A Frequency Model of a Continuously Driven Clocked CMOS Comparator
956 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—II: EXPRESS BRIEFS, VOL. 57, NO. 12, DECEMBER 2010 A Frequency Model of a Continuously Driven Clocked CMOS Comparator Shunsuke…
956 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—II: EXPRESS BRIEFS, VOL. 57, NO. 12, DECEMBER 2010
A Frequency Model of a Continuously
Driven Clocked CMOS Comparator
Shunsuke Okura, Hajime Shibata, Member, IEEE, Tetsuro Okura,
Toru Ido, Member, IEEE, and Kenji Taniguchi, Member, IEEE
Abstract—A frequency model of a continuously driven clocked
CMOS comparator with the effect of the input signal during
regeneration is presented. The model utilizes a small-signal linear
model derived from the theoretical analysis of the comparison
error caused by the transition from the tracking mode to the
regeneration mode. The comparison error voltage is a function
of input signal frequency and is represented with the transfer
function. The correctness of the model is assured by several
transistor-level simulation results. The model provides a valuable
insight for the design of high-speed comparators.
Index Terms—Clocked comparator, CMOS, frequency re-
sponse, metastability, regeneration.
FOR THE design of analog-to-digital converters (ADCs),a comparator is a crucial and often a limiting component
toward high-speed operation because of its finite accuracy
and comparison speed. Among the main building blocks of
comparators, flip-flops and latches have been well analyzed. A
formula describing the probability of a metastable state occur-
rence has been presented in . A formula for the sensitivity
of a latch in a sense amplifier has been described in . The
latch has been theoretically analyzed with small-signal models
in  and . Moreover, there have been many reports on high-
speed ADC or comparators that discuss the effect of the static
offset voltage, the metastability, the reset time of the latch –
, the frequency response of the preamp , , and design
optimizations of high-speed comparators , , .
This brief presents a frequency model of continuously driven
clocked CMOS comparators taking into account the effect of
the input signal during regeneration. A clocked comparator
has two modes, namely, the tracking mode and the regener-
ation mode. According to the study on the transition from
the tracking mode to the regeneration mode, the comparison
result depends on the input signal frequency. This dependence
is caused by the input signal change just after the beginning
of regeneration while the regeneration gain is still small. The
Manuscript received April 19, 2010; revised July 24, 2010; accepted
October 11, 2010. Date of current version December 15, 2010. This work
was supported in part by a grant from the Global COE Program, “Center
for Electronic Devices Innovation,” from the Ministry of Education, Culture,
Sports, Science and Technology of Japan. This paper was recommended by
Associate Editor J. Paramesh.
S. Okura, T. Okura, T. Ido, and K. Taniguchi are with the Graduate School
of Engineering, Osaka University, Suita 556-0871, Japan.
H. Shibata is with Analog Devices, Inc., Toronto, ON M5G 2C8, Canada.
Color versions of one or more of the figures in this paper are available online
Digital Object Identifier 10.1109/TCSII.2010.2087972
Fig. 1. Schematic of a clocked CMOS comparator.
comparison error that depends on the input signal frequency is
named ac error in this brief. Based on the small-signal analysis,
the ac error voltage is given by a simple function of input
signal frequency. The frequency response of the comparator
is modeled with the derived ac error. The presented model is
useful for designs using high-speed comparators. In previous
work, the dynamic nature of the comparator that includes the
frequency response of the latch is analyzed . This brief
delivers a frequency model based on the detailed analysis.
This brief is organized as follows. Section II shows that the
comparison result depends on the input signal frequency. The
ac error voltage is derived by the theoretical analysis with a
small-signal model, and the comparator is modeled to represent
the frequency dependence in Section III. Section IV provides
SPICE simulation results with a 0.18-μm CMOS process model
to confirm the derived model. The conclusion is drawn in
II. EFFECT OF THE INPUT SIGNAL CHANGE
Fig. 1 shows a schematic of the clocked CMOS comparator
having two parts, a preamp and a regenerative latch. The input
signal is continuously applied to the comparator. The preamp
drives the latch according to the voltage difference between vin
and vref . When CLK is high, the latch is in a reset state and
tracks the difference signal with nonzero ON-resistance of the
switch. At this tracking mode, the comparator shows a negative
exponential response since vout is shorted to ground. Just at
the time of the switch turning to open, vout is regeneratively
amplified to the rail-to-rail level by a positive feedback loop
in the amplifier of latch, and finally, the comparison result of
vin and vref is stored in the latch. Since the regeneration time
constant is a finite nonzero value, the signal generated by the
preamp during regeneration introduces the ac error voltage,
which depends on the input signal frequency.
