of 78
8/10/2019 Ch7_Sequential1
1/78
Chapter 7
Sequential Circuits
Boonchuay Supmonchai
Integrated Design Application Research (IDAR) Laboratory
August 20, 2004; Revised - July 4, 2005
8/10/2019 Ch7_Sequential1
2/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 2
Goals of This Chapter
Implementation techniques for Register: latches and flipflops
Schmitt Triggers
Oscillator, pulse generators
Static versus Dynamic Realization
Clocking Strategies
8/10/2019 Ch7_Sequential1
3/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 3
Storage Mechanisms
Positive Feedback Charge-Based
COMBINATIONAL
LOGIC
Inputs Outputs
Next stateCurrent State
Q D
State
Register
CLOCK
Sequential Logic
STATIC DYNAMIC
8/10/2019 Ch7_Sequential1
4/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 4
Static vs Dynamic Storage
Static storage preserve state as long as the power is on
have positive feedback (regeneration) with an internal
connection between the output and the input
useful when updates are infrequent (clock gating)
Dynamic storage
store state on parasitic capacitors
only hold state for short periods of time (milliseconds)
require periodic refresh
usually simpler, so higher speed and lower power
8/10/2019 Ch7_Sequential1
5/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 5
Latches versus F lipflops
Latches (with Clock) level sensiti vecircuit that passes inputs to Q when the clock is
high (or low) - transparent mode
input sampled on the falling edge of the clock is held stable
when clock is low (or high) - hold mode
Flipflops (edge-triggered)
edge sensiti vecircuits that sample the inputs on a clock
transition
positive edge-triggered: 01
negative edge-triggered: 10
built using latches (e.g., master-slave flipflops)
8/10/2019 Ch7_Sequential1
6/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 6
Vi2 Vo2Vo1Vi1
Cascaded Inverters
A
Vi1 = Vo2
B
C
Review: The Regenerative Property
Small deviation frombias point C(e.g., fromnoise) is amplified andregenerated around the
circuit loop until eitherpoint Aor Bis reached
If the gain in thetransient region is larger
than 1, only Aand Barestable operation points.Cis a metastableoperation point.
8/10/2019 Ch7_Sequential1
7/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 7
Review: Bistable Circuits
The cross-coupling of twoinverters results in abistable circui t(a circuitwith two stable states)
Have to be able to changethe stored value by making A
(or B) temporarily unstable by increasing the loop gain
to a value larger than 1
done by applying a trigger pulse at Vi1or Vi2
the width of the trigger pulse need be only a little larger than
the total propagation delay around the loop circuit (twice the
delay of an inverter)
Vi1
Vi2
8/10/2019 Ch7_Sequential1
8/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 8
Review: SR Latch
S R Q !Q Action
0 0 Q !Q memory
1 0 1 0 set
0 1 0 1 reset
1 1 0 0 disal lowed
S
RQ
!Q
8/10/2019 Ch7_Sequential1
9/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 9
Review: Clocked D Latch
clock
QD
clock
D
Q
!Q
clock
t ransparentmode
holdmode
In our courseAll latches mean
clocked latches
8/10/2019 Ch7_Sequential1
10/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 10
D
Clk
Q D
Clk
Q
Flipflopstores data when
clock r ises (fal ls)
Clk
D
Q
Clk
D
Q
Latches versus F lipf lops I I
Latchstores data when
clock is low (h igh)
8/10/2019 Ch7_Sequential1
11/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 11
Positive and Negative Latches
In
Out
Clk
Out
StableOut
Follow In
Out
Stable
Out
Follow In
D Q
G
I n Out
Clk
Positive Latch
In
Out
Clk
Out
Stable
Out
Follow In
Out
StableOut
Follow In
D Q
G
I n Out
Clk
Negative Latch
8/10/2019 Ch7_Sequential1
12/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 12
N latch is transparentwhen = 0
P latch is transparentwhen = 1
Latch-Based Design
N
Latch
P
LatchLogic
Logic
8/10/2019 Ch7_Sequential1
13/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 13
clock
Out outputstable
output
stable
time
clock
In datastable
time
time
tsu thold
tc-q
Timing Metrics
8/10/2019 Ch7_Sequential1
14/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 14
Timing Defini tions
Setup time, tsetupis the time that the data inputs(D) must be valid before the clock transition
0 to 1 transition for a positive edge-triggered device
1 to 0 transition for a negative edge-triggered device
Hold time, tholdis the time that the data inputsmust remain valid after the clock edge
Propagation Delay, tc-qis the worst casepropagation delay (with reference to the clockedge)
time to copy D to Q
8/10/2019 Ch7_Sequential1
15/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 15
T tc-q+ tplogic+ tsutcdreg+ tcdlogicthold
T (clock period)
System Timing Constraints
COMBINATIONAL
LOGIC
Inputs Outputs
Next stateCurrent State
Q D
State
Register
CLOCK
tcd: contamination delay = minimum delay
8/10/2019 Ch7_Sequential1
16/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 16
Notes on System Timing Constraints
It is important to minimize the values of the timingparameters associated with the register.
