1

Simplified Unified Analysis of Switched-RC Passive

Mixers, Samplers and N -Path Filters using the
Adjoint Network
Shanthi Pavan, Senior Member, IEEE and Eric Klumperink, Senior Member, IEEE

Abstract—Recent innovations in software defined CMOS radio R C1,2,3,4 = C

transceiver architectures heavily rely on high linearity switched- vx
RC sampler and passive-mixer circuits, driven by digitally pro- ϕ4 ϕ3 ϕ2 ϕ1
grammable multi-phase clocks. Although seemingly simple circuits, vi v4 v3 v2 v1
the frequency domain analysis of these Linear Periodically Time
C4 C3 C2 C1
Variant (LPTV) circuits is often deceptively complex. This paper
uses the properties of sampled LPTV systems and the adjoint
(inter-reciprocal) network to greatly simplify the analysis of the
τ
switched-RC circuit. We first derive the transfer function of the
equivalent linear time-invariant filter relating the input to the voltage ϕ1
sampled on the capacitor in the switched-RC kernel. We show ϕ2 Ts
τ= N
how a leakage resistor across the capacitor can easily be addressed ϕ3
using our technique. A signal-flow graph is then developed for the ϕ4
complete continuous-time voltage waveform across the capacitor, 0 Ts
and simplified for various operating regions. We finally derive the Fig. 1. A 4-path switched-capacitor network that forms the basis for a 4-path
noise properties of the kernel. The results we derive have largely passive mixer and the 4-path filter.
been reported in prior works, but the use of the adjoint network
simplifies the derivation, while also providing circuit insight.

approximated assuming RC ≫ τ . This has been labeled the

“mixing region” approximation [4] or the “passive mixer” region
I. I NTRODUCTION
[5], [6]. Under this condition, the voltage on the capacitors
High linearity samplers and passive mixers are being increas- becomes virtually constant for an input-frequency close to the
ingly used in radio transceivers, both for frequency translation in switching frequency [7]–[11]. The average power dissipation
the analog domain, as well as for conversion to and from the in the resistor R can now be evaluated and modeled by an
discrete-time and/or digital domain [1], [2]. Moreover, passive equivalent shunt impedance Zsh in parallel to the capacitor in an
mixers can be leveraged to implement frequency translated filter- Linear Time Invariant (LTI) equivalent model [7], [8]. Baseband
ing in mixer-first receivers or so called N -path filters [3]. As the resistance in parallel to the capacitor results in an extra dissipation
filter center frequency can be controlled by a digital clock, this and hence an extra loss resistance.
allows for flexibly programmable software-defined radio archi- Another analysis method targeting down-conversion mixer
tectures. The target is to replace external surface acoustic wave analysis assumes an RF current source with parallel impedance
(SAW) filters by highly integrated SAW-less CMOS implemen- driving four baseband lowpass impedances via 4 switches with
tations, while retaining interference robustness [2]. The removal 25% duty cycle [12], [13]. The RF-input frequency fRF is
of SAW filtering increases the receiver’s linearity requirements, confined to fs /2 and 3fs /2, and the RF-current is approximated
and has hence fueled research and development in high linearity in a Fourier series. The RF input voltage is then written in the
MOSFET switch-RC circuits. An example of such a circuit is form of an infinite sum. However, due to the input frequency
shown in Fig. 1. limitation, signal folding from frequencies higher than 3fs /2
Depending on how the outputs are used, the 4-path circuit cannot be analyzed.
in Fig. 1 can act as a time-interleaved sampler, a quadrature The above mentioned methods limit the frequency range, and
I/Q mixer or an N -path filter. Although the circuit may seem do not predict harmonic responses and signal folding. A large RC-
simple, the analysis of its frequency domain transfer function time is assumed, i.e. mixer region operation, which does not cover
easily becomes involved. To keep the analysis tractable, non- the sampling region. In contrast, Soer et. al. use the methods of
overlapping clock phases and simplified models for the switches Strom and Signell [14] to provide a unified analysis of samplers
are usually assumed. We briefly review some of the results of and mixers in [4], albeit with lengthy integral calculations to
prior works, the simplifying assumptions made, and then propose find exact expressions for the harmonic transfer functions of
our technique based on the impulse response of the adjoint an N -path switch-R-C circuit driven by non-overlapping ideal
network. clocks (see for example, the 4-path circuit in Fig. 1). The key
The transfer function and input impedance of circuits like design parameters are the input and switching frequency, duty
the one in Fig. 1 excited by a sinewave voltage input can be cycle D = τ /Ts = 1/N , and the RC time-constant. A sampling
and mixing region were defined, based on the symmetry found manner compared to the frequency-domain method of [4]. Section
in the noise analysis results around Γ ≡ τ /RC = 2 (Fig. 12 VI concludes the paper.
in [4]). Strongly simplified expressions for the harmonic transfer
functions in the sampling and mixing region were derived, which
also formed the basis for noise analysis. Later work, building II. LPTV N ETWORKS WITH S AMPLED O UTPUTS
on this analysis derived the filter transfer and noise for N -path The method of analysis of N -path filters and mixers pre-
band-pass [15] and notch filters [16] and an equivalent RLC sented in this paper exploits two key properties of sampled LPTV
filter model, gain-boosted N -path filters [17], and an N -path networks [19] summarized below.
filter preceded by a series-inductor for low-pass pre-filtering [18].
Such pre-filtering is important, as mixer-first receivers and N -
(a)
path filters show folding of RF-signals around multiples of fs
(similar to aliasing in discrete time circuits, albeit with more y(t) ∑
LPTV System
attenuation). x(t) y[lTs ]δ(t − lTs )
@fs
Whereas the unified analysis of mixing and sampling an- l

alyzes the continuous-time output, the unified analysis of [5], ∑

[6] focusses on the voltages sampled on the capacitors. The δ(t − lTs )
x(t) = ej2πf t
l
authors of [5], [6] show that the results from [4] in the mixing (b)
region can be reproduced assuming ideal reconstruction filtering
of the sampled capacitor voltage via a Zero-Order-Hold block LTI System ŷ(t) ∑
(ZOH). Further, their analysis allows for the incorporation a x(t) ∑ y[lTs ]δ(t − lTs )
parallel resistor to the capacitor in Fig. 1. Although the analysis Heq (f ) = Hk (f ) l
k
is certainly simpler than in [4] and provides valuable insight, it ∑
still involves considerable algebra. j2πf t
δ(t − lTs )
x(t) = e l
This paper, like [5], [6], recognizes the crucial role played Fig. 2. (a) An LPTV system (varying at a rate fs ) whose output is sampled at
by the sampled capacitor voltages on the properties of the N - fs . (b) An LTI system whose output sampled at fs yields the same sequence as
path system. Our analysis then exploits the properties of sampled the LPTV system above.
LPTV networks to derive the results from [5], [6] and [4] in a
simpler manner (using lot less algebra). The rest of this paper Consider an LPTV network varying at a frequency fs , as
is organized as follows. In Section II, we summarize the key shown in Fig. 2(a). It is excited by a complex exponential x(t) =
results on sampled LPTV networks that are relevant to this work ej2πf t . The output y(t) is sampled with the period Ts ≡ 1/fs as
[19]. It turns out that the output samples of an LPTV system shown in the figure. Since the system is LPTV, the output consists
varying at fs , and excited by an input x(t) (when the sampling of frequencies of the form (f + kfs ), where k is an integer. We
rate is also fs ) can be thought of as being obtained by exciting thus have
an linear time-invariant (LTI) filter by x(t) and sampling its ∑∞

