AIC Ch12 PDF
AIC Ch12 PDF
AIC Ch12 PDF
12.1
General Considerations
In order to understand the motivation for sampled-data circuits, let us rst consider the simple continuous-time amplier shown in Fig. 12.1(a). Used extensively with bipolar op amps, this circuit presents a difcult issue if implemented in CMOS technology. Recall that, to achieve a high voltage gain, the open-loop output resistance of CMOS op amps is maximized, typically approaching hundreds of kilo-ohms. We therefore suspect that gain, degrading the precision of the circuit. In fact, with the aid of the simple equivalent circuit
395
396
R2 V in R1 Vout V in R1 VX
(a)
R2 Vout
A vV X (b)
R out
Figure 12.1. (a) Continuous-time feedback amplier, (b) equivalent circuit of (a).
12
:1
and hence
Rout A , v Vout = , R2 R2 12:2 Vin R1 1 + Rout + Av + R2 : R1 R1 Equation (12.2) implies that, compared to the case where Rout = 0, the closed-loop gain suffers
from inaccuracies in both the numerator and the denominator. Also, the input resistance of the amplier, approximately equal to R1, loads the preceding stage while introducing thermal noise. In the circuit of Fig. 12.1(a), the closed-loop gain is set by the ratio of R2 and R1. In order to
avoid reducing the open-loop gain of the op amp, we postulate that the resistors can be replaced by capacitors [Fig. 12.2(a)]. But, how is the bias voltage at node X set? We may add a large feedback
RF C2 C1 V in X Vout C1 V in X Vout C2
(a)
(a)
Figure 12.2.
(a) Continuous-time feedback amplier using capacitors, (b) use of resistor to dene bias point.
resistor as in Fig. 12.2(b), providing dc feedback while negligibly affecting the ac behavior of the amplier in the frequency band of interest. Such an arrangement is indeed practical if the circuit senses only high-frequency signals. But suppose, for example, the circuit is to amplify a voltage
397
V in
Vout t
step. Illustrated in Fig. 12.3, the response contains a step change due to the initial amplication by the circuit consisting of C1; C2 , and the op amp, followed by a tail resulting from the loss of charge on given by 1 Vout s , RF C2 s 1 Vin RF + C1 s C1s 2 R C s F 1 = , RF C2 s + 1 ; indicating that Vout =Vin
C2 through RF .
From another point of view, the circuit may not be suited to amplify
wideband signals because it exhibits a high-pass transfer function. In fact, the transfer function is
(12.3)
(12.4)
RF C2 ,1 . The above difculty can be remedied by increasing RF C2 , but in many applications the required
values of the two components become prohibitively large. We must therefore seek other methods of establishing the bias while utilizing capacitive feedback networks. Let us now consider the switched-capacitor circuit depicted in Fig. 12.4, where three switches control the operation:
S2 C2 V in S1 C1 Vout S3
S1 and S3 connect the left plate of C1 to Vin and ground, respectively, and
Figure 12.4.
Switched-capacitor amplier.
and study the circuit in two phases. First, S1 and S2 are on and S3 is off, yielding the equivalent
398
(a)
(b)
V in VA V in0
Vout t0
(c)
V in0 t
C1 C2
Figure 12.5. Circuit of Fig. 12.4 in (a) sampling mode, (b) amplication mode. circuit of Fig. 12.5(a). For a high-gain op amp, VB
Vout 0, and hence the voltage across C1 is approximately equal to Vin . Next, at t = t0 , S1 and S2 turn off and S3 turns on, pulling node A to ground. Since VA changes from Vin0 to 0, the output voltage must change from zero to Vin0 C1 =C2 .
=
through
C2 drives the op amp input differential voltage and hence the voltage across C1 to zero (Fig. 12.6). The charge stored on C1 at t = t0 must then be transferred to C2 , producing an output
C2 C1 Q in0
0
The output voltage change can also be calculated by examining the transfer of charge. Note
= 0,
C2 C1 Vout
0
Q in0 Vout
Figure 12.6.
voltage equal to Vin0 C1 =C2 . Thus, the circuit amplies Vin0 by a factor of C1 =C2 . Several attributes of the circuit of Fig. 12.4 distinguish it from continuous-time implementations. First, the circuit devotes some time to sample the input, setting the output to zero and providing no amplication during this period. Second, after sampling, for t
the input voltage Vin , amplifying the sampled voltage. Third, the circuit conguration changes
399
considerably from one phase to another, as seen in Fig. 12.5(a) and (b), raising concern about its stability. sampling capability, we note from the waveforms depicted in Fig. 12.5 that after Vout settles, the What is the advantage of the amplier of Fig. 12.4 over that in Fig. 12.1? In addition to
current through C2 approaches zero. That is, the feedback capacitor does not reduce the open-loop gain of the amplier if the output voltage is given enough time to settle. In Fig. 12.1, on the other hand, R2 continuously loads the amplier. The switched-capacitor amplier of Fig. 12.4 lends itself to implementation in CMOS technology much more easily than in other technologies. This is because discrete-time operations require switches to perform sampling as well as a high input impedance to sense the stored quantities with no corruption. For example, if the op amp of Fig. 12.4 incorporates bipolar transistors at its input, the base current drawn from the inverting input in the amplication phase [Fig. 12.5(b)] creates an error in the output voltage. The existence of simple switches and a high input impedance have made CMOS technology the dominant choice for sampled-data applications. The foregoing discussion leads to the conceptual view illustrated in Fig. 12.7 for switchedcapacitor ampliers. In the simplest case, the operation takes place in two phases: sampling and
V in Vout
CK
Sample Amplify
Figure 12.7. General view of switched-capacitor amplier. amplication. Thus, in addition to the analog input, Vin , the circuit requires a clock to dene each phase. Our study of SC ampliers proceeds according to these two phases. First, we analyze various sampling techniques. Second, we consider SC amplier topologies.