In modern high-performance systems, the register
propagation delay and set-up times account for a
signi f icant portionof the clock period.
DEC Alpha EV6 has a maximum logic depth of 12 gates and
the register overhead accounts for about 15% of the clock
period.
Hold time becomes an issue when there is little logic
between registers or when the clocks at different
registers are somewhat out of phase due to clock skew.
8/10/2019 Ch7_Sequential1
17/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 17
Bui lding A (Static) Latch
CLK
CLK
CLK
D
Q
Cutt ing the feedback loop
(Mux -based latch)
Overpowering the feedback loo p
(as in Static RAM)
For a latch, use the clock as a decoupling signal, thatdistinguishes between the transparent and opaque states
D
CLK
CLK
D
can implement as NMOS-only
8/10/2019 Ch7_Sequential1
18/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 18
Q = !clk & Q | clk & DQ = clk & Q | !c lk & D
Negative Latch
Q
D
clk
0
1
feedback
t ransparentwhen the
clock is low
Change the stored value by cutting the feedback loop
MUX Based Latches
Posit ive Latch
Q
D
clk
1
0
feedback
t ransparentwhen the
clock is high
8/10/2019 Ch7_Sequential1
19/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 19
!clk
clk
input sampled
(t ransparentmode)
feedback
(ho ldmode)
TG MUX Based Latch Implementation
Q
D
clk
clk
!clk
Posi t ive Latch
c lkload is twotransistors (and twofor !c lk) = clock load of 4
Having to generate both c lkand !clk
(nonoverlapping clocks)
8/10/2019 Ch7_Sequential1
20/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 20
QD
clk !Q
!clk
Reduced clock load, bu tthreshold drop at output of
pass trans is tors so reduced
noise margins and per formance
PT MUX Based Latch Implementation
!clk
clk
input sampled
(t ransparentmode)
feedback
(ho ldmode)
8/10/2019 Ch7_Sequential1
21/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 21
clk
T tc -q+ tplog ic+ tsu
Thigh tc-q+ tcd log ic
clk
B BB
Which value of B is stored?
Latch Race Problem
Two-sidedclock constraint
8/10/2019 Ch7_Sequential1
22/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 22
clk
QM
Q
D
clk
QD
clk = 0 transparent hold
clk = 1 hold transparent
0
1 Q1
0
D
clk
Q
clk
SlaveMaster
QM
Master Slave Based ET F lipf lop
8/10/2019 Ch7_Sequential1
23/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 23
T1
T2 Q
D
clk
QM
I1
I2 I3
I4
I5 I6
T3
T4
MasterSlave
!clk
clk
master t ransparentslave hold
master holdslave t ransparent
20 Transistors*
8 clock loads* Ignore clk b uffer
MS ET Implementation
8/10/2019 Ch7_Sequential1
24/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 24
Assume propagation delays are tpd_invand tpd_tx, that thecontamination delay is 0, and that the inverter delay to
derive !clkis 0
Set-up time- time before rising edge of clkthat Dmust
be valid
Propagation delay- time for QMto reach Q
Hold time- time Dmust be stable after rising edge of clk
MS ET Timing Properties
tsu= 3 * tpd_inv+ tpd_tx
tpd= tpd_inv+ tpd_tx
thold= 0
8/10/2019 Ch7_Sequential1
25/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 25
Notes on MS ET Timing Properties
Set-up time
How long before the rising edge does D have to be stable suchthat QMsamples the value reliably?
