Transcript lecture-05

5 CMOS Logic Basics(1)-combinational logic
Contents
1. Dynamic NMOS logic
2. Pseudo NMOS logic

Transistor sizing

Complex logic gate layout
3. Static(complementary) CMOS logic

Inverter, NAND, NOR gates layout
4. Pass transistor logic
5. BiCMOS logic
6. Dynamic CMOS logic
5.1
1. Dynamic NMOS Circuits
PMOS Static
+area, +speed
NMOS Static
+power
-area
-speed
CMOS Static
+power
+area
(utilize para. C)
+area
+speed
NMOS Dynamic
CMOS Dynamic
Q. Why we study dynamic NMOS which is not popular ?
(compared to CMOS)
A. Simple vehicle for understanding dynamic operation.
5.2
i) 2 phase, ratioed type
VDD
 
 

2
1
2
1
2
Vi
C1

1
V2 i
VC1
VC2
VC3
C2
C3
Low level is
determined from
inverter ratio
demands
large chip area!
5.3
ii) 2 phase, ratioless A
1
VDD
2
Vi
C2
C1
C4
C3
1
2
DV = C3  VDD : Charge sharing between
C2+C3
C2 and C3
Vi
VC1
VC2
Ta
Tb
VL
Ta: DC power consmption due to
ratioed operation
Tb: ratioless operation
DV
VC3
VC4
VL
O
5.4
iii) 2 phase ratioless B
VDD
1
2
C2
T1
Vi
C1

C3
C4
Vi가 high이고 1이 on 일때 VDD-GND DC경로를 표 한 T1이 차단
DV 
C 3 C 4
V
C 2 C 3 C 4 DD

Charge Sharing에 의한 전압강하

C기생성분의 증가로 CV2f(dynamic 전력소모) 증가
5.5
iv) 2 phase ratioless C
1
2

No DC path

extremely large CV2f dissipation
and large clock driver

simple layout
Vi
1
2
5.6
v) 4 phase ratioless A
1
2
2
3
1
Type 2
1
2
3
4
3
4
2
Type 3
4
1
3
Type 4
4
Type 1
Type 1
2
3
4
1
2
3
4
E
P
H
H
H
E
P
H
H
H
E
P
P
H
H
E
E : evaluate
P : precharge
H : hold
5.7
vi) 4 phase ratioless B




1
2
1
2
T1
x


2
3
12
23
y




1
2
1
2
장 * same as 5. except T1 which prevents charge sharing between x and y
improves noise immunity
5.8
2. Pseudo NMOS logic
 A kind of ratioed logic as R-load, depletion-load.
 No body effect
VDD
IL
grounded-gate PMOS is close to constant current
source load
pseudo NMOS
const. I src
IL
depl.
Vo
PDN
VOL is given by
R
Load line curve
2

 kp
VOL
2
k n (VDD  VTn )VOL 

(
V

V
)

DD
Tp
2
2


 VOL
enh.
Vo

Kp 

, assuming VT  VTn  VTp
 (V DD  VT )1  1 

Kn


5.9
Pseudo-NMOS vs. Complementary CMOS
 (Pros.) of Pseudo -NMOS
1. Smaller number of transistors
less area, less capacitive loading to preceding gate
 (Cons.)
1. Power consumption due to DC current path:
Pav (low)  VDD  I low 
kp
2
VDD (VDD  VT ) 2
low NML( i.e., high low-output VOL=ILRPDN)
 (Comments)
Pseudo-NMOS is desirable for applications where speed is of major
importance, or where we know that the majority of outputs are high, such
as address decoder in memory
5.10
Design constraints of pseudo-NMOS
1. Power consumption : to reduce static power
2. noise margin: to obtain reasonably high NML, VOL=ILRPDN should be
low.
3. output rise time
4. output fall time
RPU large
1. Power consumption
2. Noise margin
RPDN small
3. Rise time
4. Fall time
RPU small
: contradictory
5.11
Answer to Question
Q1. How do we consider capacitance to VDD ?
A1. Capacitance to VDD and all other DC nodes are lumped as
capacitance to ground.
Charge variation on node x due to DVX
DQX = C G N D  [(V x  DV x )V x ] C VD D [(V X  DV x V D D ) (V x V D D )]
= C G N D  DV x  C VD D  DV x
=
VDD
(C G N D  C VD D ) DV x
C eff  C G N D  C VD D
x
Ceff
5.12
Q2. Name a circuit where capacitance ratio, not value, determines the
behaviour.
A2. Switched-capacitor filter
i
V1
V2
i
f
f
C1
C2
V1
V2
v1 V 2
1
i
 fc (V1 V 2 ) R 
R
fc
F (S ) 
1
SRC 2
fC 1