400
12.2
Sampling Switches
12.2.1 MOSFETS as Switches A simple sampling circuit consists of a switch and a capacitor [Fig. 12.8(a)]. A MOS transistor can serve as a switch [Fig. 12.8(b)] because (a) it can be on while carrying zero current, and (b) its
CK V in CH
(a) (b)
Vout
V in
M1 CH
Vout
Figure 12.8. (a) Simple sampling circuit, (b) implementation of the switch by a MOS device.
source and drain voltages are not pinned to the gate voltage, i.e., if the gate voltage varies, the source or drain voltage need not follow that variation. By contrast, bipolar transistors lack both of these properties, typically necessitating complex circuits to perform sampling. To understand how the circuit of Fig. 12.8(b) samples the input, rst consider the simple cases depicted in Fig. 12.9, where the gate command,
CK V in = 0 M1 Vout I D1 V DD CH CK
0
CK ,
goes high at
t0 .
In Fig.
V DD
V DD Vout t0
(a)
CK V in = +1 V I D1 M1 Vout
0
V DD CK
0 +1 V
CH
Vout
(b)
t0
Figure 12.9. Response of a sampling circuit to different input levels and initial conditions. 12.9(a), we assume that Vin = 0 for t t0 and the capacitor has an initial voltage equal to VDD . Thus, at t = t0 , M1 senses a gate-source voltage equal to VDD while its drain voltage is also equal to VDD . The transistor therefore operates in saturation, drawing a current of ID1
=
401
n Cox=2W=LVDD , VTH 2 from the capacitor. As Vout falls, at some point Vout = VDD , VTH , driving M1 into the triode region. The device nevertheless continues to discharge CH until Vout approaches zero. We note that for Vout 2VDD , VTH , the transistor can be viewed as a resistor equal to Ron = n Cox W=LVDD , VTH ,1. Now consider the case in Fig. 12.9(b), where Vin = +1 V, Vout t = t0 = 0 V, and VDD = 3 V. Here, the terminal of M1 connected to CH acts as the source, and the transistor turns on with VGS = +3 V, but VDS = +1 V. Thus, M1 operates in the triode region, charging CH until Vout approaches +1 V. For Vout +1 V, M1 exhibits an on-resistance of Ron = n Cox W=LVDD , Vin , VTH ,1 .
The above observations reveal two important points. First, a MOS switch can conduct current in either direction simply by exchanging the role of its source and drain terminals. Second, as shown
High
in Fig. 12.10, when the switch is on, Vout follows Vin and when the switch is off, Vout remains
V in CH
Vout CH
(a) Low
V in CH
Vout CH
(b)
Figure 12.10. Track and hold capabilities of a sampling circuit. constant. Thus, the circuit tracks the signal when CK is high and freezes the instantaneous value of Vin across CH when CK goes low. Example 12.1
In the circuit of Fig. 12.9(a), calculate Vout as a function of time. Assume = 0. Solution. Before Vout drops below VDD
Vout t
After
= =
D1 t VDD , IC H
t1 =
12
:7
402
,ID1
(12.8) (12.9)
12
:10
Vout
Thus,
2VDD
, VTH , Vout
ox W ,n C C L dt: H
12
:11
12
:12 :13
Cox W t , t : Vout = ,VDD , VTH n 1 2VDD , VTH , Vout CH L Taking the exponential of both sides and solving for Vout , we obtain
ln
12
Vout =
12
:14
VDD .
V DD CK
0
CH
VDD V TH Vout t
Figure 12.11.
403
out CH dV dt
= =
ID 1
12
:17 :18
12
Vout = VDD , VTH , 1 Cox W 1 : 12:19 1 t+ V ,V n 2 CH L DD TH Equation (12.19) implies that as t ! 1, Vout ! VDD , VTH . This is because as Vout approaches VDD , VTH , the overdrive voltage of M1 vanishes, reducing the current available for charging CH to negligible values. Of course, even for Vout = VDD , VTH , the transistor conducts some subthreshold current and, given enough time, eventually brings Vout to VDD . Nonetheless, as mentioned in Chapter 3, for typical operation speeds, it is reasonable to assume that Vout does not exceed VDD , VTH .
level is close to VDD , then the output provided by an NMOS switch cannot track the input. From another point of view, the on-resistance of the switch increases considerably as the input and output voltages approach VDD , VTH . We may then ask: what is the maximum input level that the switch can pass to the output faithfully? In Fig. 12.11, for Vout The foregoing analysis demonstrates a serious limitation of MOS switches: if the input signal
Vin , the transistor must operate in deep triode region and hence the upper bound of Vin equals VDD , VTH . As explained below, in practice
In the circuit of Fig. 12.12, calculate the minimum and maximum on-resistance of M1 . Assume
n Cox = 50 A/V2, W=L = 10=1; VTH = 0:7 V; VDD = 3 V, and = 0. Solution. We note that in the steady state, M1 remains in the triode region because the gate voltage is higher than both Vin and Vout by a value greater than VTH . If fin = 10 MHz, we predict
404
+3 V +0.5 V
M1
1 pF
+0.5 V
CH t
f in = 10 MHz
Figure 12.12. that Vout tracks Vin with a negligible phase shift due to the on-resistance of M1 and CH . Assuming
Vout Vin , we need not distinguish between the source and drain terminals, obtaining Ron1 =
Thus,
Ron1;max 1:11 k and Ron1;min 870 . raised to 1.5 V, then Ron1;max = 2:5 k.