D has to propagate through I1, T1, I3 and I2 before the risingedge to ensure that the node voltages on both terminals of T2
are the same value.
Propagation delay time
Since the delay of I2 is included in the set-up time, the outputof I4 is valid beforethe rising edge of clk, so the delay is
simply the delay through T3 and I6
Hold time
since T1 turns off when the clock goes high, any changes in Dafter clk goes high are not seen, so hold time is 0
8/10/2019 Ch7_Sequential1
26/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 26
Set-up Time Simulation
D clk
QM
I2out
tsetup= 0.21 ns
works correct ly
Vo
lts
Time (ns )
-0.5
0
0.5
1
1.5
2
2.5
3
0 0.2 0.4 0.6 0.8 1
Q
8/10/2019 Ch7_Sequential1
27/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 27
Set-up Time Simulation I I
-0.5
0
0.5
1
1.5
2
2.5
3
0 0.2 0.4 0.6 0.8 1
Vo
lts
Time (ns )
D clk
QM
I2out
tsetup= 0.20 ns
Q
the clock is enabled before the nodes on both sides
of the transmission gate T2 settle to the same valueFails!
8/10/2019 Ch7_Sequential1
28/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 28
Propagation Delay Simulation
tc-q (LH)= 160 psec tc-q (HL)= 180 psec
-0.5
0
0.5
1
1.5
2
2.5
3
0 0.5 1 1.5 2 2.5
Volts
Time (ns )
tc-q(LH) tc -q(HL)
prop agation delay is measured from the 50% point
of the clk edge to the 50% point of th e Q output
DClk Q
8/10/2019 Ch7_Sequential1
29/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 29
Reduced Load MS ET FF
!clkclk
QD
!clk clk
I1
I2I4
I3
QM T2
T1
Clock load per register is important since it directly
impacts the power dissipation of the clock network.
Can reduce the clock load (at the cost of robustness) by
making the circuit ratioed
to switch the state of the master, T1must be sized to overpowerI2 to avoid reverse conduction, I4must be weakerthan I1
reverse conduct ion
12 Transistors4 clock loads
8/10/2019 Ch7_Sequential1
30/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 30
Non-I deal Clocks
!clk
clk
Ideal clocks
!clk
clk
Non-Ideal clocks
1-1 Overlap 0-0 Overlap
Clkand !clkare never perfect inversions of one another
We must generate !clkand route both signals
Variations can exist in the wires used to route the two clocksignals and load capacitances may vary
Non-ideal clocks create skewresulting in clock overlap
8/10/2019 Ch7_Sequential1
31/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 31
!QD
clkX
!clk
!clk Q
clk
B
AP1
P2
P3
P4
I1 I2I3 I4
Race conditiondirect path from D to Q during the short time
when both clk and !clk are high (1-1 ov erlap)
Undefined stateboth B and D are driving A when clk and !clk
are both high
Dynamic storagewhen clk and !clk are both low (0-0 ov erlap)
Race
Example of Clock Skew Problems
8/10/2019 Ch7_Sequential1
32/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 32
clk1
clk2
mastert ransparent
slaveho ld
master hold
slave t ransparent
dynamic
storage
tnon_overlap
Pseudostatic Two-Phase ET FF
!QD
clk1X
clk2
clk2 Q
clk1
B
AP1
P2
P3
P4
I1 I2I3 I4
8/10/2019 Ch7_Sequential1
33/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 33
Two Phase Clock Generator
clk
clk1
clk2
A
B
clk
A
B
clk1
clk2
8/10/2019 Ch7_Sequential1
34/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 34
!clk
clk
Power PC F lipf lop
mastert ransparent
slavehold masterho ld
slavet ransparent
1
1 0
1 1
D Q
clk
!clk
!clk
clk
0 0 0
1
16 Transistors
8 clock loads
8/10/2019 Ch7_Sequential1
35/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 35
S
Q
Q
Cross-coupled NANDs
This is no t used in datapaths any m ore,
but is a basic bui ld ing block for m emory cel l
Overpower ing The Feedback Loop
Clocked SR Latch
SR
clkclk
!Q
Q
M1
M2
M3
M4
M5
M6
M7
M8
8/10/2019 Ch7_Sequential1
36/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 36
10
onoff
off ->onoff ->on
0
1
on
on
off
off
1
0
on
off
off
onS
R
clkclk
!Q
Q
M1
M2
M3
M4
M5
M6
M7
M8 0101
Ratioed CMOS Clocked SR Latch
8 Transistors
2 Clock loads** sized
No static power consumption, but a ratioeddevicewhere sizing is critical to ensure proper functionality
M7, M8 must overcome M4 to bring Q low, so must M5, M6over M2
8/10/2019 Ch7_Sequential1
37/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 37
Sizing I ssues
0
0.5
1
1.5
2
2 2.5 3 3.5 4
(W/L)5 and 6
!Q(
Vo
lts)
(W/L)2 and 4 = 1.5m/0.25 m(W/L)1 and 3 = 0.5m/0.25 m
so (W/L )5 and 6> 3
Output vo l tage depends on pul l -down trans is tor width
8/10/2019 Ch7_Sequential1
38/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 38
Transient Response
S
0
1
2
3
0 0.4 0.8 1.2 1.6 2
!Q(
Volts)
Time (ns)
W=1 mW=0.9 m
W=0.8 m
!Q
W=0.5 m
W=0.6 mW=0.7 m
Individual device ratio for M5 or M6 must be largerthan approx. 6.