SC 2
e-ratio
5.13
 Variations of pseudo-NMOS
i) parallel large adaptive PMOS load
In pseudo-NMOS tPLH tends to be rather high to make VOL low
enough(i.e., NML high enough).
In standby, large PMOS
pull-up, M1 is off, which is
turned on after an address
translation is detected to
result in large current drive.
5.14
Ex.) Pseudo-NMOS layout example(4-input NAND)
(W/L)NMOS = 1.8/1.2
(W/L)PMOS = 1.8/4.8
5.15
Istatic = 48.5A, VOH = 5V, VOL = 0.48V
VM = 1.62V, NMH = 3.26V, NML= 0.27V
tpHL = 0.6nsec, tpLH = 2.37nsec
Trying to make PMOS longer to increase NML yields even worse tpLH,
which is already large.
If we change to (W/L)NMOS = 7.2/1.2, (W/L)PMOS = 1.8/1.2,
Istatic
0.18mA, tpHL
0.24ns, tpLH
1.4ns
5.16
ii) CMOS Multi-drain Logic

This idea came from I2L(Integrated-Injection Logic).

Low energy due to low switching voltage

NOR-type logic only
Q
i
n
n
p
I
I
p
Q
i
5.17
Z=A+B
A
B
Q. Why is I2L not used much ?
A. 1) efficiency of PNP current source
2) CMOS is better in area, power, speed
5.18
A
Z=A+B
B

A+B
Fanout limited by NML
5.19
iii) Ganged CMOS logic
A
B
C
Z

NOR gate ; Z = A+B+C

Can be operated as NAND gate by suitably ratioing PMOS over
NMOS.

Can be operated as quarternary logic
5.20
iv) Cascode Logic(1) : CVSL

Q
CVSL(Cascode Voltage Switch Logic) or DCVSL(Differential
Cascode Voltage Switch Logic)
Q
Q
Q
a
a
b
...
...
b
a
b
c
c

CVSL : requires dual signal rail ratioed logic(DC path exists) large
current spike during switching regenerative feedback
fast(?)
5.21

Common factor sharing between Out and Out
5.22
(Another example for sharing)
Q  abcd
5.23

Dynamic CVSL
5.24

SSDL(Sample-Set Differential Logic) : modified version of dynamic CVSL

during clk low ; one at VDD, the
other at slightly below VDD

during clk high : set to VDD &
VSS quickly
5.25
v) Cascode Logic(II) : DSLL(Differential Split Level Logic)

replaces p-load in pseudo-NMOS with cross-coupled cascoded n-and
PMOS
out
out
VDD
 VTN
set VR 
2
VD D
I
N
:
VX  VTN   , when
2
IN :0 V
VDD , when IN
VX
VD D
 0 , IN :
2
PMOS, NMOS의 size를 이 조건에
맞도록 정한다.
VR
in
in
5.26
A
t1
P1
in
t3
n1
P3
out
P4
out
C1
in
t2

n2
P2
A
VR
C2
VR
V
DD
(

) 일때
in : low(0V), in : high
2
P2 is fully ON, A is pulled to VDD, C2 is discharged
P1 is not fully OFF, A is at 0.5-0.6V(just below VTN of n1)
V
C1 is charged to D D via t3, P3
2
VD D
 Switch in to
and in to 0V.
2
P1 turns on very fast(
VDD, A

A

t2 turns on very fast(
it wasn’t completely OFF)
it was just under threshold)
0.5-0.6V
5.27
Size of PMOS is set indep. of pull-down NMOS, as they are decoupled by
cascode NMOS.
Therefore, charging through PMOS is faster than in pseudo-NMOS.

Repartitioning the gate to open-drain
structure, where input and output resides
in(0,
VD D
2
), this improving speed.


( Voltage swing in large C is reduced.)
5.28
iv) SPFL(Source Follower Pull-uP Logic)

extensive use of NMOS over slower PMOS

utilizing parallel connection over slower series connection
good for NOR gate

P1 is turned off by anyone of N1-N4, connected as paralled source follower,
turned on

suffers from body effect and degraded VH in the internal node(Nload).
5.29
3. Static(Complementary) CMOS
 inverter layout

PMOS drain(p-diffusion) and NMOS drain(n-diffusion) can be connected
via metal strap and contacts.

