL tpLH
From simulation: tpHL = 39.9 psec and tpLH = 31.7 psec 20
What are those overshoots (next page)
© Digital Integrated Circuits2nd Inverter
The overshoots
The overshoots are
caused by gate-drain
capacitance of the
inverter transistors.
-0.5
0
0.5
1
1.5
2
2.5
3
0 0.5 1 1.5 2 2.5
© Digital Integrated Circuits2nd Inverter
Propagation Delay vs. Power Supply
Combining earlier equations:
For VDD>> VTn + VDSATn
Increasing VDD results in
reducing delay. Simulation
results (as shown) confirms
analytical results.
© Digital Integrated Circuits2nd Inverter
Reducing Propagation Delay Delay CL / (W/Ln k’n VDSATn )
Reduce CL
internal diffusion capacitance of the gate itself
– keep the drain diffusion as small as possible
interconnect capacitance
fanout
Increase W/L ratio of the transistor
the most powerful and effective performance optimization
tool in the hands of the designer
watch out for self-loading! – when the intrinsic capacitance
dominates the extrinsic load
Increase VDD
can trade-off energy for performance
increasing VDD above a certain level yields only very minimal
improvements
reliability concerns enforce a firm upper bound on VDD23
© Digital Integrated Circuits2nd Inverter
Inverter Propagation Delay, Revisited To see how a designer can optimize the delay of a gate
have to expand the Req in the delay equation
1
1.5
2
2.5
3
3.5
4
4.5
5
5.5
0.8 1 1.2 1.4 1.6 1.8 2 2.2 2.4VDD (V)
tpHL = 0.69 Reqn CL
= 0.69 (3/4 (CL VDD)/IDSATn )
0.52 CL / (W/Ln k’n VDSATn )
24
© Digital Integrated Circuits2nd Inverter
Inverter Sizing
25
© Digital Integrated Circuits2nd Inverter
Propagation Delay Issues
Design consideration and trade-offs
1. NMOS-to-PMOS Ratio
2. Sizing Inverters for Performance
3. Sizing a Chain of Inverters
4. Choosing the Right Number of Stages in
an Inverter Chain
5. The Rise-Fall of Input Signal
6. Delay with long interconnects
© Digital Integrated Circuits2nd Inverter
1. NMOS/PMOS Ratio
Widening the PMOS degrades the tpHL due to larger
parasitic capacitance
If speed is the only concern, reduce the width of the
PMOS device! See next slide
For equal rise and fall time: the PMOS and NMOS are sized so that the Req’s match (ratio of 3 to 3.5)
symmetrical VTC
equal high-to-low and low-to-high propagation delays
Does not lead to minimum overall delay
27
© Digital Integrated Circuits2nd Inverter
1. NMOS/PMOS Ratio: Optimum Ration
Reqp and Reqn : Ration of identically-sized PMOS and NMOS.
For our process, r=31K/13K = 2.4
For C w << C dn1 + C gn2
This is in contrast with ration of r.
PMOS/NMOS Ratio Effects
3
3.5
4
4.5
5
1 2 3 4 5
= (W/Lp)/(W/Ln)
x 10-11
of 2.4 (= 31 k/13 k)
gives symmetrical
response
Previous equation
predicts opt to be 1.6
From simulation, opt is
around 1.9! (close to our
analytical results.)
tpLH
tp
tpHL
29
© Digital Integrated Circuits2nd Inverter
2. Device Sizing for Performance Divide capacitive load, CL, into
Cint : intrinsic - diffusion and Miller effect (drain-gate overlap cap)
Cext : extrinsic - wiring and fanout
where tp0 = 0.69 Req Cint is the unloaded (intrinsic) delay of the gate
Widening both PMOS and NMOS by a factor S reduces Req by an identical factor (Req = Rref/S), but raises the intrinsic capacitance by the same factor (Cint = SCiref)
tp0 is independent of the sizing of the gate
any S sufficiently larger than (Cext/Cint) yields the best performance gains with least area impact. Yet, any sizing factor S that is sufficiently larger than (Cext/Cint) produces similar results at a substantial gain in silicon area
30
Sizing Impacts on Delay: Example
2
2.2
2.4
2.6
2.8
3
3.2
3.4
3.6
3.8
1 3 5 7 9 11 13 15
S
x 10-11 From example 5.5,
Cext/Cint = 3.15ff/3.0ff
1.05
The majority of the
improvement is already
obtained for S = 5. Sizing
factors larger than 10
barely yield any extra gain
(and cost significantly
more area).