12
:20
MOS devices operating in deep triode region are sometimes called zero-offset switches to emphasize that they exhibit no dc shift between the input and output voltages of the simple sampling circuit of Fig. 12.8(b).1 This is evident from examples of Fig. 12.9, where the output eventually becomes equal to the input. Nonexistent in bipolar technology, the zero offset property proves crucial in precise sampling of analog signals. We have thus far considered only NMOS switches. The reader can verify that the foregoing principles apply to PMOS switches as well. In particular, as shown in Fig. 12.13, a PMOS transistor fails to operate as a zero-offset switch if its gate is grounded and its drain terminal senses an input
V DD V in = 0 M1 V DD CK CH CK Vout
0
V DD Vout V THP t
Figure 12.13. Sampling circuit using PMOS switch. voltage of jVTHP j or less. In other words, the on-resistance of the device rises rapidly as the input
1
405
12.2.2 Speed Considerations What determines the speed of the sampling circuits of Fig. 12.8? We must rst dene the speed here. Illustrated in Fig. 12.14, a simple, but versatile measure of speed is the time required for
CK V in = V in0 M1 Vout
0
V DD CK
0 V
V in0 Vout
0
CH
tS
Vout would take innite time to become equal to Vin0 , we consider the output settled when it is within a certain error band, V; around the nal value. For example, we say the output settles to 0.1% accuracy after tS seconds, meaning that in Fig. 12.14, V=Vin0 = 0:1%. Thus, the speed specication must be accompanied by an accuracy specication as well. Note that after t = tS , we
can consider the source and drain voltages to be approximately equal. From the circuit of Fig. 12.14, we surmise that the sampling speed is given by two factors: the on-resistance of the switch and the value of the sampling capacitor. Thus, to achieve a higher speed, a large aspect ratio and a small capacitor must be used. However, as illustrated in Fig. 12.12, the on-resistance also depends on the input level, yielding a greater time constant for more positive inputs (in the case of NMOS switches). From Eq. (12.20), we plot the on-resistance of the switch as a function of the input level [Fig. 12.15(a)], noting the sharp rise as Vin approaches
R on,N R on,P
the output voltage to go from zero to the maximum input level after the switch turns on. Since
0 (a)
VDD V TH
V in
0 V THP (b)
V in
Figure 12.15. On-resistance of (a) NMOS and (b) PMOS devices as a function of input voltage.
VDD , VTH . For example, if we restrict the variation of Ron to a range of 4 to 1, then the maximum
406
12
:21 :22
This value falls around VDD =2, translating to severe swing limitations. Note that the device threshold voltage directly limits the voltage swings.2 In order to accommodate greater voltage swings in a sampling circuit, we rst observe that a PMOS switch exhibits an on-resistance that decreases as the input voltage becomes more positive [Fig. 12.15(b)]. It is then plausible to employ complementary switches so as to allow rail-to-tail swings. Shown in Fig. 12.16(a), such a combination requires complementary clocks, producing
R on,P CK M1 V in M2 CK
(a)
12
R on,N
Vout CH V THP
(b)
R on,eq VDD V TH V in
Figure 12.16. (a) Complementary switch, (b) on-resistance of the complementary switch.
an equivalent resistance:
Ron;eq
= =
Ron;N jjRon;P
1
(12.23)
: W , C W V , C W (12.25) n Cox W N VDD , VTHN , n Cox V L L N p ox L P in p ox L P THP Interestingly, if n Cox W=LN = p Cox W=LP , then Ron;eq is independent of the input level.3 Fig. 12.16(b) plots the behavior of Ron;eq in the general case, revealing much less variation than
=
jj
(12.24)
407
For high-speed input signals, it is critical that the NMOS and PMOS switches in Fig. 12.16(a) turn off simultaneously so as to avoid ambiguity in the sampled value. If, for example, the NMOS device turns off t seconds earlier than the PMOS device, then the output voltage tends to track the
input for the remaining t seconds, but with a large, input-dependent time constant (Fig. 12.17). This effect gives rise to distortion in the sampled value. For moderate precision, the simple circuit
CK
t
CK V in
Vout
Figure 12.17.
shown in Fig. 12.18 provides complementary clocks by duplicating the delay of inverter I1 through
VDD G2 CK CK in CK I1
Figure 12.18. Simple circuit generating complementary clocks. the gate G2 . 12.2.3 Precision Considerations Our foregoing study of MOS switches indicates that a larger W=L or a smaller sampling capacitor results in a higher speed. In this section, we show that these methods of increasing the speed degrade the precision with which the signal is sampled. Three mechanisms in MOS transistor operation introduce error at the instant the switch turns off. We study each effect individually.
408
Channel Charge Injection Consider the sampling circuit of Fig. 12.19 and recall that for a MOSFET to be on, a channel must exist at the oxide-silicon interface. Assuming Vin
CK V in M1 Vout CH
Vout , we
Figure 12.19.
use our derivations in Chapter 2 to express the total charge in the inversion layer as
12
:26
source and drain terminals, a phenomenon called channel charge injection. error. On the other hand, the charge injected to the right side is deposited on CH , introducing an error in the voltage stored on the capacitor. For example, if half of resulting error equals V
=
The charge injected to the left side on Fig. 12.19 is absorbed by the input source, creating no
:27
Illustrated in Fig. 12.20, the error for an NMOS switch appears as a negative pedestal at the
Figure 12.20. Effect of charge injection. output. Note that the error is directly proportional to WLCox and inversely proportional to CH . half of the channel charge in injected onto CH ? In reality, the fraction of charge that exits through the source and drain terminals is a relatively complex function of various parameters such as the impedance seen at each terminal to ground and the transition time of the clock [1, 2]. Investigations An important question that arises now is: why did we assume in arriving at (12.27) that exactly
409
of this effect have not yielded any rule of thumb that can predict the charge splitting in terms of such parameters. Furthermore, in many cases, these parameters, e.g., the clock transition time, are poorly controlled. Also, most circuit simulation programs model charge injection quite inaccurately. As a worst-case estimate, we can assume that the entire channel charge is injected onto the sampling capacitor. How does charge injection affect the precision? Assuming all of the charge is deposited on the capacitor, we express the sampled output voltage as
12
:28 :29
12
suggesting that the output deviates from the ideal value through two effects: a non-unity gain equal
12.21). In
other words, since we have assumed channel charge is a linear function of the input voltage, the
V in
Figure 12.21. Input/output characteristic of sampling circuit in the presence of charge injection. In the foregoing discussion, we tacitly assumed that VTH is constant. However, for NMOS switches (in an n-well technology), body effect must be taken into account.5 Since VTH =
VTH 0 + Vout
4 5
p2
B + VBS ,
p2
B , and VBS
,Vin , we have
q
= =
ox Vin , WLC CH VDD , Vin , VTH 0 , 2 B + Vin , 2 B ; WLCox q2 + V , WLCox V , V , ox Vin 1 , WLC , B in CH CH CH DD TH 0
q
(12.30)
q
2 (12.31) B :
The voltage gain is greater than unity because the pedestal becomes smaller as the input level rises. Even for PMOS switches, the n-well is connected to the most positive supply voltage because the source and drain terminals of the switch may interchange during sampling.