Power and ground are run in metal and connected to the sources of PMOS
& NMOS.
5.30
 Various layout styles of inverter
5.31
 Various layout styles of large-W/L inverter(driver)
5.32
 Static CMOS NAND gate

default : horizonal PWR, GND/vertical polysilicon gate
5.33
 Static CMOS NOR gate

(b) is faster than (a) Why ?
5.34
 Transistor sizing for 8-input NAND gate
Delay(ns) (before/after sizing) area(before/after sizing)
Appr. 1
2.82 + 3.37 = 6.2  4.6
216  120
Appr. 2
0.88 + 4.36 = 5.24  4.8
216  136
Appr. 3 0.31+0.4+0.31+2.17=3.193.5
360  124
5.35
 Complex gate layout
i) Linear Array

Think of an AOI(And-OR-Invert) cell : f = X  X  X  X  X
1
2
3
4
5
1
2
3
4
5
VDD
f
f
In static CMOS circuit
5.36

Since each P-N pair(sharing the same input) is represented by
P-MOS
poly-Si
Symbolize by
N-MOS
a’
a
active
area
contact
aluminum
N
N
P-well
5.37

VDD
Symbolic Layout of AOI gate
1
2
3
4
5
VDD
1
2
3
4
5
out
VSS
3
5
2
4
1
(Layout ‘a’)
VSS
Noting that
5.38

Layout ‘a’ can be simplified
1
VDD
2
3
4
5
VDD
P
N
1
out
2
3
VSS
out
4
5
VSS
out
(Layout ‘b’)

If we interchange 2 and 3
Observation : 1-3-2-4-5 is and
Euler path for N-graph. But
not for P-graph.
VDD
out
VSS
1
3
2
(Layout ‘c’)
4
Ques. How to find a sequence
which is an Euler path for
both P-and N-graph.
5
5.39
Ans. Sequence 2-3-1-4-5- is it!
VDD
2
Ques. How do we find it consistently?
3
1
2
3
1
4
4
out
5
3
out
5
1
5
2
(a)
4
VSS
(b)
VDD
2
VSS
3
2
3
1
4
5
VDD
out
out
1
4
P-part
N-part
5
VSS
(Layout ‘d’)
5.40

Euler path methold
2
2
3
1
4
5
3
1
4
5
2
2
3
2
3
5
4
Stretch
3
5
2
3
1
1
Split
contact
4
1
4
4
5
5
(rubber band)
5.41
1. Find all Euler paths that cover the graph.
2. Find a p- and an n-Euler path that have identical labeling(a labeling is
an ordering of the gate labels on each vertex).
3. If the paths in step 2 are not found, then break the gate in the minimum
number of places to achieve step 2 by separate Euler paths.
5.42
관찰 1) Euler path가 존재하면 그 path에 포함된 모든 edge에 대응하는
gate는 diffusion 영역을 공유하면서 연결된다.
관찰 2) 전체 edge를 지나는 Euler path가 존재하지 않으면 전체 graph
를 Euler path를 갖는 subgraph로 분할한다.(여기서 모든 Euler
path는 p-와 N-graph 모두에게 존재하는 것을 뜻한다.) 이때,
각 subgraph는 layout에서 각 diffusion island(혹은, interlace)를
형성하고, 각 interlace 사이에는 gap이 있다.
관찰 3) Reduction of graph ; 홀수개의 edge로 된 series, 혹은 parallel
graph는 single edge로 줄여서 Euler path를 찾은 후, 원래 graph
상의 Euler path로 복원시킬 수 있다.
관찰 4) make the # of inputs to all gates to be odd, by inserting dummy
input, so that E-graph can always be found.
5.43

Graph reduction fro finding Euler path
(복원 2)
(복원 1)
(축약 1)
example.
1
2
3
4
5
6
7
(축약 2)
,
8
10
9
5.44

Graph model
reduction phase
(축약 2)
(축약 1)
E-path reconstruction
1
1
4 3
2
(복원 1)
(복원 2)
10
8
5
7
9
6
5.45

Appending pseudo input
ex.
1
P1
2
3
P2
4
5

8
P1, P2 : pseudo input
10
9
Heuristic algorithm ;
1. Add pseudo-input to every gate with an even number of inputs.
2. Rearrange the input sequence such that # of interlaces between real
and pseudo inputs is minimal. Top and bottom pseudo inputs
does not contribute to separation areas.
P1 2 3
P1
2
3
1
4
5
P2
8
1
10
9
4
5
P2
5.46

Minimal Interlace Algorithm
White ; pseudo input
Black ; real input
White & Black ;
gate with both pseudo
and real input
5.47
 When a signal is applied to multiple transistors.
ex) CMOS XNOR gate
5.48
ii) Gate Matrix
5.49
 Gate Matrix layout algorithm

Transistors are grouped in strips to allow maximum source/drain
connection by abutment. To achieve better grouping, polysilicon columns
are allowed to interchange to increase abutment.