for a fixed load
self-loading effect
(intrinsic capacitance
dominates)31
© Digital Integrated Circuits2nd Inverter
3. Sizing a Chain of Inverters
Sizing up transistor reduces its delay, but it increases
the Cext of the preceding gate.
Therefore, it is important to determine the optimum size
for inverters embedded in (for example) a chain of
inverters.
First determine the input loading effect of the inverter. Both Cg and Cint are proportional to the gate sizing:
• Cint = Cg So is independent of gate sizing (almost equal to 1)
• Define Effective Fanout (f ) = Cext/Cg
• tp = tp0 (1 + Cext/Cint)
= tp0 (1 + Cext/ Cg)
= tp0 (1 + f/)
• The delay of an inverter is a function of f. 32
© Digital Integrated Circuits2nd Inverter
Inverter Chain
If CL is given
How should the inverters be sized?
How many stages are needed to minimize the delay?
– May need some additional constraints.
Real goal is to minimize the delay through an inverter chain
the delay of the j-th inverter stage and the total delay are:
33
© Digital Integrated Circuits2nd Inverter
Sizing the Inverters in the Chain The optimum size of each inverter is the geometric mean
of its neighbors – meaning that if each inverter is sized up
by the same factor f wrt the preceding gate, it will have
the same effective fan-out and the same delay
where F represents the overall effective fan-out of the circuit
(F = CL/Cg,1)
and the minimum delay through the inverter chain is:
The relationship between tp and F is linear for one
inverter, square root for two, etc.
34
© Digital Integrated Circuits2nd Inverter
Example of Inverter Chain Sizing
CL/Cg,1 has to be evenly distributed over N = 3 inverters
CL/Cg,1 = 8/1
f =
In Out
CL = 8 Cg,1Cg,1
Cg= Cg,1 Cg= 2 Cg,1 Cg= 4 Cg,1
38 = 2
36
© Digital Integrated Circuits2nd Inverter
4. Optimum Number of Stages
Examining the above delay equation reveals the the trade-off’s
in choosing the number of stages (N) for a given F (=fN).
• When the number of stages is too large, the first
component of the equation, representing the intrinsic delay
of the stages, becomes dominant.
• If the number of stages is too small, the effective fanout of
each stage becomes large, and the second component is
dominant.
37
© Digital Integrated Circuits2nd Inverter
4. Optimum Number of Stages
The optimum value can be found by differentiating
the minimum delay expression by the number of
stages, and setting the result to 0.
ff 1exp
38
0ln/1
/1 N
FFF
NN
This equation only has a closed-form solution for = 0
• = 0 No self-loading (i.e. Cint=0)
•This is when the self-loading is ignored and the load capacitance
only consists of the fanout.
• Under these simplified conditions, it is found that the optimal
number of stages equals N = ln(F), and the effective fanout of
each stage is set to f = 2.71828 = e
© Digital Integrated Circuits2nd Inverter
Optimum Effective Fanout f: ≈1When self-loading is included,
•The equation can only be
solved numerically.
•The results are plotted below.
• For ≈ 1, fopt = 3.6
39
© Digital Integrated Circuits2nd Inverter
Optimum Effective Fanout f: ≈1
40
The plots the (normalized)
propagation delay of the inverter
chain as a function of the effective
fanout ≈1 .
• Choosing values of the fanout that are
higher than the optimum does not impact the
delay that much, and reduces the required
number of buffer stages and the
implementation area. A common practice is
to select an optimum fanout of 4.
• The use of too many stages (f < fopt), on the
other hand, has a substantial negative
impact on the delay, and should be avoided.
