Analysis and Design of A Low-Voltage Low-Power Double-Tail Comparator

Download as pdf or txt
Download as pdf or txt
You are on page 1of 10

IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, VOL. 22, NO.

2, FEBRUARY 2014 343


Analysis and Design of a Low-Voltage Low-Power
Double-Tail Comparator
Samaneh Babayan-Mashhadi, Student Member, IEEE, and Reza Lot, Member, IEEE
AbstractThe need for ultra low-power, area efcient, and
high speed analog-to-digital converters is pushing toward the
use of dynamic regenerative comparators to maximize speed and
power efciency. In this paper, an analysis on the delay of the
dynamic comparators will be presented and analytical expres-
sions are derived. From the analytical expressions, designers can
obtain an intuition about the main contributors to the comparator
delay and fully explore the tradeoffs in dynamic comparator
design. Based on the presented analysis, a new dynamic com-
parator is proposed, where the circuit of a conventional double-
tail comparator is modied for low-power and fast operation
even in small supply voltages. Without complicating the design
and by adding few transistors, the positive feedback during
the regeneration is strengthened, which results in remarkably
reduced delay time. Post-layout simulation results in a 0.18-m
CMOS technology conrm the analysis results. It is shown that in
the proposed dynamic comparator both the power consumption
and delay time are signicantly reduced. The maximum clock
frequency of the proposed comparator can be increased to 2.5
and 1.1 GHz at supply voltages of 1.2 and 0.6 V, while consuming
1.4 mW and 153 W, respectively. The standard deviation of the
input-referred offset is 7.8 mV at 1.2 V supply.
Index TermsDouble-tail comparator, dynamic clocked
comparator, high-speed analog-to-digital converters (ADCs),
low-power analog design.
I. INTRODUCTION
C
OMPARATOR is one of the fundamental building blocks
in most analog-to-digital converters (ADCs). Many high-
speed ADCs, such as ash ADCs, require high-speed, low-
power comparators with small chip area. High-speed compara-
tors in ultra deep submicrometer (UDSM) CMOS technologies
suffer from low supply voltages especially when considering
the fact that threshold voltages of the devices have not
been scaled at the same pace as the supply voltages of the
modern CMOS processes [1]. Hence, designing high-speed
comparators is more challenging when the supply voltage is
smaller. In other words, in a given technology, to achieve
high speed, larger transistors are required to compensate the
reduction of supply voltage, which also means that more
die area and power is needed. Besides, low-voltage opera-
tion results in limited common-mode input range, which is
important in many high-speed ADC architectures, such as ash
Manuscript received September 4, 2012; revised December 7, 2012;
accepted January 15, 2013. Date of publication February 11, 2013; date of
current version January 17, 2014.
The authors are with the Electrical Engineering Group of Engineering
Department, Ferdowsi University of Mashhad, Mashhad 91775-111, Iran
(e-mail: babayan@ieee.org; rlot@ieee.org).
Color versions of one or more of the gures in this paper are available
online at http://ieeexplore.ieee.org.
Digital Object Identier 10.1109/TVLSI.2013.2241799
ADCs. Many techniques, such as supply boosting methods [2],
[3], techniques employing body-driven transistors [4], [5],
current-mode design [6] and those using dual-oxide processes,
which can handle higher supply voltages have been developed
to meet the low-voltage design challenges. Boosting and
bootstrapping are two techniques based on augmenting the
supply, reference, or clock voltage to address input-range
and switching problems. These are effective techniques, but
they introduce reliability issues especially in UDSM CMOS
technologies. Body-driven technique adopted by Blalock [4],
removes the threshold voltage requirement such that body-
driven MOSFET operates as a depletion-type device. Based
on this approach, in [5], a 1-bit quantizer for sub-1V
modulators is proposed. Despite the advantages, the body-
driven transistor suffers from smaller transconductance (equal
to g
mb
of the transistor) compared to its gate-driven counter-
part while special fabrication process, such as deep n-well is
required to have both nMOS and pMOS transistors operate
in the body-driven conguration. Apart from technological
modications, developing new circuit structures which avoid
stacking too many transistors between the supply rails is
preferable for low-voltage operation, especially if they do not
increase the circuit complexity. In [7][9], additional circuitry
is added to the conventional dynamic comparator to enhance
the comparator speed in low supply voltages. The proposed
comparator of [7] works down to a supply voltage of 0.5 V
with a maximum clock frequency of 600 MHz and consumes
18 W. Despite the effectiveness of this approach, the effect
of component mismatch in the additional circuitry on the
performance of the comparator should be considered. The
structure of double-tail dynamic comparator rst proposed
in [10] is based on designing a separate input and cross-
coupled stage. This separation enables fast operation over a
wide common-mode and supply voltage range [10].
In this paper, a comprehensive analysis about the delay of
dynamic comparators has been presented for various architec-
tures. Furthermore, based on the double-tail structure proposed
in [10], a new dynamic comparator is presented, which does
not require boosted voltage or stacking of too many transistors.
Merely by adding a few minimum-size transistors to the
conventional double-tail dynamic comparator, latch delay time
is profoundly reduced. This modication also results in con-
siderable power savings when compared to the conventional
dynamic comparator and double-tail comparator.
The rest of this paper is organized as follows. Section II
investigates the operation of the conventional clocked
regenerative comparators and the pros and cons of each
1063-8210 2013 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
344 IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, VOL. 22, NO. 2, FEBRUARY 2014
Fig. 1. Schematic diagram of the conventional dynamic comparator.
structure is discussed. Delay analysis is also presented and
the analytical expressions for the delay of the comparators are
derived. The proposed comparator is presented in Section III.
Section IV discusses the design issues. Simulation results are
addressed in Section V, followed by conclusions in Section VI.
II. CLOCKED REGENERATIVE COMPARATORS
Clocked regenerative comparators have found wide appli-
cations in many high-speed ADCs since they can make
fast decisions due to the strong positive feedback in the
regenerative latch. Recently, many comprehensive analyses
have been presented, which investigate the performance of
these comparators from different aspects, such as noise [11],
offset [12], [13], and [14], random decision errors [15], and
kick-back noise [16]. In this section, a comprehensive delay
analysis is presented; the delay time of two common struc-
tures, i.e., conventional dynamic comparator and conventional
dynamic double-tail comparator are analyzed, based on which
the proposed comparator will be presented.
A. Conventional Dynamic Comparator
The schematic diagram of the conventional dynamic com-
parator widely used in A/D converters, with high input
impedance, rail-to-rail output swing, and no static power
consumption is shown in Fig. 1 [1], [17]. The operation of
the comparator is as follows. During the reset phase when
CLK = 0 and M
tail
is off, reset transistors (M
7
M
8
) pull
both output nodes Outn and Out p to V
DD
to dene a start
condition and to have a valid logical level during reset. In the
comparison phase, when CLK = V
DD
, transistors M
7
and M
8
are off, and M
tail
is on. Output voltages (Out p, Outn), which
had been pre-charged to V
DD
, start to discharge with different
discharging rates depending on the corresponding input volt-
age (INN/INP). Assuming the case where V
INP
> V
INN
, Out p
discharges faster than Outn, hence when Out p (discharged
by transistor M
2
drain current), falls down to V
DD
|V
thp
|
before Outn (discharged by transistor M
1
drain current), the
corresponding pMOS transistor (M
5
) will turn on initiating the
latch regeneration caused by back-to-back inverters (M
3
, M
5
Fig. 2. Transient simulations of the conventional dynamic comparator for
input voltage difference of V
in
= 5 mV, V
cm
= 0.7 V, and V
DD
= 0.8 V.
and M
4
, M
6
). Thus, Outn pulls to V
DD
and Out p discharges
to ground. If V
INP
< V
INN
, the circuits works vice versa.
As shown in Fig. 2, the delay of this comparator is com-
prised of two time delays, t
0
and t
latch
. The delay t
0
represents
the capacitive discharge of the load capacitance C
L
until
the rst p-channel transistor (M
5
/M
6
) turns on. In case, the
voltage at node INP is bigger than INN (i.e., V
INP
> V
INN
),
the drain current of transistor M
2
(I
2
) causes faster discharge
of Out p node compared to the Outn node, which is driven by
M
1
with smaller current. Consequently, the discharge delay
(t
0
) is given by
t
0
=
C
L