The resultant groups are then placed in rows with groups maximally
connected to the VSS and VDD rails placed toward these signals. Row
placement is then based on the density of other connections.

Routing is achieved by vertical diffusion or Manhattan(horizontal and
vertical) metal routing. This normally would require a maze router.
5.50
4. Pass Transistor Logic
(or Transmission Gate Logic)
 AND, OR function can be implemented using series- and parallelconnected switches.
 Other combinational logic: series and parallel connection
 General concept
 example
A
Out
Switch
Network
X
B
necessary for grounding X when B=0
 MOSFET is a “switch”, but with some nonideality,
i.e. finite threshold voltage(+body effect), and conduction dependent
on Vgs
5.51
 Threshold loss problem
C=5V
A=5V
B can be charged up to | VDD-VT |,
where VT 1.5V due to body effect

Moreover, charging B is very slow
at the end due to reducing | VGS |
B
CL
C=5V
A=5V

B
M2
M1

restoring level using inverter

voltage at B is not high enough to
turn M2 off. Therefore, leakage
current flows through M2-M1
5.52
 Solution
Conductance
C=5V
total
B
A=5V
gDS,P
CL
gDS,N
C=0V
PMOS: transmits HIGH well
NMOS: transmits LOW well
0
VTP
Vin ,Vout
VTN
VDD

As B goes HIGH(C=HIGH), VGSP is -5V invariably, while VGSN
approaches VTN

When A is LOW(C=HIGH), VGSN is 5V invariably, while VGSP
approaches VTP
5.53
 Cascade connection of pass transistors
V
V
V
V
Vo
V
R
R
C


R
C
R
C
C
= RC • n(n+1)/2, n: number of stages(transistors)
 1/2 • n2 • RC
Vo can reach up to V-VTN
V
V
V
V

Vo(high) = V-3VTN

Output of pass transistor better
not be used as control variable
for another gate.
Vo
5.54
 Cases where pass tr. is appropriate

Multiplexer
S
F=A • S+B • S
A
requires 8 tr’s vs 6 here
S
B
S

XOR gate
A
B
B
A
F
B
B
5.55
 4-input Multiplexer using CMOS switch concept
5.56
 Rules for transmission gate logic construction 1
•A conducting path can never exist between two
different inputs which could take different logic
levels.
•If there is an overlap between 1 and 2, the
intermediate node “y” will take an undefined potential,
1
X1
2
X1
Y
located between the 1 and 0. This potential will give
rise to an erratic behavior of the logic, although it may
not be detected by a switch-level simulator
•This problem can be solved by designing control
signal with a mutual
exclusion feature
5.57
 Rules for transmission gate logic construction 2
•When a branch has several transmission-gates in series, internal nodes
can behave as a dynamic memory
•Such a gate cannot be considered as a pure static combinational logic gate,
because the memory effect can give rise to false outputs, according to the
history of successive inputs and control levels, Moreover, the output could
be at high impedance if no buffer has been provided.
•This behavior dramatically increases the simulation and test problems
5.58
 Rules for transmission gate logic construction 3
•To avoid undesired high impedance states, care should be taken to
always provide at least one conducting path between an input and
the output
•The input variable sources must be low-impedance sources for the
same reason
a
a
Ex) 1-to-2 decoder
1
a
X1
X1
1
a
X2
0
wrong
X2
good
5.59
 Formal design procedure for transmission-logic

Generalized form of a transmission-gate logic
V1
V2
A1
A2
Aj
B1
B2
Bj
X1
X2
Xj
Z1
Z2
Zj
F
Vl
Vn
V1
P1
V2
P2
Vl
Pl
Vn
Pn
F
F=P1(V1)+P2(V2)+…+Pn(Vn)
with Pi=X1 • X2 • … • Xj
5.60
 Design procedure using Karnaugh map
5.61
 Design procedure using Karnaugh map
5.62
 Design procedure using Karnaugh map
5.63