Analysis results give 2.26 (instead of 3) since it doesnt take into account
channel length modulat ionand DIBL (drain ind uced b arr ier loading).
B S h i
8/10/2019 Ch7_Sequential1
39/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 39
clkclk
SR
M1
M2
M3
M4
M5
M6 S
R
clk
!Q
Q
clk
6 Transistor CMOS SR Latch
6 Transistors
2 Clock loads
Problems with noisemarginsand staticpower consumptiondue to threshold dropacross pass transistors
Once again, sizing isimportant - especiallyM5 and M6
B S h i
8/10/2019 Ch7_Sequential1
40/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 40
Review: Storage Mechanisms
D
CLK
CLK
Q
Dynamic(charge-based)
CLK
CLK
CLK
D
Q
Static(Positive Feedback)
Useful when update is infrequent Simpler, Faster, and Lower Power
B S h i
8/10/2019 Ch7_Sequential1
41/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 41
!clk clk
T1 T2I1 I2 Q
QMD
C
1
C
2clk !clk
!clk
clk
mastert ransparent
slavehold
masterhold
slavet ransparent
master slave
tsu=
thold=
tc -q
=
tpd_txzero2 t
pd_inv+ t
pd_tx
Dynamic ET F lipf lop
8 Transistors
4 Clock loads
B S h i
8/10/2019 Ch7_Sequential1
42/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 42
0-0 overlaprace condition
toverlap0-0< tT1+ tI1+ tT2
1-1 overlaprace condition
toverlap1-1 < thold
Dynamic ET FF Race Conditions
!clk clk
T1 T2I1 I2 Q
QM
D
C
1
C
2clk !clk
!clk
clk
B Supmonchai
8/10/2019 Ch7_Sequential1
43/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 43
clk2
clk1 tnon_overlap
master t ransparent
slave hold
master hold
slave t ransparent
Dynamic Two-Phase ET FF
clk1 clk2
T1 T2I1 I2 Q
QM
D
C
1
C
2!clk1 !clk2
B Supmonchai
8/10/2019 Ch7_Sequential1
44/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 44
Pseudostatic Dynamic Latch
Robustness considerations limit the use of dynamic FFs
Coupling between signal nets and internal storage nodes caninject significant noise and destroy the FF state
Leakage currents cause state to leak away with time
Internal dynamic nodes dont track fluctuations in VDD that
reduces noise margins
A simple fix is to make the circuit pseudostatic
clk
T1D
!clk
Slight in crease in delay
(adds to the capacit iveload) and power
consump t ion, but i t
improv es noise immuni ty
signi f icant ly
B Supmonchai
8/10/2019 Ch7_Sequential1
45/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 45
clk
!clk
!clk
clk
QM
C1 C2
QD
M1
M3
M4
M2 M6
M8
M7
M5
Master Slave
master t ransparentslave hold
master hold
slave t ransparent
on
on
off
off
on
onoff
off
C2MOS (Clocked CMOS) ET F lipf lop
!clk
clk
8 Transistors
4 Clock loads
Insens i t ive to cloc k
over lap as long as the
r ise and fall t imes of
the clock edges are
suff ic ient ly smal l
B Supmonchai
8/10/2019 Ch7_Sequential1
46/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 46
C2MOS FF 0-0 Over lap Case
0 0QM
C1 C2
QD
M1
M3
M4
M2 M6
M8
M7
M5
!clk
clk
!clk
clk
B Supmonchai
8/10/2019 Ch7_Sequential1
47/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 47
Notes on C2MOS FF 0-0 Overlap Case
Does any new data sampled during the overlap window
propagate to Q (race)?