V
thp

I
2

= 2
C
L

V
thp

I
tail
. (1)
In (1), since I
2
= I
tail
/2 + I
in
= I
tail
/2 + g
m1,2
V
in
, for
small differential input (V
in
), I
2
can be approximated to be
constant and equal to the half of the tail current.
The second term, t
latch
, is the latching delay of two cross-
coupled inverters. It is assumed that a voltage swing of
V
out
= V
DD
/2 has to be obtained from an initial output
voltage difference V
0
at the falling output (e.g., Out p). Half
of the supply voltage is considered to be the threshold voltage
of the comparator following inverter or SR latch [17]. Hence,
the latch delay time is given by, [18]
t
latch
=
C
L
g
m,eff
ln
_
V
out
V
0
_
=
C
L
g
m,eff
ln
_
V
DD
/2
V
0
_
(2)
where g
m,eff
is the effective transconductance of the back-to-
back inverters. In fact, this delay depends, in a logarithmic
manner, on the initial output voltage difference at the begin-
ning of the regeneration (i.e., at t = t
0
). Based on (1), V
0
can be calculated from (3)
V
0
=

V
out p
(t = t
0
) V
outn
(t = t
0
)

V
thp

I
2
t
0
C
L
=

V
thp

_
1
I
2
I
1
_
. (3)
The current difference, I
in
= |I
1
I
2
|, between the
branches is much smaller than I
1
and I
2
. Thus, I
1
can be
BABAYAN-MASHHADI AND LOTFI: ANALYSIS AND DESIGN OF A LOW-VOLTAGE LOW-POWER DOUBLE-TAIL COMPARATOR 345
approximated by I
tail
/2 and (3) can be rewritten as
V
0
=

V
thp

I
in
I
1
2

V
thp

I
in
I
tail
= 2

V
thp

1,2
I
tail
I
tail
V
in
= 2

V
thp

1,2
I
tail
V
in
. (4)
In this equation,
1,2
is the input transistors current factor
and I
tail
is a function of input common-mode voltage (V
cm
)
and V
DD
. Now, substituting V
0
in latch delay expression and
considering t
0
, the expression for the delay of the conventional
dynamic comparator is obtained as
t
delay
= t
0
+t
latch
= 2
C
L