Fig. 2 shows the transient simulation results of the clocked
comparator, in which the ac error voltage is described. The
1549-7747/$26.00 © 2010 IEEE
OKURA et al.: FREQUENCY MODEL OF A CONTINUOUSLY DRIVEN CLOCKED CMOS COMPARATOR 957
Fig. 2. Comparator transient response depending on the input signal fre-
quency. (a) Input signal vin. (b) Comparator output vout.
input signal is a sine wave whose frequency is 200, 160, 120,
and 80 MHz, which is continuously applied to the comparator
input [see Fig. 2(a)]. The reference signal is a 40-mV dc
voltage. The comparator switches from the tracking mode to the
regeneration mode at 50 ns. At this instance, all input signals
are 0 mV, and all tracked signals are negative. As shown in
Fig. 2(b), the outputs of 200- and 160-MHz cases increase,
whereas the outputs of 120- and 80-MHz cases decrease. This
means that the comparison result depends on both the difference
signal and the ac error voltage, which becomes larger as the
frequency increases. The ac error voltage is a function of input
signal frequency because it is introduced by the input signal
changes in a short period just after the clock edge while the
regeneration gain is still small.
III. SMALL-SIGNAL ANALYSIS
A. Theoretical Derivation
The ac error voltage is studied with a small-signal circuit on
the transition from the tracking mode to the regeneration mode.
For simplicity, the comparator switches from the tracking mode
to the regeneration mode at 0.0 s. The comparator is first
analyzed in the tracking mode to derive the initial value in the
regeneration mode. The regeneration process is then analyzed.
The ac error voltage is derived from the zero-to-one transition
point V0to1, which is the reference voltage when the comparator
is in a metastable state for a given input signal. The comparison
result is logic high if vref < V0to1, whereas it is logic low if
vref > V0to1. The regeneration response is represented with the
positive exponential term. When the coefficient of the positive
Fig. 3. Equivalent circuit of the comparator.
exponential term is 0.0, the comparator results in a metastable
state, and V0to1 is derived. The ac error voltage is given by
Verr,ac = V0to1 − vin(0) (1)
where vin(0) is the sampled input signal.
The comparator is simplified as the equivalent circuit shown
in Fig. 3. The preamp is composed of an amplifier and a
transconductance gmin . The amplifier is not broken down
into components since it is conventionally well known. The
switching conductance is gsw. The negative conductance −gml
provides regenerative feedback of the latch. The variable γ,
which is the ratio of gsw to gml, is kept above 1.0 so that the
latch can reset the prior data stored. The static dc offset voltage
Vos,static is in series to the negative conductance because the
comparator has a potentially large offset voltage originating
from the latch whose device size is kept small to minimize the
regeneration time constant and power consumption. The output
capacitance is represented as CL. A sinusoidal input signal
applied to the comparator is given by
vin(t) = sin(ωint + θin) (2)
where ωin and θin are the frequency and the phase, respectively.
The input signal amplitude is normalized to 1.0 V. The transfer
function of the amplifier in the preamp Hpreamp(s) is given by
1 + sτpreamp
where Apreamp and τpreamp are the dc gain and the time
constant, respectively. The amplifier output is given by the
Laplace transform of (2) and (3) as
Based on KCL, the latch output during the tracking mode is
958 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—II: EXPRESS BRIEFS, VOL. 57, NO. 12, DECEMBER 2010
which is presented as
(γ − 1)gml
K1K2 sin(θin + θ1 + θ2)
− vref − gml
preamp + 1
θ1 ≡ −arctan(ωinτpreamp) (10)
track + 1
θ2 ≡ −arctan(ωinτtrack). (12)
The tracking error, which has been analyzed in , , and
, is not considered in this analysis, so that the comparator is
supposed to be at steady state until 0.0 s.