New data is sampled on QM, but cannotpropagate to Q sinceM7 is off (slave is in hold).
Any new data sampled on the falling clock edge is not seen at Q
For clocking on the left: at the end of the overlap period!clk= 1 and both M7 and M8 turn off, putting the slavein the hold mode
For the clocking on the r ight: at the end of the overlapperiod clk= 1 and both M3 and M4 turn off, putting themaster in the hold mode (affects setup time as well)
The result: the FF is slower(slower tc-qtime)
B Supmonchai
8/10/2019 Ch7_Sequential1
48/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 48
!clk
clk
!clk
clk
1-1 overlapconstraint: toverlap1-1 < thold
11
QM
C1 C2
QD
M1
M3
M4
M2 M6
M8
M7
M5
C2MOS FF 1-1 Over lap Case
B Supmonchai
8/10/2019 Ch7_Sequential1
49/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 49
Notes on C2MOS FF 1-1 Overlap Case
New data is sampled on QM, but cannot propagate to Q
since M8 is off (slave is in hold).
A bit more problematic than 0-0 overlap.
It must enforce a hold timeon D, so that changing D whichreaches QM is not copied to Q when overlap time is over -
f irst clocking condition.
By imposing a hold time on D - that D must be stable duringclock overlap - overcome this problem as well
However, possible race can occur if the rise/fall times ofthe clock are sufficiently slow.
Works correctly as long as the clock rise/fall times is smallerthan approximately five timesthe propagation delay of theflipflop.
B.Supmonchai
8/10/2019 Ch7_Sequential1
50/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 50
C2MOS Transient Response
Q(3)
Q(0.1)
-0.5
0
0.5
1
1.5
2
2.5
3
0 2 4 6 8
Time (nsec)
clk(0.1 ns)
QM(3)
clk(3 ns)
For slow clocks, potent ial for a race condi t ion exists
B.Supmonchai
8/10/2019 Ch7_Sequential1
51/78
B.Supmonchai
2102-545 Digital ICs Sequential Logic 51
clk clkIn
Q
Positive LatchNegative Latch
t ransparentwhen clk = 1
ho ldwhen clk = 0
clk clkIn Q
ho ldwhen clk = 1
t ransparentwhen clk = 0
True Single Phase Clocked (TSPC) Latches
Uses only a single clock
No clock over lap (skew)to worry about ; reduced clock load
B.Supmonchai
8/10/2019 Ch7_Sequential1
52/78
p
2102-545 Digital ICs Sequential Logic 52
clk clkIn
Q
PUN
PDN
clk clk
A
Q
B
BA
Embedding Logic in TSPC Latch
Logic can be embedded into latch (or FF)
Reduce delay overhead associated wi th the latch
A AND B
B.Supmonchai
8/10/2019 Ch7_Sequential1
53/78
p
2102-545 Digital ICs Sequential Logic 53
Notes on Embedding Logic in TSPC Latch
Set-up time increased, but overall performanceimproved
The increase in the set-up time is typically smallerthan the
delay of an AND gate.
For example, using minimum size devices set-up of ANDlatch is 140 psec.
Using the conventional approach of AND gate followed by
latch has an effective set-up time of 600 psec.
Technique used extensively in the design of the EV4DEC Alpha microprocessor and many other high
performance processors.
B.Supmonchai
8/10/2019 Ch7_Sequential1
54/78
p
2102-545 Digital ICs Sequential Logic 54
master hold
slave t ransparent
master t ransparentslave hold
ononoffoff
on on
off
off
clk clkD
Master Slave
clk clkQQM
clk
12 Transistors
4 Clock loads
TSPC ET FF
Vir tual ly al l constraints remo ved - no cloc ks to over lap, no race
B.Supmonchai
8/10/2019 Ch7_Sequential1
55/78
p
2102-545 Digital ICs Sequential Logic 55
Notes on TSPC ET FF
Warning!- similar to C2
MOS, TSPC flipflopsmalfunction when the slope of the clock is not
suff iciently steep.