410
It follows that the nonlinear dependence of VTH upon Vin introduces nonlinearity in the input/output characteristic. In summary, charge injection contributes three types of errors in MOS sampling circuits: gain error, dc offsets, and nonlinearity. In many applications, the rst two can be tolerated or corrected whereas the last cannot. It is instructive to consider the speed-precision trade-off resulting from charge injection. Representing the speed by a simple time constant injection, we dene a gure of merit as F
= = =
V ,1 . Writing 1
Ron CH
(12.32) (12.33)
and
WLCox V , V , V ; CH DD in TH
n F= L2 :
12
:34 :35
we have
12
Thus, to the rst order, the trade-off is independent of the switch width and the sampling capacitor. Clock Feedthrough In addition to channel charge injection, a MOS switch couples the clock transitions to the sampling capacitor through its gate-drain or gate-source overlap capacitance. Depicted in Fig. 12.22, the effect introduces an error in the sampled output voltage. Assuming the
V CK
0
V in
M1 CH
Vout
Figure 12.22.
where Cov is the overlap capacitance per unit width. The error V is independent of the input level, manifesting itself as a constant offset in the input/output characteristic. As with charge injection, clock feedthrough leads to a trade-off between speed and precision as well.
WCov ; VCK WC +C
ov H
12
:36
411
kT=C Noise
Recall from Example 7.1 that a resistor charging a capacitor gives rise to a total rms
q
noise voltage of
kT=C .
V in
Vin + Vn
Figure 12.23.
The on-resistance of the switch introduces thermal noise at the output and, when the switch turns off, this noise is stored on the capacitor along with the instantaneous value of the input voltage. It
q
can be proved that the rms voltage of the sampled noise in this case is still approximately equal to
The problem of KT=C noise limits the performance in many high-precision applications. In
order to achieve a low noise, the sampling capacitor must be sufciently large, thus loading other circuits and degrading the speed. 12.2.4 Charge Injection Cancellation The dependence of charge injection upon the input level and the trade-off expressed by (12.35) make it necessary to seek methods of cancelling the effect of charge injection so as to achieve a higher F . We consider a few such techniques here. To arrive at the rst technique, we postulate that the charge injected by the main transistor can be removed by means of a second transistor. As shown in Fig. 12.24, a dummy switch, driven by
CK is added to the circuit such that after M1 turns off and M2 turns on, the channel
CK CK
M2 ,
M1 V in
M2
q 1
q 2
CH
Vout
Figure 12.24. Addition of dummy device to reduce charge injection and clock feedthrough. charge deposited by the former on CH is absorbed by the latter to create a channel. Note that both the source and drain of M2 are connected to the output node.
412
How do we ensure that the charge injected by M1 , q1 , is equal to that absorbed by M2, q2 ?
W1 L1 Cox V , V , V : CK in TH 1 2
if we choose
12
:37
q1 . Unfortunately, the assumption of equal splitting of charge between source and drain is
W2
0:5W1 and
L2
L1 ,
then
generally invalid, making this approach less attractive. Interestingly, with the choice
W2 = 0:5W1 and L2 = L1 , the effect of clock feedthrough is suppressed. As depicted in Fig. 12.25, the total charge in Vout is zero because
CK CK CK CK
V in
M1 CH
M2
Vout
W 1 C ov CH
2W 2 C ov
Figure 12.25.
1 Cov , VCK W C +W C + 2W C 1
ov
ov
VCK W C
1
ov + CH
2W2 Cov
+2
W2Cov
0:
12
:38
Another approach to lowering the effect of charge injection incorporates both PMOS and NMOS devices such that the opposite charge packets injected by the two cancel each other (Fig. 12.26). For q1 to cancel q2 , we must have W1L1 Cox VCK
CK M1 V in M2 CK
Electrons q 1
q 2 Holes
CH
Figure 12.26.
Thus, the cancellation occurs for only one input level. Even for clock feedthrough, the circuit does not provide complete cancellation because the gate-drain overlap capacitance of NFETs is not equal to that of PFETs.
413
Our knowledge of the advantages of differential circuits suggests that the problem of charge injection may be relieved through differential operation. As shown in Fig. 12.27, we surmise that the charge injection appears as a common-mode disturbance. But, writing q1
CK M1 CH
WLCoxVCK ,
V in1
q 1
M2
Vout1
V in2
q 2
Vout2 CH
Vin1 , VTH 1, and q2 = WLCoxVCK , Vin2 , VTH 2, we recognize that q1 = q2 only if Vin1 = Vin2 . In other words, the overall error is not suppressed for differential signals. Nevertheless,
this technique both removes the constant offset and lowers the nonlinear component. This can be understood by writing q1 , q2
= =
(12.39)
Vin1 :
(12.40)
body effect now appears in both square-root terms of (12.40), leading to only odd-order distortion The problem of charge injection continues to limit the speed-precision envelope in sampled-data systems. Many cancellation techniques have been introduced but each leading to other trade-offs. One such technique, called bottom-plate sampling, is widely used in switched-capacitor circuits and is described later in this chapter.
12.3
Switched-Capacitor Ampliers
As mentioned in Section 12.1 and exemplied by the circuit of Fig. 12.4, CMOS feedback ampliers are more easily implemented with a capacitive feedback network than a resistive one. Having examined sampling techniques, we are now ready to study a number of switched-capacitor ampliers. Our objective is to understand the underlying principles as well as the speed-precision trade-offs encountered in the design of each circuit.
414
Before studying SC ampliers, it is helpful to briey look at the physical implementation of capacitors in CMOS technology. A simple capacitor structure is shown in Fig. 12.28(a), where the top plate is realized by a polysilicon layer and the bottom plate by a heavily-doped n+ region.
B Poly A SiO 2
C AB CP
n+
p substrate
(a)
CP
(b)
Figure 12.28.
(a) Monolithic capacitor structure, (b) circuit model of (a) including parasitic capacitance to the substrate.