output at fs . Further, the impulse response of the equivalent LTI y(t) = Hk (f )ej2π(f +kfs )t . (1)
system, denoted by heq (t), is obtained using the adjoint (or inter- k=−∞

reciprocal) LPTV network. The samples of y(t) are given by

Section III uses the results of Section II to derive the
∞
∑ ∞
∑
expressions for the equivalent LTI filter that relates the input
to the voltages held on the capacitors in the N -path structure y(nTs ) = Hk (f )ej2π(f +kfs )nTs = ej2πf nTs Hk (f ).
k=−∞ k=−∞
of Fig. 1. We show that the effect of leakage resistance can be (2)
incorporated using our technique. We also reflect on the apparent
Consider now the system of Figure 2(b). It is a linear time-
simplicity that characterizes our derivation when compared to
invariant
∑ system, whose frequency response Heq (f ) is chosen to
prior works.
be k Hk (f ), where Hk (f ) are the harmonic transfer functions
Section IV then derives a block diagram for the complete of the LPTV system of Figure 2(a). If this LTI system is excited
waveform of the voltage across one of the capacitors in the N - by ej2πf t , its output is
path structure of Fig. 1. We then show the equivalence of our
results to those in [4] and [5]. While the block diagram derived ∞
∑
in this paper appears more complex compared to that in [4], it ŷ(t) = ej2πf t Hk (f ). (3)
turns out that it yields extra insight into the relationship between k=−∞

mixer operation and the sampled capacitor voltage. In addition, When sampled at a rate fs = 1/Ts , we obtain
the model also allows analysis of N -path filters. We then derive
simplified expressions for the operation of the N -path switch-RC ∞
∑
network in the N -path filter and passive-mixer modes, and place ŷ(nTs ) = ej2πnf Ts Hk (f ). (4)
them in context of prior work. Section V exploits the adjoint k=−∞

impulse-response technique to derive the mean-square noise and

From (2) and (4), we see that as far as output samples are
noise spectral density of the switched-RC kernel in a simpler
concerned, an LPTV system whose output is sampled at fs is
equivalent to an LTI system with output sampled at fs . The input port of N̂ is heq (t + to ), as shown in Fig. 3(b).†
equivalent LTI filter has a frequency response [19] How does one determine N̂ ? The adjoint network N̂ has the
∞
same graph as N , and can be derived from N by applying the
∑ following element-by-element substitution rules shown in Table I
Heq (f ) = Hk (f ). (5)
[20]–[22]:
k=−∞
1) A branch in N that is a linear resistor, capacitor, or inductor
Since any arbitrary input x(t) can be represented as a sum remains unchanged in N̂ .
of complex exponentials via the Fourier transform, it follows that 2) A periodically operated switch in N controlled by a wave-
the output samples of an LPTV system (when the sampling rate is form ϕ(t) is replaced in N̂ by a switch that is controlled
the same as that at that the system is varying) can be thought of as by ϕ(−t).
being obtained by exciting an LTI filter by x(t) and sampling its 3) Linear controlled sources in N are replaced by appropriate
output at a rate fs . We emphasize that the equivalence holds only linear controlled sources in N̂ . For instance, a CCCS in
for samples, and not necessarily for the waveforms. Referring to N is replaced by a VCVS in N̂ , with the controlling and
Figs. 2(a) and 2(b), we note that y(nTs ) = ŷ(nTs ), but y(t) need controlled ports interchanged, as seen in Table I.
not equal ŷ(t). TABLE I
How do we determine the transfer function of the equivalent T RANSFORMATIONS OF LINEAR CONTROLLED SOURCES AND PERIODICALLY
OPERATED SWITCHES FROM N TO N̂ .
LTI system Heq (f ), given an LPTV system? A possible way of
doing this is to determine Hk (f ) of the latter, and then use (5).
This, however, leads to tedious algebra. N N̂
Fortunately, it turns out that the concept of adjoint (or
inter-reciprocal) networks can be used to determine the impulse v1 µv1 µi2 i2
response heq (t) corresponding to Heq (f ) in only a few steps.
The derivation and details are given in [19], we give the results
here for the convenience of the reader.
i1 µi1 µv2 v2

(a)
1 ϕ1 2
vi (t) + ϕ2
v2 (lTs + to ) v1 gm v1 gm v2 v2
−
1′ N 2′

i1 Ri1 Ri2 i2
(b) Adjoint Network
1 ϕ̂1 2 ϕ(t) ϕ(−t)
ϕ̂2 δ(t + to )
heq (t + to )
1′ N̂ 2′
As we show in the next section, the use of the adjoint
Fig. 3. (a) An LPTV network varying at fs , whose output v2 (t) is periodically network greatly simplifies the analysis of the switch-RC N -path
sampled at fs . (b) Determining the impulse response heq (t) of the equivalent structure. Adjoint (or inter-reciprocal) networks are well known in
LTI filter using the inter-reciprocal (or adjoint) network. An impulse current is
injected into the “output” port of the adjoint, and the current in the input port is circuit theory (see for instance [22]), and form the workhorse for
determined, yielding heq (t). transfer function (.XF), sensitivity and noise analysis in circuit
simulators [20], [21]. The adjoint network concept has earlier
been applied to the determination of transfer functions and noise
Fig. 3(a) shows a 2-port LPTV network N being excited by
in CT∆ΣMs [23]–[25], where the input is in continuous-time,
a voltage input vi (t). Suppose we are interested in samples of
while the output is a discrete-time sequence.
v2 (t), taken at a time offset to from the LPTV sampling clock,
namely the sequence v2 (lTs + to ). As discussed previously, one III. T HE S WITCH -RC N-PATH C IRCUIT
can find an equivalent LTI filter with impulse response heq (t),
Fig. 1 shows an N -path switched-capacitor network, with
which when excited by vi (t), yields an output whose samples
N = 4. While the capacitors are labeled C1,··· ,4 to aid under-
are identical to v2 (lTs + to ). heq (t) can be determined using the
standing, they are all equal. Only one of clocks ϕ1,··· ,4 is high
adjoint (or interreciprocal) network N̂ corresponding to N , as
shown in Fig. 3(b). † Note that exciting output port 2 with a current source and sensing the current