Slow clock cause both the NMOS and PMOS
clocked transistors to be ON simultaneously,
resulting in undefined values of the states and race
conditions.
Clock slopes thus must be carefully engineered. Ifnecessary, local buffers must be introduced to ensure
the quality of the clock signal
B.Supmonchai
8/10/2019 Ch7_Sequential1
56/78
2102-545 Digital ICs Sequential Logic 56
clkD clkQ
clk
clk
X
Y
M1
M2
M3 M6
M5
M4 M7
M8
M9
I1 I2 I3
Simpli f ied TSPC ET FF
I1 hold
I2 evaluate
I3 samp le (transp arent)
I1 sample (transparent)
I2 precharged
I3 hold
onoff
on
off
D
D
clk
on
on
off
off
1
!D
9 Transistors*
4 Clock loads
*(11 if Q is needed)
B.Supmonchai
8/10/2019 Ch7_Sequential1
57/78
2102-545 Digital ICs Sequential Logic 57
Notes on TSPC ET FF
On the positive edge of the clock, note that the node Xtransitions to a low if D is high. Therefore, the input
must be kept stable until the value on node X before the
rising edge of the clock propagates to Y
Hold time of the register (less than 1 inverter delay since ittakes 1 inverter delay for the input to affect node X).
Propagation delay is essentially three inverters since
the value on node X must propagate to output Q
Set-up time is the time for node X to be validone
inverter delay
B.Supmonchai
8/10/2019 Ch7_Sequential1
58/78
2102-545 Digital ICs Sequential Logic 58
Sizing Issues in Simplif ied TSPC ET FF
0
1
2
3
0 0.2 0.4 0.6 0.8 1
Time (nsec)
clk
!Qorig
Qorig
!Qmod
Qmod
Transistor sizing
Original width
M4, M5= 0.5m
M7, M8= 2m
Modified width
M4, M5= 1m
M7, M8= 1m
Sizing is cr i t ical w ith improper s izing gl i tches may occu r due
to race condi t ion w hen the clock trans i t ions from low to h igh
B.Supmonchai
8/10/2019 Ch7_Sequential1
59/78
2102-545 Digital ICs Sequential Logic 59
Positive Latch Negative Latch
t ransparentwhen clk = 1holdwhen clk = 0
holdwhen clk = 1t ransparentwhen clk = 0
clk
In
Q
A
clkIn Q
A
When In= 0, A= VDD- VTn When In= 1, A= | VTp |
Spli t-Output TSPC Latches
B.Supmonchai
8/10/2019 Ch7_Sequential1
60/78
2102-545 Digital ICs Sequential Logic 60
Spli t-Output TSPC ET FF
8 Transistors*
2 Clock loads*(10 if Q is needed)
Which edge-triggered?
Downside is not all node voltages in the latch experiencefull logic swing due to threshold drop.
E.g., for positive latch when D=0 and clk=1, A=Vdd-Vth (Also
limits the amount of Vddscaling possible with this latch).
clkDQ
clkQM
A
B.Supmonchai
8/10/2019 Ch7_Sequential1
61/78
2102-545 Digital ICs Sequential Logic 61
Master-Slave Flipflop Pulse-Triggered Flipf lop
D
Clk
Q D
Clk
Q
Clk
Data
L1 L2
Pulse-Tr iggered F lipf lops
Another approach to design an edge-triggeredflipflop is to use pulse-triggered.
LData
D
Clk
Q
Clk
B.Supmonchai
8/10/2019 Ch7_Sequential1
62/78
2102-545 Digital ICs Sequential Logic 62
0 ONVdd
OFF OFF
11 0 ON
Xclk
D
Q
M1
M2
M3
M4
M5
M6
P1
P2
P3
!clkd
1/0ON/
OFF
0/Vdd ON/OFF
1/0
0OFF
1
1
OFF
ONON
ON
Pulsed FF (AMD-K6)
Pulse registers - a short pulse (glitch clock) is generatedlocally from the rising (or falling) edge of the system
clock and is used as the clock input to the flipflop
B.Supmonchai
8/10/2019 Ch7_Sequential1
63/78
2102-545 Digital ICs Sequential Logic 63
Notes on Pulsed FF
Race conditions are avoided by keeping the transparentmode time very short(during the pulse only)
Reduce clock load but substantially increase complexityin verification
The transparency period determines the hold time. The window must be wide enoughfor the input data to
propagate to Q.