V
thp

I
tail
+
C
L
g
m,eff
ln
_
V
DD
4

V
thp

V
in
_
I
tail

1,2
_
. (5)
Equation (5) explains the impact of various parameters. The
total delay is directly proportional to the comparator load
capacitance C
L
and inversely proportional to the input dif-
ference voltage (V
in
). Besides, the delay depends indirectly
to the input common-mode voltage (V
cm
). By reducing V
cm
,
the delay t
0
of the rst sensing phase increases because lower
V
cm
causes smaller bias current (I
tail
). On the other hand, (4)
shows that a delayed discharge with smaller I
tail
results in
an increased initial voltage difference (V
0
), reducing t
latch
.
Simulation results show that the effect of reducing the V
cm
on
increasing of t
0
and reducing of t
latch
will nally lead to an
increase in the total delay. In [17], it has been shown that an
input common-mode voltage of 70% of the supply voltage is
optimal regarding speed and yield.
In principle, this structure has the advantages of high input
impedance, rail-to-rail output swing, no static power consump-
tion, and good robustness against noise and mismatch [1]. Due
to the fact that parasitic capacitances of input transistors do
not directly affect the switching speed of the output nodes,
it is possible to design large input transistors to minimize the
offset. The disadvantage, on the other hand, is the fact that due
to several stacked transistors, a sufciently high supply voltage
is needed for a proper delay time. The reason is that, at the
beginning of the decision, only transistors M
3
and M
4
of the
latch contribute to the positive feedback until the voltage level
of one output node has dropped below a level small enough to
turn on transistors M
5
or M
6
to start complete regeneration. At
a low supply voltage, this voltage drop only contributes a small
gate-source voltage for transistors M
3
and M
4
, where the gate-
source voltage of M
5
and M
6
is also small; thus, the delay time
of the latch becomes large due to lower transconductances.
Another important drawback of this structure is that there is
only one current path, via tail transistor M
tail
, which denes
the current for both the differential amplier and the latch
(the cross-coupled inverters). While one would like a small
tail current to keep the differential pair in weak inversion and
obtain a long integration interval and a better Gm/I ratio, a
Fig. 3. Schematic diagram of the conventional double-tail dynamic com-
parator.
large tail current would be desirable to enable fast regeneration
in the latch [10]. Besides, as far as M
tail
operates mostly in
triode region, the tail current depends on input common-mode
voltage, which is not favorable for regeneration.
B. Conventional Double-Tail Dynamic Comparator
A conventional double-tail comparator is shown in
Fig. 3 [10]. This topology has less stacking and therefore can
operate at lower supply voltages compared to the conventional
dynamic comparator. The double tail enables both a large
current in the latching stage and wider M
tail2
, for fast latching
independent of the input common-mode voltage (V
cm
), and
a small current in the input stage (small M
tail1
), for low
offset [10].
The operation of this comparator is as follows (see Fig. 4).
During reset phase (CLK = 0, M
tail1
, and M
tail2
are off),
transistors M
3
-M
4
pre-charge fn and fp nodes to V
DD
, which
in turn causes transistors M
R1
and M
R2
to discharge the output
nodes to ground. During decision-making phase (CLK =
V
DD
, M
tail1
and M
tail2
turn on), M
3
-M
4
turn off and volt-
ages at nodes fn and fp start to drop with the rate dened
by I
Mtail1
/C
fn(p)
and on top of this, an input-dependent
differential voltage V
fn(p)
will build up. The intermediate
stage formed by M
R1
and M
R2
passes V
fn(p)
to the cross-
coupled inverters and also provides a good shielding between
input and output, resulting in reduced value of kickback
noise [10].
Similar to the conventional dynamic comparator, the delay
of this comparator comprises two main parts, t
0
and t
latch
.
The delay t
0
represents the capacitive charging of the load
capacitance C
Lout
(at the latch stage output nodes, Outn and
Out p) until the rst n-channel transistor (M
9
/M
10
) turns on,
after which the latch regeneration starts; thus t
0
is obtained
346 IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, VOL. 22, NO. 2, FEBRUARY 2014
Fig. 4. Transient simulations of the conventional double-tail dynamic
comparator for input voltage difference of V
in
= 5 mV, V
cm
= 0.7 V,
and V
DD
= 0.8 V.
from
t
0
=
V
Thn
C
Lout
I
B1
2
V
Thn
C
Lout
I
tail2
(6)
where I
B1
is the drain current of the M
9
(assuming V
INP
>
V
INN
, see Fig. 3) and is approximately equal to the half of the
tail current (I
tail2
).
After the rst n-channel transistor of the latch turns on (for
instance, M
9
), the corresponding output (e.g., Outn) will be
discharged to the ground, leading front p-channel transistor
(e.g., M
8
) to turn on, charging another output (Out p) to the
supply voltage (V
DD
). The regeneration time (t
latch
) is achieved
according to (2). For the initial output voltage difference at
time t
0
, V
0
we have
V
0
=

V
out p
(t = t
0
) V
outn
(t = t
0
)

= V
Thn

I
B2
t
0
C
Lout
= V
Thn
_
1
I
B2
I
B1
_
(7)
where I
B1
and I
B2
are the currents of the latch left- and right-
side branches of the second stage, respectively.
Considering I
latch
= |I
B1
I
B2
| = g
mR1,2
V
fn/fp
, (7) can
be rewritten as
V
0
= V
Thn
I
latch
I
B1
2V
Thn
I
latch
I
tail2
= 2V
Thn
g
mR1,2
I
tail2
V
fn/fp
(8)
where g
mR1,2
is the transconductance of the intermediate
stage transistors (M
R1
and M
R2
) and V
fn/fp
is the voltage
difference at the rst stage outputs (fn and fp) at time
t
0
. Thus, it can be concluded that two main parameters
which inuence the initial output differential voltage (V
0
)
and thereby the latch regeneration time are the transcon-
ductance of the intermediate stage transistors (g
mR1,2
) and
the voltage difference at the rst stage outputs (fn and fp)
at time t
0
. In fact, intermediate stage transistors amplify
the voltage difference of V
fn/fp
causing the latch to be
imbalanced.
The differential voltage at nodes fn/fp (V
fn/fp
) at time t
0
can be achieved from
V
fn/fp
=

V
fn
(t = t
0
) V
fp
(t = t
0
)

= t
0

I
N1
I
N2
C
L,fn(p)
= t
0

g
m1,2
V
in
C
L,fn(p)
. (9)
In this equation, I
N1
and I
N2
refer to the discharging
currents of input transistors (M
1
and M
2
), which are dependent
on the input differential voltage (i.e., I
N
= g
m1,2
V
in
).
Substituting (9) in (8), V
0
will be
V
0
= 2V
Thn
g
mR1,2
I
tail2
V
fn/fp
=
_
2V
Thn
I
tail2
_
2