In the regeneration mode, KCL gives the output signal
=gminVx(s)+CLvout(0−) (t≥0) (13)
τregvout(0−)+ gmingml Vx(s)−
where τreg is the regeneration time constant . Substituting
(4) and (8) into (14), the inverse Laplace transform of the
equation gives the output presented as
vout(t) = ac1 · exp(−jωint) + ac1 · exp(jωint)
+ ac2 + ac3 · exp
(t ≥ 0) (16)
where the coefficients ac1, ac2, and ac3 are independent of
time (see the Appendix). vout(t) either increases or decreases
exponentially with time depending on the sign of ac3, whereas
the first three terms in the right side of (16) do not contribute to
the regeneration result. If ac3 > 0, vout(t) ends up with logic
high, whereas vout(t) does with logic low if ac3 < 0. When
ac3 = 0, the comparator is in a metastable state. From ac3 = 0,
the zero-to-one transition point is represented as
V0to1 =K1K2K3 sin(θin + θ1 + θ2 + θ3)
1 + ω2inτ
θ3 ≡ arctan (ωinτreg) . (19)
The first term in the right side of (17) shows the magnitude
(K1K2K3) and the phase shift (θ1 + θ2 + θ3) for sin(θin).
The variables K1, θ1, K2, θ2, K3, and θ3, which are given
by (9)–(12), (18), and (19), respectively, are the functions of
the input signal frequency ωin. The second term in the right
side of (17) indicates the input-referred voltage of the static dc
offset, which does not affect the ac error. The ac error voltage
is therefore represented as
Verr,ac = K1K2K3 sin(θin + θ1 + θ2 + θ3)− sin(θin) (20)
Fig. 4. Frequency model of the comparator. (a) With an additional summing
block. (b) Replacement of (a).
referring to (1). It is noted that the sampled input signal is given
by vin(0) = sin(θin). The ac error voltage is the attenuated
input signal with phase shifts.
B. Behavioral Model
The frequency response of the comparator is modeled with
the derived ac error voltage. The comparator is conventionally
modeled as a summing block with quantization noise added to
the input signal . When the ac error is taken into account,
the comparator is modeled with an additional summing block,
as shown in Fig. 4(a), where the sum of the input signal and the
ac error is noted as a zero-to-one transition point. From (20),
the sum of the sampled input signal and the ac error voltage is
V ′0to1(ωin) = vin(0) + Verr,ac
=K(ωin) sin (θin + θ(ωin)) (21)
K(ωin) ≡K1 ·K2 ·K3 (22)
θ(ωin) ≡ θ1 + θ2 + θ3 (23)
noting that the variables Ki and θi (i = 1, 2, 3) are functions
of the input signal frequency ωin. Since the sampled signal is
sin(θin), (21) replaces the model of the comparator shown in
Fig. 4(a) with that in Fig. 4(b). The transfer function H(s) is
given by the magnitude K(ωin) and the phase θ(ωin). In order
to derive H(s), magnitude Ki and phase θi (i = 1, 2, 3) are
transformed as follows: K1 and θ1, which were given by (9)
and (10), respectively, are transformed to
∣∣∣∣ 11 + jωinτpreamp
1 + jωinτpreamp
K2 and θ2, which were given by (11) and (12), respectively, are
∣∣∣∣ 11 + jωinτtrack
1 + jωinτtrack
OKURA et al.: FREQUENCY MODEL OF A CONTINUOUSLY DRIVEN CLOCKED CMOS COMPARATOR 959
Fig. 5. Plot of (30) when τpreamp = τreg.
and K3 and θ3, which were given by (18) and (19), respectively,
are transformed to
∣∣∣∣ 11− jωinτreg
Therefore, H(s) is derived as
1 + sτpreamp︸ ︷︷ ︸
1 + sτtrack︸ ︷︷ ︸
1− sτreg︸ ︷︷ ︸
The three poles of (30) correspond to the poles of the preamp,
the latch in the tracking mode, and the latch in the regeneration
The plot of (30) is shown in Fig. 5, setting τpreamp = τreg as
a simple example. It is noted that τtrack is given by τreg/(γ − 1)
according to (7) and (15). The bandwidth is saturated when
γ > 2, because the bandwidth is limited by Hpreamp(s) and
Hreg(s), whose poles are located at a lower frequency than that
of Htrack(s). On the other hand, the phase delay at −3 dB
decreases as γ increases because the sum of angles of
Hpreamp(s) and Hreg(s) is zero and the phase shift depends
only on Htrack(s). The phase shift in the comparator is kept
small by setting γ high at τpreamp = τreg. This is useful for the
design of the comparator, which is sensitive to the phase shift.