The set-up time can be NEGATIVE(if the transparency
window is longer than the delay from input to output). This is attractive, as data can arrive at the register even after
the clock goes high, meaning that time can be borrowed fromthe previous cycle.
B.Supmonchai
8/10/2019 Ch7_Sequential1
64/78
2102-545 Digital ICs Sequential Logic 64
0
0
1
1
1
1
1
0
1
0
1
Sense Amp FF (StrongArm SA100)
Sense amplif ieris a circuit that accept small swing inputsignals and amplify them to full rail-to-rail signals
clk
D
Q
!Q
M1
M2
M3
M5
M6
M4
M9
M7
M8
M10
!S
!R
X
Y
B.Supmonchai
8/10/2019 Ch7_Sequential1
65/78
2102-545 Digital ICs Sequential Logic 65
Notes on Sensed Amp FF
The key is transistor M4(in the middle of Sensed amp);
it delays signals that pass through to the other side of itsterminal, making the change on the other side slower
When D = 1, Y changes after X due to the delay of M4. By thetime M6 reacts to the change at its terminal, it is already
turned off by the terminal voltage at M4 (a 0). Thus, M6holds a 1.
M4 also provides DC-leakage path to ground for eithernode X or Y in case that the inputs change their valueafter the positive edge of CLKarrives.
Advantages are reduced clock loadand that it can beused as a receiver for reduced swing differential buses
Where does the differential signal enter?
B.Supmonchai
8/10/2019 Ch7_Sequential1
66/78
2102-545 Digital ICs Sequential Logic 66
F lipf lop Compar ison Chart
Name Type #clk ld #tr tset-up thold tpFF
Mux Static 8(clk-!clk) 20 3tpinv+tp tx 0 tpinv+tp tx
PowerPC Static 8 (clk-!clk) 16
2-phase Ps-Static 8(clk1-clk2) 16
T-gate Dynamic 4(clk-!clk) 8 tp tx to1-1 2tpinv+tp tx
C2MOS Dynamic 4(clk-!clk) 8
TSPC Dynamic 4(clk) 11 tp inv tp inv 3tp inv
S-O TSPC Dynamic 2(clk) 10
AMD K6 Dynamic 5(clk) 19
SA 100 SenseAmp 3(clk) 20
B.Supmonchai
8/10/2019 Ch7_Sequential1
67/78
2102-545 Digital ICs Sequential Logic 67
Choosing a Clocking Strategy
Choosing the right clocking scheme affects the
functionality, speed, and powerof a circuit
Two-phase designs
+ robust and conceptually simple
- need to generate and route two clock signals
- have to design to accommodate possible skew between thetwo clock signals
Single phase designs
+ only need to generate and route one clock signal + supported by most automated design methodologies
+ dont have to worry about skew between the two clocks
- have to have guaranteed slopes on the clock edges
B.Supmonchai
8/10/2019 Ch7_Sequential1
68/78
2102-545 Digital ICs Sequential Logic 68
Non-Bistable Sequential Circuits
Previously, we have defined a circuit having twostable states a bi-stable circuit
Other regenerative circuits, which are non-bistable:
Monostable
Only one stable state -> Pulse generators, One-shot circuits
Astable
No stable states -> Oscillator, On-chip clock generator
Schmi tt Tr igger
A special regenerative circuit exhibiting hysteresisin VTC.