The dielectric is the thin oxide layer used in MOS devices as well.6 An important concern in using this structure is the parasitic capacitance between each plate and the substrate. In particular, the bottom plate suffers from substantial junction capacitance to the underlying about 10 to 20% of the oxide capacitance. For this reason, we usually model the capacitor as in Fig. 12.28(b). Monolithic capacitors are described in more detail in Chapters ?? and ??. 12.3.1 Unity-Gain Sampler/Buffer While a unity-gain amplier can be realized with no resistors or capacitors in the feedback network [Fig. 12.29(a)], for discrete-time applications, it still requires a sampling circuit. We may therefore
V in V in
p region - typically
Vout
S1
Vout CH
(b)
(a)
Figure 12.29. (a) Unity-gain buffer, (b) sampling circuit followed by unity-gain buffer.
conceive the circuit shown in Fig. 12.29(b) as a sampler/buffer. However, the input-dependent charge injected by S1 onto CH limits the accuracy here. Now consider the topology depicted in Fig. 12.30(a), where three switches control the sampling
and amplication modes. In the sampling mode, S1 and S2 are on and S3 is off, yielding the topology
The oxide thickness in this type of amplier is typically thicker than that of MOS gate area because silicon dioxide grows faster on a heavily-doped material.
6
415
Vout
V0 V0 V in CH X
(b)
CH Vout X Vout
(c)
Figure 12.30.
(a) Unity-gain sampler, (b) circuit of (a) in sampling mode, (c) circuit of (a) in amplication mode.
shown in Fig. 12.30(b). Thus, Vout = VX 0, and the voltage across CH tracks Vin . At t = t0 , when Vin = V0 , S1 and S2 turn off and S3 turns on, placing the capacitor around the op amp and entering the circuit into the amplication mode [Fig. 12.30(c)]. Since the op amps high gain requires that node X still be a virtual ground and since the charge on the capacitor must be it can be processed by subsequent stages. With proper timing, the circuit of Fig. 12.30(a) can substantially alleviate the problem of channel charge injection. As Fig. 12.31 illustrates in slow motion, in the transition from the
S2 V in S1 CH X
(a)
conserved, Vout rises to a value approximately equal to V0 . This voltage is therefore frozen and
Vout
q 2
V in S1 CH X
S2
S3
Vout
CH X
Vout
(b)
(c)
Figure 12.31.
S2
S1 does.
We carefully
416
WLCoxVCK , VTH , VX . Although body effect makes VTH a function of VX , q2 is relatively constant because VX is quite independent of Vin . The constant magnitude of q2 means that channel charge of S2 introduces only an offset (rather
device whose gate voltage equals VCK , then q2
=
q2 onto CH , producing an error equal to q2 =CH . However, this charge is quite independent of the input level because node X is a virtual ground. For example, if S2 is realized by an NMOS
examine the effect of the charge injected by S2 and S1 . When S2 turns off, it injects a charge packet
than gain error or nonlinearity) in the input/output characteristic. As described below, this offset can easily be removed by differential operation. But, how about the charge injected by
CH ? Let us set Vin to zero and suppose S1 injects a charge packet q1 onto node P [Fig. 12.32(a)]. If the capacitance connected from X to ground (including the input capacitance of the op amp) is
S1 V in P CH X Vout
S1 onto
q 1
(a)
CH S1 V in P CH X Vout CX
(b)
X Vout CX A v1
(c)
q 1
Figure 12.32.
Effect of charge injected by S1 with (a) zero and (b) nite op amp input capacitance, (c) transition of circuit to amplication of mode.
zero, VP and VX jump to innity. To simplify the analysis, we assume a total capacitance equal to CX from X to ground [Fig. 12.32(b)], and we will see shortly that its value does not affect the
results. In Fig. 12.32(b), each of CH and CX carries a charge equal to q1 . Now, as shown in Fig. 12.32(c), we place CH around the op amp, seeking to obtain the resulting output voltage. To calculate the output voltage, we must make an important observation: the total charge at
node X cannot change after S2 turns off because no path exists for electrons to ow into or out of
CX is zero, it must still add up to zero after S1 injects charge because no resistive path is connected to X . The same holds true after CH is placed around the op amp. Now consider the circuit of Fig. 12.32(c), assuming the the total charge at node X is zero. We can write CX VX , Vout , VX CH = 0, and VX = ,Vout =Av1 . Thus, ,CX + CH Vout =Av1 ,
this node. Thus, if before S1 turns off, the total charge on the right plate of CH and the top plate of
417
Vout CH = 0, i.e., Vout = 0. Note that this result is independent of q1 , capacitor values, or the gain of the op amp, thereby revealing that the charge injection by S1 introduces no error if S2 turns off
rst. In summary, in Fig. 12.30(a), after S2 turns off, node X oats, maintaining a constant total charge regardless of the transitions at other nodes of the circuit. As a result, after the feedback conguration is formed, the output voltage is not inuenced by the charge injection due to
instantaneous input level across CH and yielding a charge equal to V0 CH on the left plate of CH . hence the output voltage to be approximately equal to V0 CH . After settling with feedback, node X is again a virtual ground, forcing CH to still carry V0 CH and The effect of the charge injected by S1 can be studied from yet another perspective. Suppose in
From another point of view, node X is a virtual ground at the moment S2 turns off, freezing the
S1 .
to be zero, the charge on the left plate of CH must be positive and that on its right plate negative, giving Vout
Fig. 12.32(c), the output voltage is nite and positive. Then, since VX = Vout =,Av1 , VX must be nite and negative, requiring negative charge on the top plate of CX . For the total charge at X 0. Thus, the only valid solution is Vout
=
S3 must establish an inversion layer at its oxide interface. Does the required channel charge come from CH
The third switch in Fig. 12.30(a), In order to turn on,
0.
or from the op amp? We note from the foregoing analysis that after the feedback circuit has settled, entirely supplied by the op amp, introducing no error.
the charge on CH equals V0 CH , unaffected by S3 . The channel charge of this switch is therefore Our study of Fig. 12.30(a) thus far suggests that, with proper timing, the charge injected by S1
and S3 is unimportant and the channel charge of S2 results in a constant offset voltage. Fig. 12.33 depicts a simple realization of the clock edges to ensure S1 turns off after S2 does.