C
Lout
C
L,fn(p)
g
mR1,2
g
m1,2
V
in
. (10)
This equation shows that V
0
depends strongly on the
transconductance of input and intermediate stage transistors,
input voltage difference (V
in
), latch tail current, and the
capacitive ratio of C
Lout
to C
L,fn(p)
. Substituting V
0
in latch
regeneration time (2), the total delay of this comparator is
achieved as follows:
t
delay
= t
0
+t
latch
= 2
V
Thn
C
Lout
I
tail2
+
C
Lout
g
m,eff
ln
_
V
DD
/2
V
0
_
= 2
V
Thn
C
Lout
I
tail2
+
C
Lout
g
m,eff
ln
_
V
DD
I
2
tail2
C
L,fn(p)
8V
2
Thn
C
Lout
g
mR1,2
g
m1,2
V
in
_
. (11)
From the equations derived for the delay of the double-tail
dynamic comparator, some important notes can be concluded.
1) The voltage difference at the rst stage outputs (V
fn/fp
)
at time t
0
has a profound effect on latch initial differen-
tial output voltage (V
0
) and consequently on the latch
delay. Therefore, increasing it would profoundly reduce
the delay of the comparator.
2) In this comparator, both intermediate stage transistors
will be nally cut-off, (since fn and fp nodes both
discharge to the ground), hence they do not play any
role in improving the effective transconductance of the
latch. Besides, during reset phase, these nodes have to
be charged from ground to V
DD
, which means power
consumption. The following section describes how the
proposed comparator improves the performance of the
double-tail comparator from the above points of view.
III. PROPOSED DOUBLE-TAIL DYNAMIC COMPARATOR
Fig. 5 demonstrates the schematic diagram of the proposed
dynamic double-tail comparator. Due to the better performance
of double-tail architecture in low-voltage applications, the
proposed comparator is designed based on the double-tail
structure. The main idea of the proposed comparator is to
increase V
fn/fp
in order to increase the latch regeneration
speed. For this purpose, two control transistors (M
c1
and
M
c2
) have been added to the rst stage in parallel to M
3
/M
4
transistors but in a cross-coupled manner [see Fig. 5(a)].
BABAYAN-MASHHADI AND LOTFI: ANALYSIS AND DESIGN OF A LOW-VOLTAGE LOW-POWER DOUBLE-TAIL COMPARATOR 347
(b) (a)
Fig. 5. Schematic diagram of the proposed dynamic comparator. (a) Main idea. (b) Final structure.
A. Operation of the Proposed Comparator
The operation of the proposed comparator is as follows (see
Fig. 6). During reset phase (CLK = 0, M
tail1
and M
tail2
are off,
avoiding static power), M
3
and M
4
pulls both fn and fp nodes
to V
DD
, hence transistor M
c1
and M
c2
are cut off. Intermediate
stage transistors, M
R1
and M
R2
, reset both latch outputs to
ground.
During decision-making phase (CLK = V
DD
, M
tail1
, and
M
tail2
are on), transistors M
3
and M
4
turn off. Furthermore, at
the beginning of this phase, the control transistors are still off
(since fn and fp are about V
DD
). Thus, fn and fp start to drop
with different rates according to the input voltages. Suppose
V
INP
> V
INN
, thus fn drops faster than fp, (since M
2
provides
more current than M
1
). As long as fn continues falling, the
corresponding pMOS control transistor (M
c1
in this case) starts
to turn on, pulling fp node back to the V
DD
; so another control
transistor (M
c2
) remains off, allowing fn to be discharged
completely. In other words, unlike conventional double-tail
dynamic comparator, in which V
fn/fp
is just a function of
input transistor transconductance and input voltage difference
(9), in the proposed structure as soon as the comparator detects
that for instance node fn discharges faster, a pMOS transistor
(M
c1
) turns on, pulling the other node fp back to the V
DD
.
Therefore by the time passing, the difference between fn and
fp (V
fn/fp
) increases in an exponential manner, leading to
the reduction of latch regeneration time (this will be shown in
Section III-B). Despite the effectiveness of the proposed idea,
one of the points which should be considered is that in this
circuit, when one of the control transistors (e.g., M
c1
) turns on,
a current from V
DD
is drawn to the ground via input and tail
transistor (e.g., M
c1
, M1, and M
tail1
), resulting in static power
consumption. To overcome this issue, two nMOS switches are
used below the input transistors [M
sw1
and M
sw2
, as shown in
Fig. 5(b)].
At the beginning of the decision making phase, due to the
fact that both fn and fp nodes have been pre-charged to V
DD
Fig. 6. Transient simulations of the proposed double-tail dynamic comparator
for input voltage difference of V
in
= 5 mV, V
cm
= 0.7 V, and V
DD
= 0.8 V.
(during the reset phase), both switches are closed and fn and
fp start to drop with different discharging rates. As soon as the
comparator detects that one of the fn/fp nodes is discharging
faster, control transistors will act in a way to increase their
voltage difference. Suppose that fp is pulling up to the V
DD
and fn should be discharged completely, hence the switch in
the charging path of fp will be opened (in order to prevent any
current drawn from V
DD
) but the other switch connected to fn
will be closed to allow the complete discharge of fn node. In
other words, the operation of the control transistors with the
switches emulates the operation of the latch. This will be more
discussed in the following section.
B. Delay Analysis
In order to theoretically demonstrate how the delay is
reduced, delay equations are derived for this structure as
348 IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, VOL. 22, NO. 2, FEBRUARY 2014
previously done for the conventional dynamic comparator and
the conventional double-tail dynamic comparator. The analysis
is similar to the conventional double-tail dynamic comparator,
however; the proposed dynamic comparator enhances the
speed of the double-tail comparator by affecting two important
factors: rst, it increases the initial output voltage difference
(V
0
) at the beginning of the regeneration (t = t
0
); and
second, it enhances the effective transconductace (g
meff
) of
the latch. Each of these factors will be discussed in detail.
1) Effect of Enhancing V
0
: As discussed before, we dene
t
0
, as a time after which latch regeneration starts. In other
words, t
0
is considered to be the time it takes (while both latch
outputs are rising with different rates) until the rst nMOS
transistor of the back-to-back inverters turns on, so that it will
pull down one of the outputs and regeneration will commence.
According to (2), the latch output voltage difference at time
t
0
, (V
0
) has a considerable impact on the latch regeneration
time, such that bigger V
0
results in less regeneration time.
Similar to the equation derived for the V
0
of the double-tail
structure, in this comparator we have
V
0
= V
Thn
I
latch
I
B1
2V
Thn
I
latch
I
tail2
= 2V
Thn
g
mR1,2
I
tail2
V
fn/fp
. (12)
In order to nd V
fn/fp
at t = t
0
, we shall notice that the
combination of the control transistors (M
c1
and M
c2
) with two
serial switches (M
sw1
, M
sw2
) emulates the operation of a back-
to-back inverter pair; thus using small-signal model presented
in [18], V
fn/fp
is calculated by
V
fn/fp
= V
fn(p)0
exp((A
v
1)t /). (13)
In this equation,