B.Supmonchai
8/10/2019 Ch7_Sequential1
69/78
2102-545 Digital ICs Sequential Logic 69
In Out
Schmitt Tr igger
Non-Bis table Sequential Circui ts
Vin
Vout
VOH
VOL
VM
VM+
2 important properties
Hysteresis
Fast Trans it ion Time
at the output
B.Supmonchai
8/10/2019 Ch7_Sequential1
70/78
2102-545 Digital ICs Sequential Logic 70
Noise Suppression using Schmitt Tr igger
VIN
t0 t
VM+
VM-
VOUT
tt0+ tp
Example: Switch Debouncer
B.Supmonchai
8/10/2019 Ch7_Sequential1
71/78
2102-545 Digital ICs Sequential Logic 71
CMOS Schmitt Tr igger
M1
M4M2
M3
VIN VOUTX
VDDMoves switchingthreshold of the
first inverter
Adapting the rat io between PMOS and NMOS, depending upo n the
direct ion of the trans i t ion resul ts in a shi f t in swi tch ing threshold
Low-to-High
ref f= kM1/(kM2+ kM4)
High-to-Low
ref f= (kM1+ kM3)/kM2
B.Supmonchai
8/10/2019 Ch7_Sequential1
72/78
2102-545 Digital ICs Sequential Logic 72
Schmitt Tr igger Simulated VTC
2.5
VM2
VM1
Vin
(V)
2.0
1.5
1.0
0.5
0.00.0 0.5 1.0 1.5 2.0 2.5
Vout(
V)
2.5
k= 2k= 3
k= 4
k= 1
Vin
(V)
2.0
1.5
1.0
0.5
0.00.0 0.5 1.0 1.5 2.0 2.5
Vout(
V)
Effect of varying the ratio of the
PMOS device M4
Voltage Transfer Characteristics
with hysteresis
M1 = 1 m/0.25 m, M2 = 3 m/0.25 m, M3 = 0.5 m/0.25 m
M4 = 1.5 m/0.25 m M4 = kx 0.5 m/0.25 m
B.Supmonchai
8/10/2019 Ch7_Sequential1
73/78
2102-545 Digital ICs Sequential Logic 73
CMOS Schmitt Tr igger (2)
How does the gate operate?
M2
VIN VOUT
X
M1
M5
M6
M3
M4
Sketch VTC and find expression for VM-and VM+
B.Supmonchai
O
8/10/2019 Ch7_Sequential1
74/78
2102-545 Digital ICs Sequential Logic 74
Review: Ring Oscil lator
0.0
0.0
0.5
1.0
1.5
2.0
2.5V1 V3 V5
3.0
20.50.5
time (ns)
1.0 1.5
Period: T= 2 x tpx N
tp
Different Clock Duty-Cyclesand ph ases can be der ived
using simp le logic operat ions
B.Supmonchai
V l C l l O il l (VCO)
8/10/2019 Ch7_Sequential1
75/78
2102-545 Digital ICs Sequential Logic 75
In
VDD
M3
M1
M2
M4
M5
VDD
M6
Vcontr
Current starved inverter
Iref
Iref
Schmitt Trigger
restores signal slopes
Voltage Control ler Osci l lator (VCO)
Oscillation frequency of a VCO is a function (typicallynonlinear) of a control voltage
Delay of a current starved inverter depends on th e current
l imit avai lable to d ischarge the load c apacitance of th e gate
B.Supmonchai
C t St d I t Si l ti
8/10/2019 Ch7_Sequential1
76/78
2102-545 Digital ICs Sequential Logic 76
Current-Starved I nverter Simulation
0.5 1.5 2.5Vcont r (V)
0.0
2
4
6
Vct r l(V)
tpHL
(nse
c)
The device is in the
subthreshold region
when Vctrlis smaller
than VT, resulting in
large variations of tp
as the drive current
is exponentially
dependent on the
drive voltage
Delay sensitive to
noise and variation
in Vctrl
B.Supmonchai
Diff ti l D l El t d VCO
8/10/2019 Ch7_Sequential1
77/78
2102-545 Digital ICs Sequential Logic 77
Differential Delay Element and VCO
two stage VCO
v1
v2
v3
v4
- InvertingInputs/Outputs+ Non-InvertingInputs/Outputs
Oscillator with even numberof stages can be implemented
in2
Vctrl
Vo2 Vo1
in1
delay cell
+
+
-
-
Differential-type VCO has better immuni ty to common mode
noise (e.g., supply noise) but consume more power
B.Supmonchai
2 St VCO Si l ti
8/10/2019 Ch7_Sequential1
78/78
2-Stage VCO Simulation
0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
2 0.5 1.5
V1
V2
V3
V4
time (ns)
2.5 3.5
The In-Phase and Quadrature Phase are prod uced simultaneously