S3 CK S1 CH X Vout S2 V in
Figure 12.33.
The input-independent nature of the charge injected by the reset switch allows complete cancellation by differential operation. Illustrated in Fig. 12.34, such an approach employs a differential
0 appears as a op amp along with two sampling capacitors so that the charge injected by S2 and S2
418
CH
X Vout
S eq CH Y
S2
S 3
Figure 12.34.
common-mode disturbance at nodes X and Y . This is in contrast to the behavior of the differential circuit shown in Fig. 12.27, where the input-dependent charge injection still leads to nonlinearity. switch, Seq , that turns off slightly after charge at nodes X and Y .
0 exhibit a nite charge injection mismatch, an issue resolved by adding another In reality, S2 and S2
Precision Considerations The circuit of Fig. 12.30(a) operates as a unity-gain buffer in the amplication mode, producing an output voltage approximately equal to the voltage stored across the capacitor. How close to unity is the gain here? As a general case, we assume the op amp exhibits a nite input capacitance the op amp, VX
Cin
and calculate the output voltage when the circuit goes The
from the sampling mode to the amplication mode (Fig. 12.35). Owing to the nite gain of
V in
CH , raising the charge on CH to CH V0 + CinVX .7 It follows that the voltage across CH equals CH V0 + CinVX =CH . We therefore
conservation of charge at
7
requires that
Cin VX
come from
The charge on CH increases because moving positive charge from the left plate of leads to a more positive voltage across CH .
419
Vout
(12.41)
(12.42)
1, then Vout
must be minimized even if speed is not critical. Recall from Chapter 9 that to increase
Av 1 ,
we may choose a large width for the input transistors of the op amp, but at the cost of higher input capacitance. An optimum device size must therefore yield minimum gain error rather than maximum Av1 . Example 12.3
In the circuit of Fig. 12.35, Cin = 0:5 pF and CH guarantees a gain error of 0.1%? Solution. Since Cin =CH
=
Speed Considerations Let us rst examine the circuit in the sampling mode [Fig. 12.36(a)]. What is the time constant in this phase? The total resistance in series with
CK CK S1 CH X
(a)
CH is given by Ron1
S2 IX Vout X VX
R on2
V in
G mV X
(b)
R0
Figure 12.36. (a) Unity-gain sampler in sampling mode, (b) equivalent circuit of (a).
12.36(b), where R0 denotes the open-loop output impedance of the op amp, we have
X and ground, RX .
IX , Gm VX R0 + IX Ron2 = VX ;
12
:43
420
R0 and Gm R0
R0 + Ron2 : RX = 1 +G R
1, we have RX
12
:44
sam
=
1 Ron1 + G CH :
12
:45
The magnitude of sam must be sufciently small to allow settling in the test case of Fig. 12.14 to the required precision. Now let us consider the circuit as it enters the amplication mode. Shown in Fig. 12.37 along with both the op amp input capacitance and the load capacitance, the circuit must begin
CH X C in CL
0
VX Vout
V0
V0
Vout t0 t
Figure 12.37.
VCH V0 , then VX = ,V0 at the beginning of the amplication mode. In other words, the input
difference sensed by the op amp initially jumps to a large value, possibly causing the op amp to slew. But, let us rst assume the op amp can be modeled by a linear model and determine the output response. To simplify the analysis, we represent the charge on CH by an explicit series voltage source,
with Vout 0 and eventually produce Vout V0 . If Cin is relatively small, we can assume that the voltages across CL and CH do not change instantaneously, concluding that if Vout 0 and
VS , that goes from zero to V0 at t = t0 while CH carries no charge itself (Fig. 12.38). The objective is to obtain the transfer function Vout s=VS s and hence the step response. We have
Also, since the current through Cin equals VX Cin s, 1 Vout R
+
0
12
:46 :47
12
421
VS C H X C in VX G mV X R0 CL Vout
Figure 12.38. Equivalent circuit of unity-gain circuit in amplication mode. Calculating VX from (12.47) and substituting in (12.46), we arrive at the transfer function:
Vout s = R Gm + Cin sCH 12:48 0 VS R0 CLCin + Cin CL + CH CLs + Gm R0 CH + CH + Cin : Note that for s = 0, (12.48) reduces to a form similar to (12.41). Since typically Gm R0CH CH ; Cin, we can simplify (12.48) as Vout s = Gm + Cin sCH : 12:49 VS CL Cin + Cin CL + CH CL s + Gm CH
Thus, the response is characterized by a time constant equal to
amp =
CLCin + CinCL + CH CL ; Gm CH
12
:50
which is independent of the op amp output resistance. This is because a higher R0 leads to a greater to CL =Gm , an expected result because with negligible Cin , the output resistance of the unity-gain We now study the slewing behavior of the circuit, considering a telescopic op amp as an example. Upon entering the amplication mode, the circuit may experience a large step at the inverting input (Fig. 12.37). As shown in Fig. 12.39, the tail current of the op amps input differential pair is then steered to one side, charging the capacitance seen at the output. Since M2 is off during slewing, loop gain, evetually yielding a constant closed-loop speed. If Cin
Cin is negligible and the slew rate is approximately equal to ISS =CL. The slewing continues until VX is sufciently close to the gate voltage of M1 , after which point the
Our foregoing studies reveal that the input capacitance of the op amp degrades both the speed
settling progresses with the time constant given in (12.50). and the precision of the unity-gain sampler/buffer. For this reason, the bottom plate of CH in Fig. 12.30 is usually driven by the input signal or the output of the op amp and the top plate is connected to node X (Fig. 12.40), minimizing the parasitic capacitance seen from node X to ground. This technique is called bottom-plate sampling.