Since the output voltage of N is sampled at (lTs + to ), the at port 1 by adding a short does NOT change the port termination impedance
for network N̂ compared to N . Similarly a current output will be excited by a
output port of N̂ is excited by a current impulse occurring at voltage source, while voltage sensing will be used in case of a current driven
−to . It turns out [19] that the resulting current flowing in the 2-port, as is done with the analysis of adjoint LTI networks.
at any given time. Each of the waveforms is high for a duration We are interested in determining vo (t), which can be sepa-
τ ≡ Ts /N . Ideal switches are assumed. Clearly, the network is rated into two parts; voff (t) which occurs when the switch is off,
LPTV, varying at fs . If vi (t) = ej2πf t , then, and von (t), when the switch is turned on. When the switch is
∞
∑ off, the capacitor simply holds its voltage. To determine voff (t),
v1 (t) = Hl (f )ej2π(f +lfs )t (6) therefore, we would like to know the voltages sampled on the
l=−∞ capacitor just before the falling edges of the clock waveform
(i.e., at vo (kTs )). In the notation of Fig. 3, to = 0−† .
where the Hl are the harmonic transfer functions. Since the
capacitors are equal, and the switch controls are delayed versions The switched-RC circuit is an LPTV network varying at
of one another, it is easy to show that a frequency fs , and we are interested in determining vo (kTs ),
∞
which is its output sampled at the same frequency as that at
∑ 2π(n−1)l which the network is varying. From the discussion in the previous
vn (t) = Hl (f )ej N ej2π(f +lfs )t , n = 1, · · · , 4. (7)
section [19], we know that vo (kTs ) can be thought of as being
l=−∞
the samples at the output of a linear time-invariant (LTI) filter,
As seen above, the magnitudes of the harmonic transfer functions driven by vi (t). Further, [19] shows that the impulse response of
(HTF) to the capacitor voltages are identical as all the N paths this equivalent LTI filter heq (t), can be readily obtained from the
have equal component values but only differ in clock-phase. Thus, adjoint network (also called the inter-reciprocal network).
the HTFs only differ in phase: the lth HTF of vn undergoes an The adjoint network corresponding to our example is as
additional phase shift of (2π(n − 1)/N )l radians. With N = 4, shown in Fig. 4(b). Note that the switch-control signal in the
we have 4-phase quadrature signals: original network is reversed in time. The determination of heq (t)
v1 (t) = · · · + H0 (f )ej2πf t + H1 (f )ej2π(f +fs )t + · · · proceeds as follows. The original LPTV network was sampled at
v2 (t) = · · · + H0 (f )ej2πf t + jH1 (f )ej2π(f +fs )t + · · · zero timing-offset, meaning that we are interested in vo (t) just
before kTs , which means to = 0−. Thus, the “output” port of
v3 (t) = · · · + H0 (f )ej2πf t − H1 (f )ej2π(f +fs )t + · · · the adjoint is excited by an impulsive current at t = 0+ (just
v4 (t) = · · · + H0 (f )ej2πf t − jH1 (f )ej2π(f +fs )t + · · · after the switch is closed), and the current waveform through the
input port is recorded.
With a proper choice of R and C, harmonically combining
Referring to Fig. 4(b), the current impulse causes the ca-
v1,··· ,4 can result in a frequency-selective, image-reject, harmonic
pacitor voltage to instantly increase to vx (0+) = 1/C † . For
down-conversion mixer [4]. The voltage vx , on the other hand,
0 < t < τ , the capacitor discharges through the resistor with a
has a frequency-selective characteristic centered around fs ≡
time-constant RC, as shown in Fig. 4(c). iout (t) during this in-
1/Ts and some of its multiples [12], [15].
terval, therefore, is (1/RC) exp(−t/RC). Just before t = τ , the
From the discussion above, we conclude that determining the capacitor voltage is vx (τ −) = (1/C) exp(−τ /RC) ≡ βvx (0+).
HTFs from vi to v1 is sufficient; the HTFs to the other capac- For τ ≤ t < Ts , the switch is open, and iout = 0. The voltage
itor voltages can be obtained using phase-rotational symmetry. across the capacitor does not change during this period. We
Further, since only one of the switches in Fig. 1 is on at any denote iout for 0 ≤ t < Ts by p(t), as shown in red in Fig. 4(c).
time, there is no coupling between the capacitors. This means
At t = Ts +, the switch is closed again, and the capacitor
that analyzing the circuit with one of the switched branches is
begins to discharge again, but now from a value βvx (0+). Note
all that is needed. This is the so called “switched-RC kernel” [4],
that heq (t) = iout is the response to a voltage vx (0+) on the
shown in Fig. 4(a).
capacitor at t = 0+. Therefore, the response to a capacitor
voltage βvx (0+) at t = Ts + must be βheq (t − Ts ). This is a
Sampling instant Impulse injected
consequence of linearity and the periodically time-varying nature
τ τ
of the network. Thus, we can write the recursion
heq (t) = p(t) + βheq (t − Ts ). (8)
(a) 0 Ts (b) 0 Ts
R R In the frequency domain,
vo P (f )
vx vx 0 Heq (f ) = . (9)
vi C iout C δ(t) 1 − βe−j2πf Ts
Observing Fig. 4(c) shows that p(t) can be expressed as the
( ) difference between two decaying exponentials as follows, where
(c) p(t) β = exp − RC
τ
h(t) ≡ e−t/RC u(t) and u(t) denotes the unit-step function.
1
RC 1
iout (t) β p(t) = (h(t) − βh(t − τ )). (10)
RC β2 RC
RC
on off on off
t Thus,
1
0 τ Ts 2Ts
P (f ) = (1 − βe−j2πf τ ). (11)
1 + j2πf RC
Fig. 4. (a) The switched-RC network. (b) The adjoint (inter-reciprocal) network.
The impulse response corresponding to the equivalent LTI filter Heq (f ) is the † Throughout this paper, t− and t+ denote the time instants just before and
current waveform iout obtained by injecting at impulse at the “output” port of just after t respectively.
the adjoint network; i.e., heq (t) = iout (t). (c) iout (t). † The 1/C actually has dimensions of voltage, since the 1 stands for 1 Coulomb.
 heq (t) sampler, and interpret it as being a “lossy accumulator” operating
(a) kTs
on discrete samples. This interpretation runs into difficulties when
1 − RC
t
C1,··· ,4 begin to interact through parasitic capacitance at vx and/or
RC e
vi (t) vk
β β the series inductance of the source. We therefore, advocate the
τ Ts use of Fig. 5(a) to model the relationship between vi and the
delay delay
voltage sampled on C1 .
Fig. 5(b) plots the magnitude responses of the equivalent LTI
(b) Γ= Ts
N RC filter for various values of Γ ≡ τ /RC. For large Γ (Γ ≫ 2) the
1
second term in (12) is approximately unity, and Heq (f ) reduces
to the transfer function of a first-order RC lowpass filter. Γ ≫ 2
0.8 Γ = 20 denotes the “sampling mode” [4] of operation. Γ ≪ 2, on the
other hand, results in the “passive-mixer” [5] mode of operation.
|Heq (f )|

0.6
Γ=2 In this mode, Heq (f ) has narrow peaks around multiples of fs
0.4 (i.e., harmonic selectivity), as seen in Fig. 5(b). Many of these
Γ = 0.2
peaks can be eliminated by appropriately combining v1,··· ,4 in
0.2
Fig. 1, resulting in mixers with reduced signal and noise folding
using the principles of N -path operation [26].
−8 −6 −4 −2 0 2 4 6 8

f /fs
τ
Fig. 5. (a) Block diagram form of the equivalent LTI filter that relates vi (t) to
the output samples. (b) Magnitude response of Heq (f ) for various values of Γ.
Γ ≪ 2 represents the mixing mode, while Γ ≫ 2 indicates the sampling mode. 0 Ts
R vc
(a)
vx
Combining (9) and (11), we have iout C RL δ(t)