© Digital Integrated Circuits2nd Inverter
Example of Inverter Staging ( =1 )
CL = 64 Cg,1Cg,1 = 1
1
CL = 64 Cg,1Cg,1 = 1
1 8
CL = 64 Cg,1Cg,1 = 1
1 4 16
CL = 64 Cg,1Cg,1 = 1
1 2.8 8 22.6
N f
1 64 65
2 8 18
3 4 15
4 2.8 15.3
41
/10N
pp FNtt
topt/tp0 versus x for various driver configurations
Impressive speed-ups with optimized
cascaded inverter chain for very large
capacitive loads.
F ( = 1)
Unbuffered Two Stage Chain
Opt. Inverter Chain
10 11 8.3 8.3
100 101 22 16.5
1,000 1001 65 24.8
10,000 10,001 202 33.1
/10N
pp FNtt
42
Sizing An Inverter Network N=3
CL_total = CL * 16 = 64*16 = 1024
f = (CL_total)1/N = 10.08
The effect gate size of stage2
10.08
The effective gate size of stage2
101.6
However, a stage1 inverter drives 4
stage2 inverters
Cg2 = 10.08/4 = 2.52
Similarly, the size of stage3 inverter
is
Cg3 = 101.6/16 = 6.35
CL = 64 * Cg1
© Digital Integrated Circuits2nd Inverter
5. Input Signal Rise/Fall Time In reality, the input signal
changes gradually (and both
PMOS and NMOS conduct for a
brief time). This affects the
current available for
charging/discharging CL and
impacts propagation delay.
3.6
3.8
4
4.2
4.4
4.6
4.8
5
5.2
5.4
0 2 4 6 8
ts(sec)
x 10-11
x 10-11
for a minimum-size inverter
with a fan-out of a single gate
tp increases linearly with increasing input slope, ts, once ts > tp
ts is due to the limited driving capability of the preceding gate
44
© Digital Integrated Circuits2nd Inverter
Design Challenge A gate is never designed in isolation: its performance is
affected by both the fan-out and the driving strength of
the gate(s) feeding its inputs.
tip = tistep + ti-1step ( 0.25)
Keep signal rise times smaller than or equal to the gate propagation delays.
good for performance
good for power consumption
Keeping rise and fall times of the signals small and of approximately equal values is one of the major challenges in high-performance designs - slope engineering.
45
© Digital Integrated Circuits2nd Inverter
6. Delay with Long Interconnects When gates are farther apart, wire capacitance and
resistance can no longer be ignored.
tp = 0.69RdrCint + (0.69Rdr+0.38Rw)Cw + 0.69(Rdr+Rw)Cfan
where Rdr = (Reqn + Reqp)/2
= 0.69Rdr(Cint+Cfan) + 0.69(Rdrcw+rwCfan)L + 0.38rwcwL2
cint
Vin
cfan
(rw, cw, L)Vout
Wire delay rapidly becomes the dominate factor (due to the quadratic term) in the delay budget for longer wires.
46
© Digital Integrated Circuits2nd Inverter
Power Dissipation
© Digital Integrated Circuits2nd Inverter
Why Power Matters
Packaging costs
Power supply rail design
Chip and system cooling costs
Noise immunity and system reliability
Battery life (in portable systems)
Environmental concerns
Office equipment accounted for 5% of total US commercial
energy usage in 1993
Drain leakage will increase as VT decreases to maintain noise
margins and meet frequency demands, leading to excessive battery
draining standby power consumption.