A
v
1

=
C
L,fn(p)
G
m,eff1
and V
fn(p)0
is the initial
fn/fp node difference voltage at the time when the correspond-
ing pMOS control transistor is started to be turned on. Hence,
it can be shown that V
fn(p)0
is obtained from
V
fn(p)0
= 2

V
Thp

g
m1,2
V
in
I
tail1
. (14)
Substituting (13) in (12), V
0
will be
V
0
= 2V
Thn
g
mR1,2
I
tail2
V
fn/fp
= 4V
Thn

V
Thp

g
mR1,2
I
tail2
g
m1,2
V
in
I
tail1
exp
_
G
m,eff1
t
0
C
L,fn(p)
_
.
(15)
Comparing (15) with (10), it is evident that V
0
has been
increased remarkably (in an exponential manner) in compare
with the conventional dynamic comparator.
2) Effect of Enhancing Latch Effective Transconductance:
As mentioned before, in conventional double-tail comparator,
both fn and fp nodes will be nally discharged completely.
In our proposed comparator, however, the fact that one of the
rst stage output nodes (fn/fp) will charge up back to the
V
DD
at the beginning of the decision making phase, will turn
on one of the intermediate stage transistors, thus the effective
transconductance of the latch is increased. In other words,
positive feedback is strengthened. Hence, t
latch
will be
t
latch
=
C
Lout
g
m,eff
+ g
mR1,2
ln
_
V
out
V
0
_
=
C
Lout
g
m,eff
+ g
mR1,2
ln
_
V
DD
/2
V
0
_
. (16)
Finally, by including both effects, the total delay of the
proposed comparator is achieved from
t
delay
= t
0
+t
latch
= 2
V
Thn
C
Lout
I
tail2
+
C
Lout
g
m,eff
+ g
mr1,2
ln
_
V
DD
/2
V
0
_
= 2
V
Thn
C
Lout
I
tail2
+
C
Lout
g
m,eff
+ g
mR1,2
ln
_
_
V
DD
/2
4V
Thn

V
Thp

g
mR1,2
I
tail2
g
m1,2
V
in
I
tail1
exp
_
G
m,eff1
t
0
C
L,fn(p)
_
_
_
.
(17)
By comparing the expressions derived for the delay of the
three mentioned structures, it can be seen that the proposed
comparator takes advantage of an inner positive feedback
in double-tail operation, which strengthen the whole latch
regeneration. This speed improvement is even more obvious
in lower supply voltages. This is due to the fact that for larger
values of V
Th
/V
DD
, the transconductance of the transistors
decreases, thus the existence of an inner positive feedback
in the architecture of the rst stage will lead to the improved
performance of the comparator. Simulation results conrm this
matter.
3) Reducing the Energy Per Comparison: It is not only the
delay parameter which is improved in the modied proposed
comparator, but the energy per conversion is reduced as well.
As discussed earlier, in conventional double-tail topology, both
fn and fp nodes discharge to the ground during the decision
making phase and each time during the reset phase they should
be pulled up back to the V
DD
. However, in our proposed
comparator, only one of the mentioned nodes (fn/fp) has to
be charged during the reset phase. This is due to the fact that
during the previous decision making phase, based on the status
of control transistors, one of the nodes had not been discharged
and thus less power is required. This can be seen when being
compared with conventional topologies [see Figs. 9(b) and
10(b)].
IV. DESIGN CONSIDERATIONS
In designing the proposed comparator, some design issues
must be considered. When determining the size of tail transis-
tors (M
tail1
and M
tail2
), it is necessary to ensure that the time
it takes that one of the control transistors turns on must be
smaller than t
0
(start of regeneration)
t
on,Mc1(2)
< t
0