1 1 − βe−j2πf τ − (R∥R
τ
− TRs −τ
Heq (f ) = . (12) β=e L )C ·e L C
(1 + j2πf RC) 1 − βe−j2πf Ts
(b) p(t)
1
RC
1 − (RL ∥R)C
t
Heq (f ) obtained above is identical to that derived in [5] and [4]. RC e
heq (t)
The technique based on the adjoint network that we used above
t
is seen to yield the same result in a much simpler fashion.
0 τ Ts 2Ts
It is natural to wonder why the technique above yields (12) (c)
in far fewer steps than in prior works. Our explanation is the 1/C
following. Earlier works arrive at the result that the sampled value vc (t)
of the voltage across the capacitor can be related to the input vi t
through an LTI filter during the course of analysis. In contrast, 0 τ Ts 2Ts
we do not discover this while analyzing this particular circuit; we Fig. 6. (a) Adjoint network for the switched-RC kernel with leakage resistor.
recognize that this property is fundamental to all sampled LPTV (b) Current waveform through the resistor. (c) Capacitor voltage waveform.
networks; we also know that the equivalent impulse response can
be found using the adjoint network. Using the adjoint method, the effect of a leakage resistance
Further, [4], [5] excite the switch-RC network with a sinu- RL in shunt with each of the capacitors in Fig. 1(a) can be
soid to determine Heq , while we use an impulse current. The analyzed as easily as the “non-leaky” case. Fig. 6(a) shows the
“bookkeeping” of the state for the impulse response calculation adjoint network, where RL denotes the leakage resistor. The
is simpler for an impulse input, as it “dies” immediately after initial voltage on the capacitor is 1/C, resulting in an initial
application. This is in contrast to what happens with a sinewave current heq (0+) = 1/RC. For 0 ≤ t ≤ τ , heq (t) is an
excitation. This is reflected in the far fewer steps needed to arrive exponentially decaying waveform with time-constant (R∥RL )C.
at Heq (f ) in our analysis when compared to prior works. When the switch is opened at t = τ , heq (t) goes to zero.
Fig. 5(a) expresses the relationship between the input vi However, the capacitor voltage continues to decay with a time-
and the sampled output in block diagram form. As discussed constant RL C during the interval τ ≤ t < Ts , as shown in
earlier, vi can be thought of as being filtered by an LTI system Fig. 6(c). Thus, the capacitor voltage at t = Ts + is given by
with impulse response heq (t) (or frequency response Heq (f )), (Ts −τ )
− (R∥R
τ
−
{z· e
whose output is sampled at multiples of kTs . The LTI system vx (Ts +) = |e } vx (0+). (13)
L )C RL C

can be thought of as a cascade of a first-order RC filter whose ≡β

output is delayed, scaled and subtracted from it, followed by
a transfer function 1/(1 − βe−j2πf Ts ). Since this is periodic Proceeding as we did while deriving (12), it is straightfor-
with a frequency 1/Ts , the authors of [5] move it beyond the ward to see that with leakage, β and P (f ) in (9) should be
replaced by • We first express the output voltage across the capac-
(Ts −τ )
− (R∥R
τ
− itor as the sum of two waveforms: voff (t) that is
β=e L )C ·e RL C
(14)
non-zero when the switch is open, i.e., the time in-
and − τ tervals kTs ≤ t < (k + 1)Ts − τ , and von (t), which is
RL 1 − e (R∥RL )C e−j2πf τ
P (f ) = . (15) non-zero when the switch is closed, i.e., the intervals
RL + R (1 + j2πf (R∥RL )C) (k + 1)Ts − τ ≤ t < (k + 1)Ts , as shown in blue and red
These equations are the same as those derived in [5]. Setting RL in Fig. 7(b) respectively.
to ∞ in the expressions above yields (12), as it should. Again, • Next, we determine voff (t). This is very straightforward if
our analysis to account for leakage resistance is straightforward. we know the voltages sampled on the capacitor at the falling
Moreover, the time-domain response can provide insight into the edges of the clock, namely at times t = kTs −. We denote
effect on frequency selectivity. The leakage resistance results in a the sampled capacitor voltage sequence by vk . Then, it is
reduced peak H0 , and a lower Q (reduced frequency selectivity), easily seen that voff (t) is simply the waveform obtained by
since the impulse response dies out more rapidly. Referring to exciting a zero-order-hold (ZOH) that∑holds for a duration
Figs. 4(b) and 6(b), we see that β can physically be interpreted (Ts − τ ) by the impulse waveform k vk δ(t − kTs ), as
as the fraction of charge remaining on C at the beginning of shown in Fig. 7(c).
t = Ts +. • We then invoke the results of the previous section to
determine vk . As we saw there, vk can be thought of as
IV. C OMPLETE O UTPUT WAVEFORM the sampled output of a linear time invariant filter with
impulse response heq (t), when excited by vin . heq (t) can
τ be conveniently found using the adjoint network approach.
• Next, we set out to determine von (t), which is non-zero
(a) (b) only in the intervals (k + 1)Ts − τ < t < (k + 1)Ts . Note
0 Ts
R
that von (t) must not only depend on the voltage held on
kTs (k + 1)Ts
vo 1 the capacitor when the switch is closed vk (this forms the
w(t)
0 “initial condition”) as well as vin (t). The derivation of
vi C von (t) proceeds as discussed below.
vo [k + 1]
off on vo [k] von (t) consists of two parts: a decaying portion due to past
τ inputs and a part due to the current input. The first part is due
to “previous history”, namely the voltage vk already stored on
(c)
∑ the capacitor, which will exponentially decay as it discharges
vi heq (t) vk δ(t − kTs ) through the resistor with a time-constant RC. This is modeled
k
∑ by the output of the path labeled ⃝a in Fig. 7(c). The explanation
δ(t − kTs ) for path ⃝ a is as follows. When the switch is turned on at t =
∑ k
vk δ(t − kTs ) (k+1)Ts −τ , von (t) must be vk . This starts to decay exponentially
k (Ts − τ )
until t = (k + 1)Ts , at which time the switch is opened. This
voff (t)
ZOH process can be modeled by exciting a filter with impulse response
vk
βvk h(t)−β·h(t−τ ) with an impulse of strength vk at t = (k+1)Ts −
τ
(Ts − τ ) 0 τ (recall that h(t) = e−t/RC u(t) and β = e−τ /RC ). The resulting
(k + 1)Ts ⃝
delay h(t)
a
waveform at ⃝ a is shown in the inset. Fig. 8 shows the waveforms
that comprise von (t) in more detail. va (in blue), which is non-
delay β zero only when the switch is closed, shows the exponentially
∑
vo (t) decaying waveform due to path ⃝. a At t = (k + 1)Ts − τ , va
Ts
τ
delay β
jumps from 0 to vk . At t = (k + 1)Ts −, it has decayed to βvk ,
and goes abruptly to zero at t = kTs .
h(t) The second part of von (t) is due to vi (t) exciting the RC
w(t) ⃝d circuit when the switch is closed. This part, in turn, is the sum
of two components. The first part is the output due to vi exciting
1 ⃝c ⃝b
vi RC h(t) the RC network, and can be modeled by driving a filter with an
impulse response (1/RC)e−t/RC u(t) = (1/RC)h(t)u(t) with
−
the time-windowed input w(t) · vi (t), shown as the output of
t
h(t) ≡ e RC u(t)

Fig. 7. (a) The switched-RC circuit. (b) Timing and output waveforms. (c) Signal path ⃝.
b1 w(t), as shown in Fig. 7(b), is 1 when the switch is
flow diagram. closed and 0 otherwise. Fig. 8 shows vi and w(t)vi in gray and
black respectively. The voltage vb1 shows the output of path ⃝. b1
We now determine the complete output waveform of the As seen in the figure, vb1 persists when w(t)vi = 0 due to the
switched-RC kernel (without the leakage resistor RL ). The circuit memory of the capacitor.
(not the adjoint) is shown in Fig. 7(a). The analysis proceeds as At t = (k + 1)Ts , however, the output of path ⃝ b
follows. in Fig. 7(c) must go to zero, since von (t) is zero for
 w(t) · vi initial conditions. In the opinion of the authors, the flow graph of
0.8 Fig. 7(c) is more intuitively appealing, since in physical terms,
vi
vb = vb1 + vb2 the input excites the RC-network only when the switch is closed,
0.6
i.e., w(t) = 1.
0.4 va vb1
In [4], Soer et. al propose a model for the output of the
0.2
switched-RC network. This model, derived using state-space and
vb
0 frequency domain methods, appears (very) different. It turns
vb2 out that it is equivalent to the model of Fig. 7(c), though the
-0.2
equivalence is not apparent. Soer’s model looks much simpler
-0.4
(k + 1)Ts − τ (k + 1)Ts that of Fig. 7(c). In the subsection that follows, we demonstrate
t
that our signal-flow diagram is equivalent to that of Soer’s model.
We also show that there is a price to be paid for the apparent
Fig. 8. vi (t), w · vi (t) and various components of von (t). simplicity of that model.