422
M7
M8
VDD
M5
M6
I SS Vout
M3 Vb I SS M1
M4 M2
V0
Off
CH X
CL
I SS
It is instructive to compare the performance of the sampling circuits shown in Figs. 12.29(b) and 12.30(a). In Fig. 12.29(b), the sampling time constant is smaller because it depends on only the on-resistance of the switch. More importantly, in Fig. 12.29(b), the amplication after the switch turns off is almost instantaneous whereas in Fig. 12.30, it requires a nite settling time. However, the critical advantage of the unity-gain sampler is the input-independent charge injection. 12.3.2 Noninverting Amplier In this section, we revisit the amplier of Fig. 12.4, studying its speed and precision properties. Repeated in Fig. 12.41(a), the amplier operates as follows. In the sampling mode, S1 and S2 are voltage [Fig. 12.41(b)]. At the end of the sampling mode, S2 turns off rst, injecting a constant on and S3 is off, creating a virtual ground at X and allowing the voltage across C1 to track the input
charge, q2, onto node X . Subsequently, S1 turns off and S3 turns on [Fig. 12.41(c)]. Since VP goes from Vin0 to 0, the output voltage changes from 0 to approximately Vin0 C1=C2 , providing a has the same polarity as Vin0 and the gain can be greater than unity.
voltage gain equal to C1 =C2 . We call the circuit a noninverting amplier because the nal output
423
S2 C2 V in S1 C1 X S3 Vout V in P C1 X Vout
(a)
(b)
C2 V in0 P C1 X Vout t
(c)
C1 C2
V in0
Figure 12.41. (a) Noninverting amplier, (b) circuit of (a) in sampling mode, (c) transition of circuit to amplication mode.
As with the unity-gain circuit of Fig. 12.30(a), the noninverting amplier avoids inputdependent charge injection by proper timing, namely, turning After
S2
off before
S1
(Fig. 12.42).
q 2
V in0 P C X 1
V in
Vout
S1 carefully. As illustrated in Fig. 12.43, the charge injected by S1 , q1 , changes the voltage at node P by approximately VP = q1=C1 , and hence the output voltage by ,q1 C1=C2 . However, after S3 turns on, VP drops to zero. Thus, the overall change in VP is equal to 0 , Vin0 = ,Vin0 , producing an overall change in the output equal to ,Vin0 ,C1=C2 = Vin0 C1 =C2 . The key point here is that VP goes from a xed voltage, V0 , to another, 0, with an intermediate perturbation due to S1 . Since the output voltage of interest is measured after node P is connected to ground, the charge injected by S1 does not affect the nal output. From another perspective, as shown in Fig. 12.44, the charge on the right plate of C1 at the instant S2 turns off is approximately
charge injection of Let us rst study the effect of
424
VP
VP S 3 turns on. t C1 C2
q 1
V in0
S2 C2 S1
0
V in0 Vout
V in
Vout
V in
P C X 1
Figure 12.44. Charge redistribution in noninverting amplier. equal to ,Vin0 C1. Also, the total charge at node X must remain constant after S2 turns off. Thus, when node
charge are nearly zero, and the charge ,Vin0 C1 must reside on the left plate of C2 . In other words, the output voltage is approximately equal to Vin0 regardless of the intermediate excursions at node The foregoing discussion indicates that two other phenomena have no effect on the nal output.
P.
First, from the time S2 turns off until the time S1 turns off, the input voltage may change signicantly (Fig. 12.45) without introducing any error. In other words, the sampling instant is dened by the value of VP is zero, this charge is unimportant. Neither of these effects introduces error because the total charge at node turn-off of
S2 .
Second, when
S3 turns on, it requires some channel charge but since the nal
emphasize that VP is initially and nally determined by xed voltages, we say node P is driven
X is conserved and VP
low-impedance distinguishes node P , at which charge is not conserved, from oating nodes such as X , where charge is conserved.
or node P switches from a low-impedance node to another low-impedance node. Here the term
425
Figure 12.45. Effect of input change after S2 turns off. In summary, proper timing in Fig. 12.41(a) ensures that node X is perturbed by only the charge
injection of S2, making the nal value of Vout free from errors due to S1 and S3. The constant offset due to S2 can be suppressed by differential operation (Fig. 12.46).
S3 S2 C2 S1 V in
S1
C1
X Vout
S eq C1 Y C2
S2
S 3
Figure 12.46.
0 In the differential circuit of Fig. 12.46, suppose the equalizing switch is not used and S2 and S2
exhibit a threshold voltage mismatch of 10 mV. If C1 all switches
=
Example 12.4
WLCox = 50 fF, calculate the dc offset measured at the output assuming all of the 0 is injected onto X and Y , respectively. channel charge of S2 and S2 Solution. Simplifying the circuit as in Fig. 12.47, we have Vout q=C2 , where q = WLCoxVTH . Note that C1 does not appear in the result because X is a virtual ground, i.e., the voltage across C1 changes only negligibly. Thus, the injected charge resides primarily on the left plate of C2 , giving an output error voltage equal to Vout = WLCox VTH =C2 = 1 mV.
1 pF, C2
426
q
C1
S2 C2
S1 V in
S1
X Vout
C1 Y C2 S 2
Figure 12.47.
Precision Considerations As mentioned above, the circuit of Fig. 12.41(a) provides a nominal gain equal to
voltage gain of C1=C2 . We now calculate the actual gain if the op amp exhibits a nite open-loop
Av 1 .
Depicted in Fig. 12.48 along with the input capacitance of the op amp, the
C2 C1 V in C in X Vout A v1
Figure 12.48.
Vout , VX C2s = VX Cins + VX , Vin C1 s: Vout = ,C1 Vin C2 + C2 + C1 + Cin : Av1
12
:51
Since Vout
,Av1 VX , we have
12
:52
Vout , C1 1 , C2 + C1 + Cin 1 ; 12:53 Vin C2 C2 Av1 implying that the amplier suffers from a gain error of C2 + C1 + Cin =C2 Av1 . Note that the gain error increases with the nominal gain C1=C2 . Comparing (12.42) with (12.53), we note that with CH = C2 and for a nominal gain of unity, the
noninverting amplier exhibits greater gain error than does the unity-gain sampler. This is because
427
For example, if Cin is negligible, the unity-gain buffers gain error is half that of the noninverting amplier. Speed Considerations The smaller feedback factor in Fig. 12.48 suggests that the time response of the amplier may be slower than that of the unity-gain sampler. This is indeed true. Consider the equivalent circuit shown in Fig. 12.49(a). Since the only difference between this circuit and that
C2 C1 V in X VX
(a)
Vout G mV X R0 CL
C in
C2 C eq
V in
X VX
(b)
Vout G mV X R0 CL
Figure 12.49. (a) Equivalent circuit of noninverting amplier in amplication mode, (b) circuit of (a) with Vin ; C1, and Cin replaced by a Thevenin equivalent. in Fig. 12.38 is the capacitor C1, which is connected from node X to an ideal voltage source, we But for a more rigorous analysis, we substitute Vin ; C1 , and
expect that (12.50) gives the time constant of this amplier as well if Cin is replaced by Cin + C1 .