Ex) Comparator
An-1 An-1
Cn((=0)
Dn(=0)
Aj
Aj
A0
A0
Cn Cj+1
Cj C1
C0
Dn Dj+1
Dj D1
D0
Bn-1 Bn-1
Bj
Bj
B0
B0
1) Cn=0, Dn =0
3) C0 becomes 1 when A>B
2) For cell I,
D0 becomes 1 when A<B
Ci =1 when Ci+1 =1, or
C0 and D0 are both 0 when A=B
Ai > Bi and Di+1 =0
Di =1 when Di+1 =1, or
Ai > Bi and Ci+1 =0
5.64
Ci  Ci 1  Ai  Bi  Di 1
AiBi
00
A
00
A
01
B
B
01
A
C
D
A
C
B
B
D
11
A
C
D
A
C
B
B
D
10
A
C
D
A
C
B
B
D
C
C
D
D
Ci+1Di+1
d
11
not
exist
A
10
B
d
A
C
D
B
d
A
C
D
B
d
A
C
D
B
C
D
When Ai  Bi  1, pass Ci 1
When Ai  Bi  1, pass Di 1
5.65
Di  Di 1  Ai  Bi  Ci 1
AiBi
00
A
00
A
01
B
B
01
A
C
D
A
C
B
B
D
11
A
C
D
A
C
B
B
D
10
A
C
D
A
C
B
B
D
C
C
D
D
Ci+1Di+1
d
11
not
exist
A
10
B
d
A
C
D
B
d
A
C
D
B
d
A
C
D
B
C
D
When Ai  Bi  1, pass Di 1
When Ai  Bi  1, pass Ci 1
5.66

Resultant I-th layout
Ai
Ci+1
Ai
Bi
Bi
Ci
Di+1
Di
5.67
5. BiCMOS logic
 Comparison of CMOS vs. Bipolar ECL
CMOS
CMOS
ECL
Adv.
high, symm noise margin
high input impedance
high gain in transition region
high packing density
low power consumption
high current drive per unit area
high switching speed
low I/O noise
Low speed(esp. for large C)
High power consumption
large area
low input impedance
smaller noise margin
Disad.
5.68
 ECL(Emitter-Coupled Logic)
Vo ,1  VCC  VBE ( on)  RC I C 1
V 2  VCC  VBE ( on)  RC I C 2
Vin  Vref
I C1
 exp(
)
IC 2
VT
I C1  I C 2 I
EE
5.69
 Deriving ViH, ViL for noise margin
(Define ViL as Q1 carrying 1% of IEE, while Q2 99%)
I C1
ex

 0.01( X  Vref  Vin L )
x
I EE 1  e
V iL ,iH  V ref  VT
ln (

1 
)
ViL=Vref-120mV, ViH=Vref +120mV (when VT=26mU)
narrow transition region of 240mV.
VO 2
VCC-VBE(on)(on)
 0.24V
VCC-VBE(on)-IEERC
VO1
ViL
Vref
ViH
Vin
5.70
 Cross-section of BiCMOS Process

n-epitaxial layer on P-substrate
(n-epi. For n-well of PMOS & collector of npn BJT)

n+-buried layer(before epi.)
to reduce collector resistance and to increase latchup immunity
5.71
 Basic Function of BiCMOS gate
is
better be
M2 Q2 M1 Q1

When Vin = 0
on
on


When Vin = 1


on on
Z2
Z1
off on
on off
Z2, Z1 : necessary to remove base charge for fast turn-off, also to prevent
instantaneous DC path with both Q1, Q2 on.
5.72

Increased power consumption due to reduced swing
5.73

Transient behaviour of BiCMOS inverter
5.74

Combinational logic gate in BiCMOS
VDD
i) When Vi=1, Vo must be 0
For that, f must be ON
f
Vi
A
and f must be OFF
Q1
Z1
f : PMOS, f : NMOS
ii) Then, Z1 better be ON
VO
and Z2 must be OFF
Z1 : same as f (NMOS)
f
Q2
Z 2 : Z1
control input from A
Z2
5.75
 Several Topologies of BiCMOS inverter
(a) : basic (b) : Similar(PMOS pulldown:slower) (c) Vo reaches VDD
5.76
(d) Feed Z1 into the Q2 base
 Utilize charge removed
from Q1-base to turn on Q2.
5.77
(e) BiCMOS with full rail output(GND, VDD)
VDD
speed up charge
MP
Q1
M1
R1
 M1 can be added to
removal from Q1-base
Vo
Mn
Vi
R2
Q2
5.78
 NAND2 BiCMOS gate
 For large fan-in,
discharging Q1-base is
slow due to many
parallel PMOSFET’S.
5.79
 NOR2 BiCMOS gate
 NAND is preferred over
NOR in BiCMOS logic,
because charging through
series PMOSFET’S is
very slow
5.80
 BiCMOS AOI/OAI logic
5.81
 BiCMOS to CMOS interface circuit
 One can connect PMOS
output and VDD.
 Connecting from Q1-base
to feedback inverter is
faster than connecting
from Q1-emitter
5.82
Comparison of CMOS & BiCMOS
VM
tP ,C M O S  t0 
C L
ID
VM
tP ,B iC M O S  t1 
C L
ID
t0, t1 : delay from input to output due only to internal capacitance
t1
t0
t1> t0 (
CMOS
BiMOS
CL

tP
In BiCMOS, internal node
capacitance and delay due to additional delay
are larger.)
5.83
6. Dynamic CMOS Logic
 Dynamic vs. Static scheme (for charge storage)