(k + 1)Ts < t < (k + 2)Ts − τ † . The memory effect of the ca- A. Connection to Soer’s Model
pacitor, reflected as the non-zero output of path ⃝ b1 in the interval
Observing Fig. 7(c) shows that the contributions of the
(k + 1)Ts < t < (k + 2)Ts − τ , should therefore be canceled.
β paths in ⃝ a and ⃝ b are equal and opposite, and therefore
This can be achieved as follows. If we knew the output of path
cancel. Combining the remaining portions of paths ⃝ a and ⃝ b
⃝
b1 at t = (k +1)Ts (which we denote by vb1 (k + 1)), subtracting
and recognizing that the ZOH can be represented as an integrator
vb1 (k + 1)e−(t−(k+1)Ts )/RC u(t − (k + 1)Ts ) from the output of
cascaded with a filter of impulse response δ(t) − δ(t − (Ts − τ )),
⃝
b1 will result in path ⃝ b having a zero output for the intervals
we obtain the simplified signal flow graph of Fig. 9(a). Path ⃝ 1
(k + 1)Ts < t < (k + 2)Ts − τ . This is due to the following.
generates voff (t), while the sum of paths ⃝ 2 and ⃝3 yield von (t).
At t = (k + 1)Ts , the input to the filter (with impulse response
Combining the outputs of ⃝ 1 and ⃝ 2 of this figure leads to the
(1/RC)h(t) in Fig. 7(c)) goes to zero, and vb1 (k + 1) will decay
signal flow graph in part (b). While this flow graph looks simpler,
exponentially with a time-constant RC.
observe that the separation between voff (t) and von (t) is lost.
How do we determine vb1 (k + 1)? Just when the switch
The impulse response of the integrator is the unit step u(t),
is turned on, the capacitor voltage is vk due to the past input
and h(t) = e−t/RC u(t). Thus, their difference (which forms
memorized on the capacitor. When the switch is turned off at a
the input of the upper path in Fig. 9(b)) is (1 − e−t/RC )u(t).
time τ later, this voltage will have decayed to vk e−τ /RC ≡ βvk .
This is immediately recognized as being the step-response of
At the same time, the output due to w · vi (t) alone, which is
an RC-filter with impulse response (1/RC)h(t)u(t). This can
the output of path ⃝ b1 is vb1 (k + 1). By linearity, the capacitor
be represented as a cascade of an integrator and a filter whose
voltage vo at (k + 1)Ts , therefore, has to be βvk + vb1 (k + 1).
impulse is (1/RC)h(t), as shown in Fig. 9(c). Reordering the
However, we know that, by definition, the capacitor voltage
blocks of the top-most path, we obtain the signal-flow diagram
at (k + 1)Ts is vk+1 . Thus, vb1 (k + 1) = vk+1 − βvk . From
of Fig. 9(d). The final step is to recognize the ZOH of width
the discussion in the previous paragraph, we should subtract
(Ts − τ ) in this figure, and move the summer before the filter,
vb1 (k + 1)e−(t−(k+1)Ts )/RC u(t − (k + 1)Ts ) from the output of
yielding Fig. 9(e), which is Soer’s model [4].
path ⃝b1 so that von (t) = 0 for (k + 1)Ts ≤ t < (k + 2)Ts − τ .
This can be accomplished by exciting a filter with impulse While Soer’s model seems much simpler than that of
response h(t) with (vk+1 − βvk )δ(t − (k + 1)Ts ), as shown Fig. 7(c) (and that shown in Fig. 17 in [5]), it makes it more
in path ⃝. b2 As shown in Fig. 8, the output of path ⃝, b2 difficult to understand the operation of the switched-RC kernel
namely vb2 , is the negative of vb1 when the switch is off, as part of an N -path filter, or in the passive-mixer mode. The
i.e., for (k + 1)Ts ≤ t < (k + 2)Ts − τ . As a result, their primary reason for this is that attempting to merge the various
sum is zero during that interval. von (t) is simply the sum paths in Fig. 7(c), while yielding a simpler looking signal-flow
va + vb = va + vb1 + vb2 . diagram, also makes it difficult to separate voff (t) and von (t)
(which is needed to analyze the operation as an N -path filter), or
The complete output waveform of the switched-RC circuit,
identify the key constituents of the output in the passive-mixer
therefore, is the sum of voff (t), and von (t) which in turn is
mode.
comprised of paths ⃝ a and ⃝, b as shown in Fig. 7(c).
From an input-output perspective, the signal flow diagram of
Fig. 7(c) is equivalent to that of Fig. 17 in [5]. The two approaches B. Operation as an N-Path Filter
are identical with respect to voff (t) and path ⃝. a The flow graph
of [5], however, differs from that in Fig. 7(c) with respect to When the network of Fig. 1 is used as an N -path filter, the
the output of path ⃝.b In that reference, vi excites the filter with output is taken at vx . Clearly, vx (t) is simply the sum of the von
impulse response (1/RC)h(t). The filter’s output is multiplied waveforms of the individual kernels. The signal-flow diagram that
by w(t), and an appropriate waveform is added to account for yields von (t) can be obtained from that in Fig. 7(c) by eliminating
the path that yields voff and simplifying the remainder of the
†v
off (t), which is non-zero only during (k + 1)Ts < t < (k + 2)Ts − τ ,
diagram. The result is shown in Fig. 10. When vi = ej2πf t ,
models that part of the waveform. the component of von (t) at the frequency f is comprised of
 ∑
vk δ(t − kTs )
k
∫
voff (t) contributions from paths ⃝
2 and ⃝
3 as follows.
Heq (f ) RC
⃝
1 Path ⃝
2 →− (1 − e−j2πf (Ts −τ ) ) (16)
Ts 1 + j2πf RC
delay | {z } | {z }
Ts − τ F (heq (t)) F (h(t))
and sampling
(a) ∑
h(t)
⃝
2 τ 1
Path ⃝
3 → · . (17)
delay Ts 1 + j2πf RC
w(t) |{z} | {z }
Ts − τ dc F (h(t)/RC)
component
vi 1 of w(t)
RC h(t)
⃝
3
Thus, X0 (f ), the zeroth-order harmonic transfer function of the
∑
vk δ(t − kTs ) N -path filter, given by N · Vx (f )/ej2πf t , and using N τ = Ts , is
k
∫ seen to be
[ { }]
1 RC
X0 (f ) = 1− Heq (f ) 1 − e−j2πf (Ts −τ ) .
(1 + j2πf RC) τ
delay ∑ (18)
(b) h(t) Ts − τ
Fig. 11 compares the magnitudes of simulated and analytical

vi 1
RC h(t)
0
Analytical
∑ w(t) -10 Spectre
vk δ(t − kTs )
k -20
∫
1 |X0 (f )| (dB)
RC h(t)
-30
delay ∑ -40
(c) Ts − τ
-50
vi 1
RC h(t) -60
0 1 2 3 4 5 6