48
© Digital Integrated Circuits2nd Inverter
Why worry about power? -- Power Dissipation
P6Pentium ®
486
386
2868086
80858080
80084004
0.1
1
10
100
1971 1974 1978 1985 1992 2000
Year
Po
we
r (W
att
s)
Lead microprocessors power continues to increase
Power delivery and dissipation will be prohibitive
Source: Borkar, De Intel49
Chip Power Density Distribution
Power density is not uniformly distributed across the chip
Silicon is not a good heat conductor
Max junction temperature is determined by hot-spots Impact on packaging, w.r.t. cooling
0
50
100
150
200
250
Heat
Flu
x (
W/c
m2)
Willamette Power Distribution
200-250
150-200
100-150
50-100
0-50
Power Map
40
60
80
100
120
Tem
pera
ture
(C
)
Al-SiC+ Epoxy Die Attach
100-120 80-100
60-80 40-60
On-Die Temperature
50
© Digital Integrated Circuits2nd Inverter
Power and Energy Figures of Merit Power is the rate of consuming energy
It is measured in Watts
power is a rate (e.g. speed)
Peak power
– determines power ground wiring designs
– sets packaging limits
– impacts signal noise margin and reliability analysis
Average Power determines battery life
Energy is a quantity and measured in Joules
Energy = power * time
Joules = Watts * seconds
lower energy number means less power to perform a
computation at the same frequency
51
© Digital Integrated Circuits2nd Inverter
Power versus Energy
Watts
time
Power is height of curve
Watts
time
Approach 1
Approach 2
Approach 2
Approach 1
Energy is area under curve
Lower power design could simply be slower
Two approaches require the same energy
52
© Digital Integrated Circuits2nd Inverter
PDP and EDP Power-delay product (PDP) = Pav * tp = (CLVDD
2)/2
PDP is the average energy consumed per switching event
(Watts * sec = Joule)
lower power design could simply be a slower design
allows one to understand tradeoffs better
0
5
10
15
0.5 1 1.5 2 2.5
Vdd (V)
En
erg
y-D
elay (n
orm
alized
)
energy-delay
energy
delay
Energy-delay product (EDP) = PDP * tp = Pav * tp2
EDP is the average energy consumed multiplied by the computation time required
takes into account that one can trade increased delay for lower energy/operation (e.g., via supply voltage scaling that increases delay, but decreases energy consumption)
53
© Digital Integrated Circuits2nd Inverter
Where Does Power Go in CMOS?
• Dynamic Power Consumption
• Short Circuit Currents
• Leakage
Charging and Discharging Capacitors
Short Circuit Path between Supply Rails during Switching
Leaking diodes and transistors
54
© Digital Integrated Circuits2nd Inverter
Dynamic Power
Energy = Q * VDD = CL * VDD2
Energy/transition =CL * VDD2 * P01
Pdyn = Energy/transition * f = CL * VDD2 * P01 * f
Pdyn = CEFF * VDD2 * f where CEFF = P01 CL
Vin Vout
CL
Vdd
f01
55
© Digital Integrated Circuits2nd Inverter
Transistor Sizing for Minimum
Energy
Goal: Minimize Energy of whole circuit
Design parameters: f and VDD
tp tpref of ref circuit, i.e. f=1 and VDD =Vref
1Cg1
In
fCext
Out
TEDD
DDp
pp
VV
Vt
f
Fftt
0
0 11
56
© Digital Integrated Circuits2nd Inverter
Transistor Sizing (2)
Performance Constraint (=1)
Energy for single Transition
1
3
2
3
2
0
0
F
f
Ff
VV
VV
V
V
F
f
Ff
t
t
t
t
TEDD
TEref
ref
DD
refp
p
pref
p
F
Ff
V
V
E
E
FfCVE
ref
DD
ref
gDD
4
22
11
2
1
2
57
© Digital Integrated Circuits2nd Inverter
1 2 3 4 5 6 70
0.5
1
1.5
2
2.5
3
3.5
4
f
vd
d (
V)
1 2 3 4 5 6 70
0.5
1
1.5
f
no
rma
lize
d e
ne
rgy
Transistor Sizing (3)
F=1
2
5
10
20
VDD=f(f) E/Eref=f(f)
58
© Digital Integrated Circuits2nd Inverter
Short-Circuit Consumption
(Dissipation due to Direct-Path
Current)
Finite slope of the input signal causes a direct
current path between VDD and GND for a short
period of time during switching when both the
NMOS and PMOS transistors are conducting.