IV. SIMULATION RESULT
Fig. 6 presents an example schematic of a fully differen-
tial high-speed comparator. Two differential input pairs (Mn1,
Mn2, and Mn3, Mn4) sense the difference between the analog
inputs and the reference voltages and generate currents accord-
ingly. These currents are mirrored from load transistors (Mp1
and Mp2) to p-channel transistors (Mp3 and Mp4). A pair of
p-channel cascode transistors (Mp5 and Mp6) biased at a given
voltage Vb prevents kickback noise . The latch utilizes an
Fig. 6. Example of a fully differential high-speed CMOS comparator.
VARIABLES OF THE COMPARATOR DERIVED FROM SPICE SIMULATIONS
Fig. 7. Comparison of zero-to-one transition points between the model and
simulated results for input signal frequencies.
n-channel cross-coupled transistor (Mn6 and Mn7) and a
CMOS switch (Mn5 and Mp7).
Table I shows the variables of the comparator derived
from SPICE simulations with the transistor-level netlist and a
0.18-μm CMOS process model. It is noted that the two poles of
the preamp, namely, the mirror pole and the cascode pole, are
represented with τpreamp1 and τpreamp2.
The simulated zero-to-one transition points versus the input
signal frequency ωin are shown in Fig. 7, where the input signal
phase is 0 rad. The dots are found with transient simulation
runs. The solid line shows the presented frequency model of the
comparator, where the variables listed in Table I are used. The
solid line is close to the dots, so that the presented model pre-
cisely estimates the zero-to-one transition points. The dashed
line shows the conventional model only with the frequency
response of the preamp. The conventional model significantly
deviates from the dots because the effect of the input signal
during regeneration is large at high frequencies.
960 IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS—II: EXPRESS BRIEFS, VOL. 57, NO. 12, DECEMBER 2010
Fig. 8. Comparison of zero-to-one transition points between the model and
simulated results for input signal phases at ωin = 3.14 Grad/s.
Fig. 8 shows the zero-to-one transition points versus the input
signal phase θin at ωin = 3.14 Grad/s. The ac error voltage is
not negligible at this high frequency. The presented model of
the solid line shows good fitting to the simulation results of dots
and provides fine estimation of the zero-to-one transition points
at each input signal phase, whereas the mismatch between
the conventional model and simulation results is considerably
large. These simulation results demonstrate the accuracy of the
A novel frequency model of a continuously driven clocked
comparator has been presented. The model includes a transfer
function to represent the effect of the input signal during regen-
eration, which was derived from the small-signal analysis on
the transition from the tracking mode to the regeneration mode.
The transfer function is given by the product of three poles,
which are functions of basic design parameters. In addition, a
comparison between the model estimation and simulated results
for an example design has been described. The presented model
shows a precise estimation of a comparator response with
high-frequency input signals. Therefore, the presented model
provides a valuable insight for high-speed comparator design.
The time response of the comparator during the regeneration
mode is given by (16) with coefficients ac1, ac2, and ac3. These
coefficients are respectively given by
ac1 = −ApreampK1 gmin
sin(θin + θ1) + j cos(θin + θ1)
2(1 + jωinτreg)
vref + Vos,static (32)
(K1K3 sin(θin + θ1 + θ3)− vref)
− Vos,static + vout(0−). (33)
The first two terms in the right side of (16) is represented
with (31) as
ac1 · exp(−jωint) + ac1 · exp(jωint)
= −ApreampK1K3 gmin
sin(ωint + θin + θ1 + θ3). (34)
S. Okura would like to thank T. Wakimoto and C. W.
Mangelsdorf of Analog Devices, K. K., for making possible the
internship during which this work was performed and for their
assistance, support, and encouragement throughout this project.