V
Thp

C
L,fn(p)
I
n1,2
<
V
Thn
C
Lout
I
B1

V
Thp

C
L,fn(p)
I
Tail1
2
<
V
Thn
C
Lout
I
Tail2
2
. (18)
BABAYAN-MASHHADI AND LOTFI: ANALYSIS AND DESIGN OF A LOW-VOLTAGE LOW-POWER DOUBLE-TAIL COMPARATOR 349
This condition can be easily achieved by properly designing
the rst and second stage tail currents. Even if possible in the
fabrication technology, low-threshold pMOS devices can be
used as control transistors leading to faster turn on.
In designing the nMOS switches, located below the input
transistors, the drain-source voltage of these switches must
be considered since it might limit the voltage headroom,
restricting the advantage of being used in low-voltage applica-
tions. In order to diminish this effect, low-on-resistance nMOS
switches are required. In other words, large transistors must
be used. Since the parasitic capacitances of these switches do
not affect the parasitic capacitances of the fn/fp nodes (delay
bottlenecks), it is possible to optimally select the size of the
nMOS switch transistors in a way that both low-voltage and
low-power operations are maintained.
The effect of mismatch between controlling transistors on
the total input-referred offset of the comparator is another
important issue. When determining the size of controlling
transistors (M
C1
M
C2
), two important issues should be
considered. First, the effect of threshold voltage mismatch
and current factor mismatch of the controlling transistors
on the comparator input-referred offset voltage. Second, the
effect of transistor sizing on parasitic capacitances of the
fn/fp nodes, i.e., C
L,fn(p)
, and consequently the delay of the
comparator. While larger transistors are required for better
matching; however, the increased parasitic capacitances are
delay bottlenecks. In order to study the effect of threshold
and current factor mismatch of control transistors on the
total input-referred offset voltage, a brief mismatch analysis
is presented here.
A. Mismatch Analysis
In principle, the effect of threshold voltage mismatch and
current factor mismatch of controlling transistors is almost
negligible in most cases except for the situation where input
differential voltage (V
in
) is very small where fn and fp have
approximately similar discharging rates. This is true because
by the time that the controlling transistor (M
c1
or M
C2
) turns
on, the differential input signal is already amplied to large
amplitude compared to the mismatches. In other words, offset
due to the controlling transistor mismatches is divided by the
gain from the input to the output. However, in case of small
V
in
, when fn and fp follow each other tightly, the mismatch
of the controlling transistors might inuence the result of the
comparison. Hence, the following brief analyzes the effect
of threshold and current factor mismatches of controlling
transistors on the total input-referred offset voltage.
1)Effect of Threshold Voltage Mismatch of M
C1
, M
C2
, i.e.,
V
ThC1,2
:
The differential current due to the threshold voltage mis-
match can be obtained from
i
diff
= g
mc1,2
V
Thc1,2
(19)
where g
mc1,2
is the transconductance of the controlling tran-
sistors. So, the input-referred offset voltage due to the M
c1,2
threshold voltage mismatch is obtained as follows:
V
eq,dueV
Thc1,2
=
g
mc1,2
V
Thc1,2
g
m1,2
=

p
W
C1,2
V
ODC1,2

n
W
1,2
V
OD1,2
V
Thc1,2
(20)
where V
OD
refers to the overdrive voltage of the transistors.
2)Effect of Current-Factor Mismatch M
C1
, M
C2
, i.e.,

C1,2
:
In order to calculate the input-referred offset due to the
current factor mismatch of M
C1,2
,
C1,2
is modeled as a
channel width mismatch W, i.e., / = W/W. The
differential current that W generates can be obtained as
expressed in (21).
i
diff
=
1
2

p
C
ox
W
L
(V
gsc1,2
V
thc1,2
)
2
. (21)
Note that the controlling transistors are in saturation since
|V
GDc1,2
| = |V
fn
V
fp
| < |V
thp
|. So the input-referred offset
voltage due to the current factor mismatch is calculated from
V
eq,due
C1,2
=
i
diff
g
m1,2
=
0.5
p
W
C1,2
(V
gsc1,2
V
thp
)
2

n
W
1,2
(V
gs1,2
V
thn
)
=
0.5
p
W
C1,2
V
2
OD
C1,2

n
W
1,2
(V
cm
R
CLK
K
nC1,2
V
2
OD
C1,2
V
thn
)
(22)
where V
cm
is the input common mode voltage and R
clk
is the
equivalent on resistance of the tail transistor.
Assuming both mismatch factors, the total input-referred
offset due to the mismatch of the controlling transistors can
be found from