w(t) f /fs
∑
vk δ(t − kTs ) Fig. 11. Simulated and analytical X0 (f ) for N = 4, R = 50 Ω, C = 50 pF
k
∫ and fs = 1 GHz.
1
RC h(t)
X0 (f ) for N = 4, R = 50 Ω, C = 50 pF and fs = 1 GHz.
delay ∑
(d)
Ts − τ
For input frequencies around fs , and RC ≫ Ts , X0 (f )
can be simplified, using 2πf RC ≈ 2πfs RC ≫ 1 and τ /Ts =
vi 1 1/N ≡ D, as follows† .
RC h(t)
Heq (f ) ( j2πfs τ )
w(t) X0 (f ) ≈ e − 1 = Heq (f )ejπD sinc (D) . (19)
j2πfs τ
As seen from the discussion above, the transfer function of the
∑
vk δ(t − kTs ) ZOH 1 equivalent LTI filter has a central role to play in our analysis of
RC h(t)
k
Ts − τ
the N -path filter output.
(e)
vi C. Input Admittance:
w(t) Let the N -path circuit be driven by a voltage vi (t) =
Fig. 9. Connection with Soer’s model. ej2π(fs +∆f )t , where ∆f ≪ fs . The fundamental component of
∑ the current drawn from the source for input frequencies around
δ(t − kTs ) fs is simply the component of the voltage across the resistor (at
k
the input frequency) divided by R, and is given by
vi heq (t) h(t)
⃝
2 ej2π(fs +∆f )t
delay ∑ (1 − X0 (f )) .
I(f ) = (20)
w(t) von (t) R
Ts − τ
For RC ≫ τ and f ≈ fs , (12) can be simplified as
1
RC h(t)
⃝
3 sinc(D)e−jπD
Heq (f ) ≈ . (21)
(1 + j2π∆f N RC)
Fig. 10. Model of the switch-RC kernel needed to determine von (t), which is
relevant when operating the structure of Fig. 1 as an N -path filter. † In this paper, we use the normalized sinc function, i.e., sinc(x) =
sin(πx)
πx
Using this in (19), we have where H(f ) = RC/(1+j2πf RC). To determine H4 , we denote
( ) the input tone by (−4fs + f ) and proceed in a similar fashion.
ej2π(fs +∆f )t sinc2 (D) We thus have
I(f ) ≈ 1− (22)
R 1 + j2π RCD ∆f
H−4 (4fs + f ) = Heq (4fs + f ) · H(f )(1 − e−j2πf (Ts −τ ) )
The admittance, therefore, is
( ) H4 (−4fs + f ) = Heq (−4fs + f ) · H(f )(1 − e−j2πf (Ts −τ ) )
1 sinc2 (D)
Y (j2π(f + ∆f )) ≈ 1− (23) The above expressions are in excellent agreement with simula-
R 1 + j2π RCD ∆f tions, as shown in Fig. 13. From the discussion above, we see
In continued fraction form, the above expression can be written
0
as
1 Analytical
H0 (f )
Y = (24) Spectre
-10
1
R + 1−sinc2 (D) 2π∆f C
+ j D·sinc

|H|
R·sinc2 (D) 2 (D) -20 H4 (−4fs + f )

This is readily identified with the network shown in Fig. 12,

in agreement with [6], [8]† . The controlled source is necessary -30
H−4 (4fs + f )
due to the following. The input has a frequency (fs + ∆f ). The
-40
current flowing through C2 in the network of Fig. 12 would be 0.8 0.85 0.9 0.95 1 1.05 1.1 1.15 1.2
proportional to (fs + ∆f ). However, as seen from the discussion
f /fs
above, the current is seen to be dependent only on ∆f . The
Fig. 13. Analytical and simulated |H0 (f )|, H−4 (4fs + f ), H4 (−4fs + f ) for
controlled source, therefore, is necessary to remove the current a 4-path filter: R = 50 Ω, C = 50 pF, fs = 1 GHz.
component at fs .
that the folding of signals from higher frequencies into the desired
R band is largely controlled by the shape of Heq , (again) placing
vo in evidence the important role played by the voltages sampled on
the capacitors.
vi R2 C2 j2πfs C2 vo

E. Simplified Model in the Passive-Mixer Mode

2
sinc (D) C (a)
R2 = R 1−sinc2 (D) C2 = Dsinc2 (D)
(Ts − τ )
Fig. 12. Approximate LTI network for frequencies close to f = fs . ∑ voff (t)
vk δ(t − kTs ) ZOH vo (t)
k

τ ⃝
a
(Ts − τ )
D. High-order Harmonic Transfer Functions: delay ZOH
Higher-order harmonic transfer functions can be similarly
determined. Thanks to N -path operation, only the N th order (b)
Ts
HTFs and their multiples will be non-zero. Referring to Fig. 10,
every N th coefficient of the Fourier expansion of w(t) is zero, vi (t) heq (t) ZOH vo (t)
since w(t) has a width Ts /N . As a result, path ⃝ 3 does not ∑
δ(t − kTs )
contribute to H±4 . We now determine H−4 . We assume that the k
input frequency is expressed as 4fs + f . This way of denoting Fig. 14. (a) Simplified model in the passive-mixer mode and (b) Final model
for the passive mixer.
the input frequency is convenient, since the output frequency
associated with H−4 is then (4fs + f ) − 4fs = f. vi =
In the passive-mixer (P-M) mode, fin ≈ fs +∆f , RC ≫ Ts
ej2π(f +4fs )t is filtered by the LTI filter with frequency response
and fin ≫ 1/RC. The output voltage of Fig. 7(c) can then be
Heq , before being sampled by the impulse train. The (complex)
simplified by recognizing the following. The output of path ⃝ b
sinusoid at the sampler’s input is Heq (f + 4fs )ej2π(f +4fs )t .
is very small, since the input frequency is very large in relation
Sampling translates this sinusoid to all frequencies shifted by
to the bandwidth of the RC filter. Further, h(t) is almost a unit-
integer multiples of fs . Since we are interested in the output
step function, since RC ≫ Ts , and β ≈ 1. As a result, h(t)
frequency f , the output of the sampler at frequency f is of
and the subsequent block in path ⃝ a can be approximated by a
interest. Clearly, this is (1/Ts )Heq (f + 4fs )ej2πf t . Denoting the
ZOH, with initial delay of (Ts − τ ) and width τ , as shown in
Fourier transform of h(t) by H(f ), the output of path ⃝ 2 at f is
Fig. 14(a). This is readily simplified as a single ZOH with width
seen to be (1/Ts )Heq (f + 4fs )H(f )(1 − e−j2πf (Ts −τ ) )ej2πf t ,
Ts . Thus, as concluded in [5], the performance in the P-M mode
† Unlike in a LTI network, we note that impedance and admittance are not is dependent mostly on the sampled voltage across the capacitor.
reciprocals of each other. Note that such a simplification is not immediately apparent from
Soer’s model, or from the model of Fig. 9(a). Fig. 14(c) shows When the capacitor is periodically switched, as in Fig. 15(b),
the simplified model of the switched-RC circuit in the P-M mode, the mean square noise across the capacitor is given by
including the prefilter. ∫ ∞
Let the switched-RC circuit be operating in the P-M mode, vn2 = 2kT R |heq (t)|2 dt. (28)
0
with vi (t) = ej2π(mfs +∆f )t , where m is an integer and ∆f ≪ fs .
1/(2πN RC) should be chosen so that ∆f , which is the desired In Section III, we determined heq (t) by injecting a current
baseband tone, suffers very little attenuation. impulse into the capacitor of the adjoint network, and observing
the current in the resistor. There, we saw that the capacitor voltage
Referring to Fig. 14(b), the output of the prefilter is
decays exponentially with a time-constant RC when the switch
Heq (mfs + ∆f )ej2π(nfs +∆f )t . Due to impulse sampling and
is closed, and remains unchanged when the switch is open. As a
the subsequent hold operation, the desired output tone has an
result, the resistor current, which is heq (t), decays exponentially
amplitude Heq (mfs + ∆f ). In a down-conversion mixer m is
for 0 ≤ t < τ . At t = Ts , when the switch is closed again,
typically chosen to be unity. Using (12), it is easy to show that
heq (t) begins from “where it left off”, and decays again. One
the ratio of the conversion gain from an undesired frequency
can therefore think of h2eq (t) as being the result of cutting up
(mfs + ∆f ), m ̸= 1 to that from the desired input at frequency
the squared impulse response of Fig. 15(a) into slivers of width
(fs + ∆f ) can be approximated as
    τ , and spacing them Ts apart. Thus, the area under the squared
 Heq (mfs + ∆f )   sinc(mD)  impulse response h2eq (t) is exactly the same as that in Fig. 15(a).
 ≈ 
 Heq (fs + ∆f )   sinc(D)  . (25) Therefore, the mean square noise across the capacitor in the
switch-RC kernel is (not surprisingly) kT /C, irrespective of the
V. K ERNEL N OISE sampling or mixing mode of operation and independent of the
duty cycle. This result is identical to that in [4] (and [6])– where it
was derived using frequency domain methods. It is thus seen that
1 2
(RC)2 h (t) working in the time-domain yields insights that are less apparent
1 R in the frequency domain.
(RC)2