This work was partly carried out at the Frontier Research Base
for Global Young Researchers, Osaka University, on the Pro-
gram of Promotion of Environmental Improvement to Enhance
Young Researchers’ Independence, the Special Coordination
Funds for Promoting Science and Technology, Japan Ministry
of Education, Culture, Sports, Science and Technology.
 H. J. M. Veendric, “The behavior of flip-flops used as synchronizers and
prediction of their failure rate,” IEEE J. Solid-State Circuits, vol. SSC-15,
no. 2, pp. 169–176, Apr. 1980.
 R. Sarpeshkar, J. L. Wyatt, Jr., N. C. Lu, and P. D. Gerber, “Mismatch
sensitivity of a simultaneously latched CMOS sense amplifier,” IEEE J.
Solid-State Circuits, vol. 26, no. 10, pp. 1413–1422, Oct. 1991.
 W. A. M. Van Noije, W. T. Liu, and S. J. Navarro, Jr., “Precise final
state determination of mismatched CMOS latches,” IEEE J. Solid-State
Circuits, vol. 30, no. 5, pp. 607–611, May 1995.
 G. M. Yin, F. O. Eynde, and W. Sansen, “A high-speed CMOS comparator
with 8-b resolution,” IEEE J. Solid-State Circuits, vol. 27, no. 2, pp. 208–
211, Feb. 1992.
 B. Zojer, R. Petschacher, and W. A. Luschnig, “A 6-bit/200-MHz full
Nyquist A/D converter,” IEEE J. Solid-State Circuits, vol. SSC-20, no. 3,
pp. 780–786, Jun. 1985.
 B. Peetz, B. D. Hamilton, and J. Kang, “An 8-bit 250 Megasample per
second analog-to-digital converter: Operation without a sample and hold,”
IEEE J. Solid-State Circuits, vol. SSC-21, no. 6, pp. 997–1002, Dec. 1986.
 R. J. V. D. Plassche and P. Baltus, “An 8-bit 100-MHz full-Nyquist analog-
to-digital converter,” IEEE J. Solid-State Circuits, vol. 23, no. 6, pp. 1334–
1344, Dec. 1988.
 C. W. Mangelsdorf, “A 400-MHz input Flash converter with error
correction,” IEEE J. Solid-State Circuits, vol. 25, no. 1, pp. 184–191,
 I. Mehr and D. Dalton, “A 500-MSample/s, 6-bit Nyquist-rate ADC
for disk-drive read-channel applications,” IEEE J. Solid-State Circuits,
vol. 34, no. 7, pp. 912–920, Jul. 1999.
 M. Choi and A. A. Abidi, “A 6-b 1.3-Gsample/s A/D converter in
0.35-μm CMOS,” IEEE J. Solid-State Circuits, vol. 36, no. 12, pp. 1847–
1858, Dec. 2001.
 K. Uyttenhove, J. Vandenbussche, E. Lauwers, G. Gielen, and
M. Steyaert, “Design techniques and implementation of an 8-bit 200-MS/s
interpolating/averaging CMOS A/D converter,” IEEE J. Solid-State
Circuits, vol. 38, no. 3, pp. 483–494, Mar. 2003.
 B. Razavi and B. A. Wooley, “Design techniques for high-speed, high-
resolution comparators,” IEEE J. Solid-State Circuits, vol. 27, no. 12,
pp. 1916–1926, Dec. 1992.
 S. Okura, T. Ohkura, K. Taniguchi, and H. Shibata, “Frequency response
analysis of latch utilized in high-speed comparator,” in Proc. IEEE Int.
Conf. Electron., Circuits, Syst., 2006, pp. 1077–1080.
 R. J. Baker, H. W. Li, and D. E. Boyce, CMOS Circuit Design, Layout,
and Simulation. New York: Wiley-IEEE Press, 1997.
 R. J. Baker, CMOS, Mixed-Signal Circuit Design. New York: Wiley-
IEEE Press, 2002.
 K. Sushihara, H. Kimura, Y. Okamoto, K. Nishimura, and A. Matsuzawa,
“A 6b 800 MSample/s CMOS A/D converter,” in Proc. Int. Solid State
Circuits Conf., Feb. 2000, pp. 428–429.