Static storage due to static FF(positive feedback)

Stored value remains valid as long as power is supplied

In order to change state, a pulse from lower Zo(output impedance) source
(than the Zi of FF) for a minimal length equal to the propagation delay of
2-stage inverter chain must be applied

Dynamic charge storage(on capacitor) depends on leakage current 
requires periodic refresh

Dynamic charge storage requires high input impedance device for good
readout, which is available in CMOS, not in Bipolar

Therefore, dynamic is impossible in Bipolar, while in CMOS, both static
and dynamic are possible
5.84
 Pseudo-static latch

D
in
D



 high : transparent mode
 low : storage mode(bistable action)
5.85
 Static latch

in
R
R > Ron (pass Tr.) for overwrite
 Condition for static charge storage
i) Cross-coupling with  1 gain
ii) Sufficient node capacitance (otherwise corrupted by small noise)
5.86
 Master-slave D-FF is built by cascading two pseudo-static latches
clocked opposite

Problem occurs when  &  overlaps due to clock skew;
i) node A is undefined as driven by both In and B
ii) Input signal can race through both master and slave FF if  and  are
both high for long enough time
5.87
 Using nonoverlapping clock(1 , 2)

Nonoverlap interval(t 12) needs to be long enough to guarantee non
overlap despite clock skew. (Maintaining two nonoverlapping clocks all
over the chip is very hard)

During nonoverlap period, storage mechanism is dynamic, while static
during feedback on. (hence the name pseudo-static)
5.88
 Purely dynamic 2-phase FF
5.89
 Simple dynamic logic using 1- clock
in
out
n < delay < N
(RonCin)max < TW < n
N < TP < min(Trefresh)
 This doubled-sided

relation is too difficult to
satisfy for all signal cases.
TW
TP
5.90
 Dynamic logic using 2-phase clocking
in
out
1 > delay  RonCin
C/L
2 > preset
1
1 t12 2 t21
(i)
(ii) (iii)
TP
2
t12 > (C/L)delay,min0
t21 > 
for clock skew
Tp = delay+C/L delay
+ preset +t21
2  t12  (C / L) delay.max
5.91
Observations on Dynamic Logic vs. Static Logic
 Occupies less area : only one type of MOSFET is mainly used

comparable to NMOS or pseudo-NMOS
 Offers higher speed: less input capacitance, i.e. 0.5 and lower
switching threshold, i.e. device threshold rather than 1 Vdd
2
 Vulnerable to problems such as charge sharing, glitch, clock skew : as
it requires clock circuits.
 Difficult to operate at low frequency, making it unsuitable for
portable(low-power) systems unless some modifications are made
 Needs pre-charging anyway, causing unnecessary charging and
discharging regularly for quiet input signals
5.92
Ripple-through logic
 DC current path exists during clocked transistor on
P-logic
clk
c
Vo
clk
a
a
b
c
b
Vo
clk
clk=0 : prech.
clk=1 : eval
clk=0 : eval
clk=1 : prech.
clk
N-logic
poor
signal swing
5.93
 It is impossible to cascade ripple-through gates of the same kind.
When clk=0, ‘out’ node cannot be fully precharged.
 Node X is high, connecting ‘out’ to GND thus forming VDD-GND DC
path
5.94
 Cascading different
kinds (ideal clock)
clk = 0 : pre-charge
clk = 1 : evaluate
5.95
 Cascading different kinds (with clock skew)