2kT R C
Ts
x y
(a) ∫ ∞ 2kT R heq (t) ZOH Svo (f )
1 kT
vn2 = 2kT R h2 (t) dt =
0 (RC)2 C ∑
a1 a2 a3 a4 a5 a6 a7
δ(t − kTs )
k
t
Fig. 16. Determining the output noise spectral density in the passive-mixer
mode.
R ϕ1
Next, we determine the noise spectral density of the switch-
2kT R C
h2eq (t) RC kernel in the passive-mixer mode. The signal flow diagram is
1
(RC)2
∫ ∞
kT
shown in Fig. 16. The double-sided noise spectral density due to
vn2 = 2kT R h2eq (t) dt =
0 C the resistor is 2kT R. The resistor noise is filtered by an LTI filter
with impulse response heq (t) and sampled, before exciting a ZOH
(b) τ
(which holds for a duration Ts ). The autocorrelation function of
the noise at x is given by
Rxx (t) = 2kT R · heq (t) ∗ heq (−t)
a1 a2 a3 a4 a5 a6 a7

t (29)

Fig. 15. (a) Mean square noise of the capacitor voltage, which is proportional
where ∗ denotes convolution. Due to sampling, the autocorrela-
to the integral of the square of the impulse response, is kT /C. (b) tion function of the noise at y is therefore
∑∞
Ryy (t) = Rxx (lTs )δ(t − lTs ). (30)
In this section, we address kernel noise. First consider the
l=−∞
LTI network of Fig. 15(a). The impulse response corresponding to
the transfer function from the noise source (with a double-sided The noise spectral density of the output waveform, denoted by
voltage noise spectral density of 2kT R) to the capacitor voltage Svo (f ), is thus given by
is (1/RC)e−t/RC u(t) = (1/RC)h(t). The mean-square noise ( ) ZOH
1 ∑
∞ z }| {
across the capacitor can be obtained using Parseval’s theorem as Svo (f ) = Rxx (lTs )e−j2πf lTs Ts2 sinc2 (f Ts ) .
follows. Ts
l=−∞
∫ ∞ | {z }
vn2 = 2kT R |H(f )|2 df (26) Syy (f )=F (Ryy (t))
−∞ (31)
∫ ∞ In a down-conversion mixer, the spectral density around f = 0
1 kT
= 2kT R 2
|h(t)|2 dt = . (27) is of interest. This is given by
0 (RC) C
∞
∑
Clearly, the mean square noise is proportional to the area under Svo (0) = Ts Rxx (lTs ) (32)
the squared impulse response. l=−∞
We therefore need to compute the sum of the samples of Rxx (t), [7] C. Andrews and A. C. Molnar, “A passive mixer-first receiver with digitally
sampled at fs . Rxx (0) is simply the mean-square noise across the controlled and widely tunable RF interface,” IEEE Journal of solid-state
circuits, vol. 45, no. 12, pp. 2696–2708, 2010.
capacitor, and as we saw earlier in this section, is given by kT /C. [8] ——, “Implications of passive mixer transparency for impedance matching
Rxx (Ts ) is proportional to the area under heq (t) · heq (t − Ts ): and noise figure in passive mixer-first receivers,” IEEE Transactions on
observing Fig. 4(c), it is easy to see that Rxx (Ts ) = βRxx (0). Circuits and Systems I: Regular Papers, vol. 57, no. 12, pp. 3092–3103,
2010.
Proceeding similarly, we can write [9] A. Molnar and C. Andrews, “Impedance, filtering and noise in N -phase
kT |l| passive CMOS mixers,” in Proceedings of the Custom Integrated Circuits
Rxx (lTs ) = β (33) Conference (CICC). IEEE, 2012, pp. 1–8.
C [10] C. Andrews, C. Lee, and A. Molnar, “Effects of LO harmonics and overlap
Thus, shunting on N-phase passive mixer based receivers,” in Proceedings of the
∞
∑ ( ) ESSCIRC (ESSCIRC). IEEE, 2012, pp. 117–120.
kT 1+β [11] D. Yang, C. Andrews, and A. Molnar, “Optimized design of N-phase passive
Rxx (lTs ) = . (34)
C 1−β mixer-first receivers in wideband operation,” IEEE Transactions on Circuits
l=−∞
and Systems I: Regular Papers, vol. 62, no. 11, pp. 2759–2770, 2015.
In the passive-mixer mode, β = e−τ /RC ≈ 1 − (τ /RC). Using [12] A. Mirzaei, H. Darabi, J. C. Leete, and Y. Chang, “Analysis and optimization
of direct-conversion receivers with 25% duty-cycle current-driven passive
the equation above in (32), we obtain mixers,” IEEE Transactions on Circuits and Systems I: Regular Papers,
vol. 57, no. 9, pp. 2353–2366, 2010.
Svo (0) ≈ 2kT RN (35) [13] A. Mirzaei and H. Darabi, “Analysis of imperfections on performance of 4-
phase passive-mixer-based high-Q bandpass filters in SAW-less receivers,”
which is identical to the result obtained in eq.52 of [4] and IEEE Transactions on Circuits and Systems I: Regular Papers, vol. 58, no. 5,
eq.21 of [6] using frequency-domain methods. Thanks to the pp. 879–892, 2011.
[14] T. Strom and S. Signell, “Analysis of periodically switched linear circuits,”
adjoint impulse-response method, the determination of noise is IEEE Transactions on Circuits and Systems, vol. 24, no. 10, pp. 531–541,
considerably less involved than that in [4], [6]. 1977.
[15] A. Ghaffari, E. A. Klumperink, M. C. Soer, and B. Nauta, “Tunable high-
Q N-path band-pass filters: Modeling and verification,” IEEE Journal of
VI. C ONCLUSIONS Solid-State Circuits, vol. 46, no. 5, pp. 998–1010, 2011.
The operation and properties of switched-RC N -path filters, [16] A. Ghaffari, E. A. Klumperink, and B. Nauta, “Tunable N-path notch filters
for blocker suppression: Modeling and verification,” IEEE Journal of Solid-
samplers and passive mixers are intimately tied to the sampled State Circuits, vol. 48, no. 6, pp. 1370–1382, 2013.
voltages across the capacitors. In this work we exploited two [17] Z. Lin, P. I. Mak, and R. P. Martins, “Analysis and modeling of a gain-
key properties of sampled LPTV networks that greatly simplify boosted N-path switched-capacitor bandpass filter,” IEEE Transactions on
Circuits and Systems I: Regular Papers, vol. 61, no. 9, pp. 2560–2568, 2014.
the determination of the sampled capacitor voltages: first, that [18] L. Duipmans, R. E. Struiksma, E. A. Klumperink, B. Nauta, and F. E. van
the sampled output of an LPTV network (when the sampling Vliet, “Analysis of the signal transfer and folding in n-path filters with a
rate is the same as that at which the network is varying) can series inductance,” IEEE Transactions on Circuits and Systems I: Regular
Papers, vol. 62, no. 1, pp. 263–272, 2015.
be thought of as the sampled output of an appropriately chosen [19] S. Pavan and R. S. Rajan, “Interreciprocity in linear periodically time-
linear time-invariant system. Next, the adjoint network can be varying networks with sampled outputs,” IEEE Transactions on Circuits
used to determine the impulse response of the equivalent LTI and Systems II: Express Briefs, vol. 61, no. 9, pp. 686–690, 2014.
[20] J. Vandewalle, H. De Man, and J. Rabaey, “The adjoint switched capacitor