total
=
_

2
V
ThC1,2
+
2

C1,2
. (23)
From (20) and (22), it can be concluded that the ratio of
the controlling transistor sizes to the input transistor size, i.e.,
(W
C1,2
/W
1,2
), is effective in reducing the offset. Due to the
fact that the transconductance of the input transistors (g
m1,2
)
is important in amplifying the input differential voltage and
due to the dominant role of the size of these transistors on
total input-referred offset, usually large input transistors are
designed, which results in diminishing the effect of controlling
transistors mismatch.
B. Kickback Noise
Principally in latched comparators, the large voltage vari-
ations on the regeneration nodes are coupled, through the
parasitic capacitances of the transistors, to the input of the
comparator. Since the circuit preceding it does not have zero
output impedance, the input voltage is disturbed, which may
degrade the accuracy of the converter. This disturbance is
usually called kickback noise. In [16], it has been shown
that the fastest and most power efcient comparators gen-
erate more kickback noise. This is true about our proposed
dynamic comparator. Although it improves the double-tail
350 IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, VOL. 22, NO. 2, FEBRUARY 2014
Fig. 7. Peak input voltage error due to kickback noise.
Fig. 8. Layout schematic diagram of the proposed dynamic comparator.
topology in terms of operation speed and thus energy per
comparison, the kickback noise is increased in comparison to
conventional double-tail structure (Fig. 3). Fig. 7 presents the
peak disturbance as a function of differential input voltage of
the comparator in three studied architectures. While double-
tail structure takes advantage of input-output isolation and
thus the minimum kickback noise, the conventional dynamic
comparator and our proposed structure has nearly similar
kickback noise. However, in our proposed comparator since
control transistors are not supposed to be as strong as the latch
transistors in conventional dynamic comparator, it is possible
to determine the size of those transistors in a way that keeps
the advantages of the speed enhancement and power reduction,
while reducing kickback noise. Besides, for some applications
where kickback becomes important, it is possible to apply sim-
ple kickback reduction techniques, such as neutralization [16]
to remarkably reduce the kickback noise (See Fig. 7, proposed
dynamic comparator with neutralization).
V. SIMULATION RESULTS
In order to compare the proposed comparator with the
conventional and double-tail dynamic comparators, all cir-
cuits have been simulated in a 0.18-m CMOS technology
with V
DD
= 1.2 V. The comparators were optimized and
the transistor dimensions were scaled to get an equal offset
standard variation of
OS
= 8 mV at the input common-
mode voltage of V
cm
= 1.1 V (the same conditions that are
(b)
(a)
Fig. 9. (a) Post-layout simulated delay and (b) energy per conversion as a
function of supply voltage (V
in
= 50 mV, V
cm
= V
DD
0.1).
(a)
(b)
Fig. 10. (a) Post-layout simulated delay and (b) energy per conversion as a
function of input common-mode voltage (V
in
= 50 mV, V
DD
= 1.2).
TABLE I
SUMMARY OF THE COMPARATOR PERFORMANCE
Item Value
Technology 180-nm CMOS
Supply voltage 1.2 V
Average power dissipation per conversion
@ freq. = 500 MHz
329 W
Worst case delay (V
cm
= 0.6 V,
V
in
= 1 mV)
550 ps
Delay/log(V
in
) 69 ps/dec
Offset standard deviation
(1-sigma) (
os
)
7.8 mV
Energy efciency 0.66 pJ
found in [10]). Fig. 8 shows the layout of the comparator.
Particular care was taken in the layout to avoid affecting delay
and power of the comparator. Fig. 9(a) and (b) demonstrates
BABAYAN-MASHHADI AND LOTFI: ANALYSIS AND DESIGN OF A LOW-VOLTAGE LOW-POWER DOUBLE-TAIL COMPARATOR 351
TABLE II
PERFORMANCE COMPARISON
Comparator structure Conventional Dynamic
Comparator
Double-tail Dynamic
Comparator
Proposed Dynamic Comparator
Technology CMOS 180 nm 180 nm 180 nm
Supply voltage (V) 0.8 V 0.8 V 0.8 V
Maximum sampling frequency 900 MHz 1.8 GHz 2.4 GHz
Delay/log(V
in
) (ps/dec.) 940 358 294
Peak transient noise voltage
at regeneration time(nV)
215 n 221 n 219 n
Kickback noise voltage (at V
in
= 10 mV) 51.3 mV 5.3 mV 43 mV With neutralization: 13 mV
Energy per conversion (J) 0.3 p 0.27 p 0.24 p Without M
sw1
and M
sw2
: 0.265 p
Input-referred offset voltage (mV) 7.89 mV 7.91 mV 7.8 m
Estimated area 16 16 28 12 28 14
the post-layout simulation results of the delay and the energy
per conversion of the mentioned dynamic comparators versus
supply voltage variation. As shown in Fig. 9(a), in comparison
with the other two structures, the delay of the proposed double-
tail dynamic comparator is signicantly reduced in low-voltage
supplies. It is obvious that at high supply voltages, all struc-
tures have approximately similar performances, about 200 ps
clk-to-output delay (including clock buffer) with 0.65 pJ/bit-
conversion for 8-mV offset. However, by decreasing the supply
voltage, three structures start to behave differently. It is evident
that the double-tail topology can operate faster and can be used
in lower supply voltages, while consuming nearly the same
power as the conventional dynamic comparator. The case is
even much better for the proposed comparator when compared
to the conventional double-tail topology. For instance, the
proposed comparator can operate in 0.6 V supply at the
cost of 106 fJ/conversion with 840 ps delay versus 1.81 ns
for the conventional double-tail comparator and 3.5 ns for
the conventional topology. Our simulations show that if the
circuit is optimized for V
DD
= 0.6 V, the results would
be even better for the proposed circuit. Fig. 10 shows the
simulated performance as a function of input common-voltage
(V
cm
). Generally in the double-tail topologies, the delay of
the comparator is less inuenced by the variation of the input
common-mode voltage in comparison with the conventional
dynamic topology and thus has a wider common-mode range.
The power consumption is nearly equal.
Fig. 11 depicts the dependence of the comparator delay
on power supply level at various differential input voltages.
For V
in
= 10 mV, the delay is 460 ps at V
DD
= 0.9 V.
This delay drops from 460 to 162 ps when V
DD
changes
from 0.9 to 1.5 V. In addition, at a given V
DD
, the larger the
differential input voltage, the smaller the comparator delay will
be. Fig. 12 shows the simulated delay of the comparator versus
differential input voltage under different conditions of input
common-mode voltage (V
cm
) at V
DD
= 1.2 V. The delay of the
comparator at V
in
= 1 mV and V
cm
= 700 m is 413 ps. For
a given value of V
cm
, the delay decreases as differential input
voltage increases. Furthermore, the delay is also dependent