t1 - t2 (0-0 overlap);
Stage 2 in eval. mode is affected by prech. of stage1

t3 - t4 (1-1 overlap);
Stage 2 output can be determined be prech. of stage2,
regardless of, or less affected by the evaluated signal of stage1.
5.96
Clocked CMOS(C2MOS) Inverter
 Functionally equivalent to each other
 C2MOS, (a) requires less area than (b).
5.97
 What would happen if clocked transistors are connected at the
PWR/GND ends?
5.98
 Output degradation occurs due to charge sharing
clk
input
output
hi-Z state
charge sharing
5.99
C2MOS register (C2MOS Master Slave D-f/f)
 Insensitive to   
or 1  2 overlap
 C2MOS D-f/f consists of
two cascaded C2MOS
latches oppositely clocked
1
0
section
eval.
hold
section
hold
eval.
5.100
 C2MOS register with (  )  clocking is insensitive to overlap as
long as the rise and fall times of clock is sufficiently small.
 While signal propagates via alternative charging and discharging in
the cascaded logic stages, it is impossible in C2MOS D-f/f
5.101
 When the clock rise and fall times are very slow, race condition can
occur. ( due to the simultaneous turning on M2 and Mg.)
5.102
Pipelining Insights
Tmin : min. clock period
Tmin : tp,reg + tp,logic + tsetup,reg
This includes thold,reg
Tmin,pipe = tp,reg + max(tp,add , tp,abs ,
tp,log ) + tsetup,reg
5.103
Latches vs. Registers
 Latch is level-sensitive, i.e., output follows input during (clock) = 1,
while sampled input is stored during  = 1. (positive latch)
D
0
1
0
Q
clk()
D
( -ve D-latch )
1
Q
clk()
( +ve D-latch )
D
Q

D
Q



5.104
Positive edge-triggered D-f/f or D-Reg.
D
QM
-ve latch
(master)
+ve latch
Q
(slave)
clk
shaded region has no effect
D
(i) setup time
QM
(ii) hold time
Q
(iii) C-Q delay
(i)
(ii)
(iii)
5.105
 Setup and hold time violations (due to clock skew)
M1(register)
g1
clk
logic
CL
M2(register)
delay(skew)

hold time violation occurs when
g1+ CL< skew

: M2 samples one cycle too late!
setup time violation occurs when
g1+ CL> skew + p : Next M2 sample will not have been updated
5.106
 Double-sided constraint on skew and delay(g1+ CL)

difficult to satisfy
skew < g1+ CL < skew+ p
(tight-rope walk)
g1+ CL < p (when skew  0)
g1 : clk-to-Q delay of M1,
CL : logic delay
p : clock period,
skew : clock skew
X0
d1
X1 d2 X2
X2-d2 < X1 < X0+d1
l1
l2
d2
d1
< l2
< l1
5.107
 Timing constraints in latch-based system ((b) and (c))
(a)
clk
Reg.
A
clk
Latch
A
(b)
(c)
clk
Latch
Tq
A
Tq
Comb.Logic
Td
Tq
Comb.Logic
Td
Comb.Logic
Ts
Tda
Latch
B
Ts
Reg.
B
Ts
Latch
B
Comb.Logic
Tdb
Latch
C
5.108
 Timing constraints in latch-based pipelined system (cont’d)
Tc0
Tc1
Tp
In (c), min(clk=high) time is given by
Tc1 > Tda + Tga + Tsb
Tda : delay of first-part logic
Tga : clk-Q delay of latch A
Tsb : setup time of latch B
Similarly,
Tc0 > Tdb + Tgb + Tsa
Assuming 50% clock cycle and identical latches,
Tc = Tc1 + Tc0 > Tda + Tdb + 2(Tg + Ts)
5.109
 If we use dynamic latch in latch-based pipeline, we obtain the
following circuit;
Vin
A
Here, “race” occurs if     0

(i.e. finite overlaps)

Vc1
Vc2
B
During D11 , “race” occurs
between input(A) and Vc1(B).
(Vc1 is the legal affector of Vc2 )
D11
5.110
Rule iii)
 C2MOS latch-based (rather than dynamic latch-based) pipelined circuit
is race-free as long as all static logic functions between C2MOS latch
are noninverting
If F is inverting, NMOS in the 1st & 2nd stage can both conduct if     0
5.111
Rule iii)
 This circuit is insensitive to clock overlap.
For (1-1) overlap i.e., when  =  = 1, the only way race can occur is
when logic function F is inverting as the following;
5.112
 NORA(NO-RAce) CMOS and rules to prevent RACE
i) consists of i) C2MOS pipeline registers, and ii) NP-CMOS dynamic
functional blocks modules.
ii) Each block module is a combinational logic followed by a C2MOS
latch.
(Generalization) Each combinational logic is a cascade of an appropriate
number(being limited by clock period) of either static or dynamic logic stages.
iii) Number of static inversions between C2MOS latches should be
even
iv) Same(opposite) type of two dynamic logic stages need an odd
(even) number of static inversions between them.
v) Number of static inversions between the last dynamic and the
C2MOS latch must be even for race-free.
5.113
Rule ii)
 Examples of NORA blocks(modules)
 -block consisting of two dynamic stage(N and P) followed by C2MOS
latch;
5.114
Rule v) O.K. case
 Examples (Cont’d)
 -block consisting of a dynamic stage followed by two(even) static
inversions until the C2MOS latch
 block
=0
=1
Logic
Precharge
Evaluate
Latch
Hold
Evaluate
 block
Logic
Evaluate
Precharge
Latch
Evaluate
Hold
5.115
Rule v) violation case
 Race occurs in a circuit violating condition(v)
5.116
 Domino logic(Special part of NORA)
 only noninverting outputs are available
 only low-to-high transition during evaluation (therefore, no glitches)
5.117
 Charge compensation scheme in Domino logic(I)
5.118
 Charge compensation scheme(II)
5.119
 Internal node precharge to prevent charge sharing
5.120
 MODL(Multiple Output Domino Logic)
 Useful when the target logical expression F can be expressed as
F = f1 f2      fn
where fi’s are subfunctions also required as output
5.121
 MODL carry lookahead adder
ci+1 = gi + pi ci
c1 = g0 + p0c0 (c0 = cin)
c2 = g1 + p1(g0+p0c0)