system in a simple manner. We derived a model for the complete network and its application to frequency, noise and sensitivity analysis,”
capacitor voltage in a switched-RC kernel. It turned out that International journal of circuit theory and applications, vol. 9, no. 1, pp.
this model was very convenient to understand the operation in 77–88, 1981.
[21] F. Yuan and A. Opal, “Adjoint network of periodically switched linear
the passive-mixer, sampler and N -path filter modes. Finally, we circuits with applications to noise analysis,” IEEE Transactions on Circuits
showed that noise analysis of the kernel, usually accomplished and Systems I: Fundamental Theory and Applications, vol. 48, no. 2, pp.
in the frequency domain, can also benefit from the simplicity of 139–151, 2001.
[22] R. Rohrer, L. Nagel, R. Meyer, and L. Weber, “Computationally efficient
the adjoint-network approach. electronic-circuit noise calculations,” IEEE Journal of Solid-State Circuits,
vol. 6, no. 4, pp. 204–213, 1971.
R EFERENCES [23] S. Pavan and R. S. Rajan, “Simplified analysis and simulation of the STF,
NTF, and noise in continuous-time ∆Σ modulators,” IEEE Transactions on
[1] A. A. Abidi, “The path to the software-defined radio receiver,” IEEE Journal Circuits and Systems II: Express Briefs, vol. 61, no. 9, pp. 681–685, 2014.
of Solid-State Circuits, vol. 42, no. 5, pp. 954–966, 2007. [24] A. Sukumaran and S. Pavan, “Design of continuous-time ∆Σ modulators
[2] H. Darabi, A. Mirzaei, and M. Mikhemar, “Highly integrated and tunable with dual switched-capacitor Return-to-Zero DACs,” IEEE Journal of Solid-
RF front ends for reconfigurable multiband transceivers: A tutorial,” IEEE State Circuits, vol. 51, no. 7, pp. 1619–1629, 2016.
Transactions on Circuits and Systems I: Regular Papers, vol. 58, no. 9, pp. [25] S. Pavan, R. Schreier, and G. C. Temes, Understanding Delta-Sigma Data
2038–2050, 2011. Converters, Second Edition. John Wiley & Sons, 2016.
[3] L. Franks and I. Sandberg, “An alternative approach to the realization of [26] L. Franks and I. Sandberg, “An alternative approach to the realization of
network transfer functions: The n-path filter,” Bell Labs Technical Journal, network transfer functions: The N -path filter,” Bell Labs Technical Journal,
vol. 39, no. 5, pp. 1321–1350, 1960. vol. 39, no. 5, pp. 1321–1350, 1960.
[4] M. C. Soer, E. A. Klumperink, P.-T. De Boer, F. E. Van Vliet, and B. Nauta,
“Unified frequency-domain analysis of switched-series-RC passive mixers
and samplers,” IEEE Transactions on Circuits and Systems I: Regular
Papers, vol. 57, no. 10, pp. 2618–2631, 2010.
[5] T. Iizuka and A. A. Abidi, “FET-RC Circuits: A unified treatment - Part
I: Signal transfer characteristics of a single-path,” IEEE Transactions on
Circuits and Systems I: Regular Papers, vol. 63, no. 9, pp. 1325–1336,
2016.
[6] ——, “FET-RC Circuits: A unified treatment - Part II: Extension to multi-
paths, noise figure, and driving-point impedance,” IEEE Transactions on
Circuits and Systems I: Regular Papers, vol. 63, no. 9, pp. 1337–1348,
2016.
 Shanthi Pavan obtained the B.Tech degree in Electron-
ics and Communication Engg from the Indian Institute
of Technology, Madras in 1995 and the M.S and Sc.D
degrees from Columbia University, New York in 1997
and 1999 respectively. From 1997 to 2000, he was with
Texas Instruments in Warren, New Jersey, where he
worked on high speed analog filters and data converters.
From 2000 to June 2002, he worked on microwave
ICs for data communication at Bigbear Networks in
Sunnyvale, California. Since July 2002, he has been with
the Indian Institute of Technology-Madras, where he is
now a Professor of Electrical Engineering. His research interests are in the areas
of high speed analog circuit design and signal processing.
Dr. Pavan is the recipient of the IEEE Circuits and Systems Society Darlington
Best Paper Award (2009), the Shanti Swarup Bhatnagar Award (2012) and the
Swarnajayanthi Fellowship (2009) (from the Government of India), the Mid-
career Research Excellence Award and the Young Faculty Recognition Award
from IIT Madras (for excellence in teaching), the Technomentor Award from the
India Semiconductor Association and the Young Engineer Award from the Indian
National Academy of Engineering (2006). He is the author of Understanding
Delta-Sigma Data Converters (second edition), with Richard Schreier and Gabor
Temes). Dr. Pavan has served as the Editor-in-Chief of the IEEE Transactions on
Circuits and Systems: Regular Papers, and on the editorial boards of both parts
of the IEEE Transactions on Circuits and Systems. He has been on the technical
program committee of the International Solid State Circuits Conference, and a
Distinguished Lecturer of the Solid-State Circuits Society. He is a fellow of the
Indian National Academy of Engineering.

Eric Klumperink was born on April 4th, 1960, in

Lichtenvoorde, The Netherlands. He received the B.Sc.
degree from HTS, Enschede (1982), worked in industry
on digital hardware and software, and then joined the
University of Twente in 1984, shifting focus to ana-
log CMOS circuit research. This resulted in several
publications and his Ph.D. thesis “Transconductance
Based CMOS Circuits” (1997). In 1998, Eric started
as Assistant Professor at the IC-Design Laboratory in
Twente and shifted research focus to RF CMOS circuits
(e.g. sabbatical at the Ruhr Universitaet in Bochum,
Germany). Since 2006, he is an Associate Professor, teaching Analog & RF
IC Electronics, and guiding PhD and MSc projects related to RF CMOS circuit
design with focus on software-defined radio, cognitive radio, and beamforming.
He served as an Associate Editor for the IEEE Transactions on Circuits and
Systems: Express Briefs (2006-2007), IEEE Transactions on Circuits and Systems:
Regular Papers (2008-2009) and the IEEE Journal of Solid State Circuits(2010-
2014), as IEEE Solid-State Circuit Society Distinguished Lecturer (2014/2015),
and as member of the technical program committees of ISSCC (2011-2016) and
the IEEE RFIC Symposium (2011-..). He holds several patents, authored and co-
authored more than 150 internationally refereed journal and conference papers,
and was recognized as 20+ ISSCC paper contributor over 1954-2013. He is a
co-recipient of the ISSCC 2002 and the ISSCC 2009 “Van Vessem Outstanding
Paper Award”.