on the variation of common-mode voltage. For example, at
V
in
= 10 mV, the delay increases by 64 ps, from 239 to
303 ps, as V
cm
decreases from 900 to 700 mV. As shown in
Fig. 11. Delay of the proposed comparator versus supply voltage (V
DD
).
Fig. 12. Delay of the proposed comparator versus input voltage difference
(V
in
).
Fig. 13, the standard deviation of the offset of the proposed
comparator is achieved to be
OS
= 7.8 mV using Monte
Carlo simulations for a run of 300 samples. Table I summarizes
the performance of the proposed dynamic comparator. Finally,
352 IEEE TRANSACTIONS ON VERY LARGE SCALE INTEGRATION (VLSI) SYSTEMS, VOL. 22, NO. 2, FEBRUARY 2014
Fig. 13. Histogram diagram of Monte Carlo simulation of the offset in
proposed dynamic comparator.
Table II compares the performance of the proposed comparator
with the conventional dynamic and double-tail comparators.
In 0.18-m CMOS, the proposed comparator provides the
maximum sampling of 2.4 GHz at 0.8-V supply voltage.
VI. CONCLUSION
In this paper, we presented a comprehensive delay analy-
sis for clocked dynamic comparators and expressions were
derived. Two common structures of conventional dynamic
comparator and conventional double-tail dynamic comparators
were analyzed. Also, based on theoretical analyses, a new
dynamic comparator with low-voltage low-power capability
was proposed in order to improve the performance of the
comparator. Post-layout simulation results in 0.18-m CMOS
technology conrmed that the delay and energy per conversion
of the proposed comparator is reduced to a great extent in
comparison with the conventional dynamic comparator and
double-tail comparator.
REFERENCES
[1] B. Goll and H. Zimmermann, A comparator with reduced delay time in
65-nm CMOS for supply voltages down to 0.65, IEEE Trans. Circuits
Syst. II, Exp. Briefs, vol. 56, no. 11, pp. 810814, Nov. 2009.
[2] S. U. Ay, A sub-1 volt 10-bit supply boosted SAR ADC design in
standard CMOS, Int. J. Analog Integr. Circuits Signal Process., vol. 66,
no. 2, pp. 213221, Feb. 2011.
[3] A. Mesgarani, M. N. Alam, F. Z. Nelson, and S. U. Ay, Supply
boosting technique for designing very low-voltage mixed-signal circuits
in standard CMOS, in Proc. IEEE Int. Midwest Symp. Circuits Syst.
Dig. Tech. Papers, Aug. 2010, pp. 893896.
[4] B. J. Blalock, Body-driving as a Low-Voltage Analog Design Technique
for CMOS technology, in Proc. IEEE Southwest Symp. Mixed-Signal
Design, Feb. 2000, pp. 113118.
[5] M. Maymandi-Nejad and M. Sachdev, 1-bit quantiser with rail to rail
input range for sub-1V modulators, IEEE Electron. Lett., vol. 39,
no. 12, pp. 894895, Jan. 2003.
[6] Y. Okaniwa, H. Tamura, M. Kibune, D. Yamazaki, T.-S. Cheung,
J. Ogawa, N. Tzartzanis, W. W. Walker, and T. Kuroda, A 40G-
b/s CMOS clocked comparator with bandwidth modulation technique,
IEEE J. Solid-State Circuits, vol. 40, no. 8, pp. 16801687, Aug. 2005.
[7] B. Goll and H. Zimmermann, A 0.12 m CMOS comparator requiring
0.5V at 600MHz and 1.5V at 6 GHz, in Proc. IEEE Int. Solid-State
Circuits Conf., Dig. Tech. Papers, Feb. 2007, pp. 316317.
[8] B. Goll and H. Zimmermann, A 65nm CMOS comparator with modied
latch to achieve 7GHz/1.3mW at 1.2V and 700MHz/47W at 0.6V, in
Proc. IEEE Int. Solid-State Circuits Conf. Dig. Tech. Papers, Feb. 2009,
pp. 328329.
[9] B. Goll and H. Zimmermann, Low-power 600MHz comparator for
0.5 V supply voltage in 0.12 m CMOS, IEEE Electron. Lett., vol. 43,
no. 7, pp. 388390, Mar. 2007.
[10] D. Shinkel, E. Mensink, E. Klumperink, E. van Tuijl, and B. Nauta,
A double-tail latch-type voltage sense amplier with 18ps Setup+Hold
time, in Proc. IEEE Int. Solid-State Circuits Conf., Dig. Tech. Papers,
Feb. 2007, pp. 314315.
[11] P. Nuzzo, F. D. Bernardinis, P. Terreni, and G. Van der Plas, Noise
analysis of regenerative comparators for recongurable ADC archi-
tectures, IEEE Trans. Circuits Syst. I, Reg. Papers, vol. 55, no. 6,
pp. 14411454, Jul. 2008.
[12] A. Nikoozadeh and B. Murmann, An analysis of latched comaprator
offset due to load capacitor mismatch, IEEE Trans. Circuits Syst.
II, Exp. Briefs, vol. 53, no. 12, pp. 13981402, Dec. 2006.
[13] S. Babayan-Mashhadi and R. Lot, An offset cancellation technique
for comparators using body-voltage trimming, Int. J. Analog Integr.
Circuits Signal Process., vol. 73, no. 3, pp. 673682, Dec. 2012.
[14] J. He, S. Zhan, D. Chen, and R. J. Geiger, Analyses of static and
dynamic random offset voltages in dynamic comparators, IEEE Trans.
Circuits Syst. I, Reg. Papers, vol. 56, no. 5, pp. 911919, May 2009.
[15] J. Kim, B. S. Leibowits, J. Ren, and C. J. Madden, Simulation and
analysis of random decision errors in clocked comparators, IEEE Trans.
Circuits Syst. I, Reg. Papers, vol. 56, no. 8, pp. 18441857, Aug. 2009.
[16] P. M. Figueiredo and J. C. Vital, Kickback noise reduction technique
for CMOS latched comapartors, IEEE Trans. Circuits Syst. II, Exp.
Briefs, vol. 53, no. 7, pp. 541545, Jul. 2006.
[17] B. Wicht, T. Nirschl, and D. Schmitt-Landsiedel, Yield and speed
optimization of a latch-type voltage sense amplier, IEEE J. Solid-State
Circuits, vol. 39, no. 7, pp. 11481158, Jul. 2004.
[18] D. Johns and K. Martin, Analog Integrated Circuit Design, New York,
USA: Wiley, 1997.
Samaneh Babayan-Mashhadi was born in Mash-
had, Iran, in 1984. She received the B.S. and M.S.
degrees (Hons.) in electrical engineering from the
Ferdowsi University of Mashhad, Mashhad, Iran, in
2006 and 2008, respectively, where she is currently
pursuing the Ph.D. degree, from the Electrical Engi-
neering Department, with research on design of low-
power high-resolution, high-speed analog-to-digital
converters.
Her current research interests include low-power
low-voltage analog and mixed-signal integrated
circuits.
Reza Lot received the B.Sc. degree from Ferdowsi
University of Mashhad, Mashhad, Iran, the M.Sc.
degree from the Sharif University of Technology,
Tehran, Iran, and the Ph.D. degree from the Uni-
versity of Tehran, Tehran, in 1997, 1999, and 2004,
respectively, all in electrical engineering.
Since 2004, he has been with the Ferdowsi Univer-
sity of Mashhad, where he is currently an Associate
Professor. From 2008 to 2009, he was with the
Electronics Research Laboratory, Delft University
of Technology, Delft, the Netherlands, as a Post-
Doctoral Scientic Researcher involved in research on ultralow-power analog
and mixed-signal integrated circuits for biomedical applications. His current
research interests include low-voltage low-power analog integrated circuit
design for biomedical applications and high-speed data converters for telecom-
munication systems.
Dr. Lot has been an Associate Editor of the IEEE TRANSACTIONS ON
CIRCUITS AND SYSTEMSPART I: REGULAR PAPERS since 2010.

You might also like