Vin Vout
CL
Isc
© Digital Integrated Circuits2nd Inverter
Short Circuit Currents Determinates
Duration and slope of the input signal, tsc
tsc (Vdd- 2 Vt)/Vdd x tr/0.8
Ipeak determined by
the saturation current of the P and N transistors which
depend on their sizes, process technology, temperature, etc.
strong function of the ratio between input and output slopes
– a function of CL
Esc = tsc VDD Ipeak P01
Psc = tsc VDD Ipeak f01
60
© Digital Integrated Circuits2nd Inverter
Impact of CL on Psc
Vin Vout
CL
Isc 0
Vin Vout
CL
Isc Imax
Large capacitive load
Output fall time significantly larger than
input rise time.
Input moves through the transient region
before the output starts to change. As a
result, the PMOS shuts-off before
delivering much current.
Small capacitive load
Output fall time substantially smaller than
the input rise time.
The drain-source voltage of PMOS is equal
to Vdd most of the transient period,
guaranteeing maximum current. 61
Ipeak as a Function of CL
-0.5
0
0.5
1
1.5
2
2.5
0 2 4 6
time (sec)
x 10-10
x 10-4
CL = 20 fF
CL = 100 fF
CL = 500 fF
500 psec input slope
Short circuit dissipation
is minimized by
matching the rise/fall
times of the input and
output signals - slope
engineering.
When load capacitance
is small, Ipeak is large.
62
Psc as a Function of Rise/Fall Times
0
1
2
3
4
5
6
7
8
0 2 4
P(normalized wrt zero input rise-time dissipation)
tsin/tsou
t
VDD= 3.3 V
VDD = 2.5 V
VDD = 1.5V
* tsin/tsou > 2
- Small cap loade
- High Psc
* tsin/tsou < 2
-Large cap loade
-Low Psc
If VDD < VTn + |VTp| then
Psc is eliminated since
both devices are never
on at the same time.
W/Lp = 1.125 m/0.25 m
W/Ln = 0.375 m/0.25 m
CL = 30 fF 63
Leakage (Static) Power Consumption
It is static power dissipation caused by
Pstat = Istat Vdd
Sources of the static current:
1. Reverse-biased diode junction between source/drain and substrate
• Devices with high Vt leaks less.
2. Sub-threshold current: Drain-to-source current when -0<Vgs <Vt.
• Devices with higher threshold has less sub-threshold leakage current.
Choosing the right Vt is trade-off between performance and leakage power
© Digital Integrated Circuits2nd Inverter
Leakage (Static) Power Consumption
Sub-threshold current is the dominant factor.
All increase exponentially with temperature!
VDD Ileakage
Vout
Drain junction
leakage
Sub-threshold currentGate leakage
65
Leakage as a Function of VT
0 0.2 0.4 0.6 0.8 1
VGS (V)
ID (
A)
VT=0.4V
VT=0.1V
10-2
10-12
10-7
Continued scaling of supply voltage and the subsequent scaling of threshold voltage will make subthreshold conduction a dominate component of power dissipation.
An 90mV/decade VT
roll-off - so each 255mV increase in VT gives 3 orders of magnitude reduction in leakage (but adversely affects performance)
66
Exponential Increase in Leakage Currents
1
10
100
1000
10000
30 40 50 60 70 80 90 100 110
0.25
0.18
0.13
0.1
Temp(C)
I lea
kag
e(n
A/
m)
From De,1999
67
© Digital Integrated Circuits2nd Inverter
CMOS Energy & Power Equations
E = CL VDD2 P01 + tsc VDD Ipeak P01 + VDD Ileakage
P = CL VDD2 f01 + tscVDD Ipeak f01 + VDD Ileakage
Dynamic
power
Short-circuit
power
Leakage
power
f01 = P01 * fclock
68
© Digital Integrated Circuits2nd Inverter
Power measures and metrics Power-Delay Product:
PDP = Pav tp== CL * VDD2 /2
Measures the average energy per switching event
This metric can be made to “ look good” by lowering Vdd on
the expense of performance.
Energy-Delay Product
EDP = PDP tp * = tp CL * VDD2 /2
In EDP, higher voltage improves performance but hurt energy,
and the opposite is true for low voltage.
The optimum voltage can be obtained by taking the derivative
of EDP with respect to Vdd,
VDDopt = 1.5 * VTE
where VTE = VT + VDSAT/2
© Digital Integrated Circuits2nd Inverter
Analyzing Power Using Spice