5.122
 MODEL CLA Carry generation circuit
5.123
 Latched Domino Logic (Ldomino)
 offers dual rail output, i.e., NOT operations are possible.
 can only be used as input stages to a chain of standard domino gates, as
it cannot be driven by a standard domino gate.(WHY?)
5.124
 Clocked Ldomino
 separates coupled latch from logic
5.125
 NORA with Domino logic
5.126
 NORA with Domino logic
5.127
 Pipeline(Synchronous) operation (inter-block) +
Domino-mode (Asynchronous) operation (intra-block)
5.128
 True Single-Phase Clocked logic(TSPC logic)
i) Doubled n-C2MOS latch & doubled p-C2MOS latch
 6 transistors / latch unlike
 no race with clock skew (as C2MOS latch)
 no limit on the number of inversions whatsoever (unlike C2MOS
latch in NORA)
5.129
 Logic can be inserted between latches, or in the n-C2MOS/p-C2MOS
as below;
5.130
ii) Simplified TSPC (called split-output)
 5 Tr’s / latch
 Voltage at node A(A’) swings up to VDD-VTn (down to VTp)
5.131
iii) D-flipflops using TSPC
5.132
iv) Some remarks on TSPC;
 Successfully used for very high-speed CMOS circuits
such as Alpha chip (200MHz in 0.75m technology)
 Clock slope should be sufficiently steep to prevent both
PMOS and NMOS on, resulting in undefined outputs and
race conditions.
 TSPC is vulnerable to noise, as TSPC register is
dynamic and the impedance of the storage node is high.
 Often a feedback transistor is added to make it possible
5.133
 Schmitt Trigger
 converts slowly-changing input to fast-transition output
 difference in switching threshold according to the direction of input
change

V(hysteresis voltage) = | VM+  VM  |
 noise suppression
5.134
 Emitter-coupled Schmitt Trigger
VM  
RE
VCC  VBE ( on)
RE  ( R1 // R2)
VM  
RE
 VCC  VBE ( on)
RE  R1
5.135
 CMOS Schmitt Trigger
Switching threshold of CMOS inverter depends on W/L-ratio between
PMOS and NMOS transistors
5.136
 Alternate CMOS Schmitt Trigger
5.137
 Monostable Sequential Circuit (One-Shot)
 used in ATD(address transition detection) circuit
(in)
(out)
delay(td)
td
 short pulse generation
N1
In

X

N2
Out
In
X
Out
5.138
 Astable Circuits
 Ring counter with adjustable frequency(VCO)
 M5&M6 form current mirror generating Iref and can be shared among
inverters.
 Shmitt trigger is for when Iref is small.
5.139
길 잃은 사공
세계 시장과 기술의 흐름을 보지 않고
공학을 가르치고 배우는 사람은
바람과 물결의 방향도 모르고
나침반도 없이 생각날 때 마다
노를 젓는 한심한 뱃사공이다.
5.140
도전의 높이
도전한 만큼만 성공할 수 있다.
오르려고 계획한 산의 높이보다
더 높이 올라서 있을 수는 없다.
5.141
技流파악
당신이 기술의 흐름을 아는데
필요한 3요소 :
1) 흐름을 보고 싶어하는 열망,
2) 전문가로서의 당신의 수준,
기초가 든든한지?
3) 세계의 top들이 모여드는 모임에
1년에 2번 이상 참석
5.142