Cin in Fig. 12.49(a) by a Thevenin = C1 =C1 + Cin , and Ceq = C1 + Cin , and note that
eq
+
2
eq VX = Vin , Vout C C+ C
Thus,
Vout :
12
:54
eq Vin , Vout C C+ C eq
+
2
1 Vout Gm + Vout R
12
:55
and hence
1 ,Ceq C C Gm , C2 sR0 Vout s = 1 + Cin Vin C2Gm R0 + Ceq + C2 + R0 CLCeq + C2 + Ceq C2 s :
12
:56
428
12
:57
amp
12
:58
which is the same as the time constant of Fig. 12.37 if Cin is replaced by Cin + C1. Note the direct dependence of amp upon the nominal gain, C1=C2 . It is instructive to examine the ampliers time constant for the special case CL
=
(12.58) yields amp = C1 + Cin =Gm , a value independent of the feedback capacitor. This is because, while a larger C2 introduces heavier loading at the output, it also provides a greater feedback factor. 12.3.3 Precision Multiply-by-Two Circuit The circuit of Fig. 12.41(a) can operate with a relatively high closed-loop gain, but it suffers from speed and precision degradation due to the low feedback factor. In this section, we study a topology that provides a nominal gain of two while achieving a higher speed and lower gain error [5]. Shown in Fig. 12.50(a), the amplier incorporates two equal capacitors, the sampling mode, the circuit is congured as in Fig. 12.50(b), establishing a virtual ground at X
0. Equation
C1 = C2 = C .
In
and allowing the voltage across C1 and C2 to track Vin . In the transition to the amplication mode,
S3 turns off rst, C1 is placed around the op amp, and the left plate of C2 is switched to ground [Fig. 12.50(c)]. Since at the moment S3 turns off, the total charge on C1 and C2 equals 2Vin0 C (if the charge injected by S3 is neglected), and since the voltage across C2 approaches zero in the amplication mode, the nal voltage across C1 and hence the output voltage are approximately equal to 2Vin0 . This can also be seen from the slow motion illustration of Fig. 12.51. The reader can show that the charge injected by S1 and S2 and absorbed by S4 and S5 is unimportant and that injected by S3 introduces a constant offset. The offset can be suppressed by
differential operation. The speed and precision of the multiply-by-two circuit are expressed by (12.58) and (12.53), respectively, but the advantage of the circuit is the higher feedback factor for a given closed-loop gain. Note, however, that the input capacitance of the multiply-by-two circuit in the sampling mode is higher.
429
S4 S3 S1 V in S2 C2 S5 C1 X Vout
(a)
C1 V in C2
(b)
C1 X Vout C2
(c)
Vout
Figure 12.50. (a) Multiply-by-two circuit, (a) circuit of (a) in sampling mode, (b) circuit of (a) in amplication mode.
V in0 V in C1 C2 V in0
S3
2V in0
C1 C2
0
Vout
Figure 12.51.
12.4
Switched-Capacitor Integrator
Integrators are used in many analog systems. Examples include lters and oversampled analog-todigital converters. Fig. 12.52 depicts a continuous-time integrator, whose output can be expressed
CF V in R X Vout
as
Vout = , RC
1
Vin dt;
12
:59
430
if the op amp gain is very large. For sampled-data systems, we must devise a discrete-time counterpart of this circuit. connected between two nodes [Fig. 12.53(a)], carrying a current equal to VA , VB =R. The role
A VA R I B VB VA
(b)
Before studying SC integrators, let us rst point out an interesting property. Consider a resistor
S1
S2 CS
B VB
(a)
Figure 12.53. (a) Continuous-time and (b) discrete-time resistors. of the resistor is to take a certain amount of charge from node A every second and move it to node Can we perform the same function by a capacitor? Suppose in the circuit of Fig. 12.53(b),
B.
capacitor CS is alternately connected to nodes A and B at a clock rate fCK . The average current owing from A to B is then equal to the charge moved in on clock period:
IAB
= =
(12.60) (12.61)
We can therefore view the circuit as a resistor equal to CS fCK ,1 . Recognized by James Clark Maxwell, this property formed the foundation for many modern switched-capacitor circuits. integrator of Fig. 12.54(a). We note that in every clock cycle, C1 absorbs a charge equal to C1 Vin
C2 V in S1 P S2 X C1 t
(a) (b)
R in Fig.
C1 Vout Vout C2
V in
Figure 12.54. (a) Discrete-time integrator, (b) response of circuit to a constant input voltage.
when
S1 is on and deposits the charge on C2 when S2 is on (node X is a virtual ground). For example, if Vin is constant, the output changes by Vin C1 =C2 every clock cycle [Fig. 12.54(b)].
Approximating the staircase waveform by a ramp, we note that the circuit behaves as an integrator.
431
12
:62
where the gain of the op amp is assumed large. Note that the small-signal settling time constant as The integrator of Fig. 12.54(a) suffers from two important drawbacks. First, the input-
S1 introduces nonlinearity in the charge stored on C1 and hence the output voltage. Second, the nonlinear capacitance at node P resulting from the source/drain junctions of S1 and S2 leads to a nonlinear charge-to-voltage conversion when C1 is switched to X . This can be understood with the aid of Fig. 12.55, where the charge stored on the total junction capacitance, Cj , is not equal to Vin0 Cj , but rather equal to
C2 P Vin0 C1 Cj S2 X Vout
qcj =
Z V in0
0
Cj dV:
12
:63
Since Cj is a function of voltage, qcj exhibits a nonlinear dependence on Vin0 , thereby creating a nonlinear component at the output after the charge is transferred to the integration capacitor. An integrator topology that resolves both of the foregoing issues is shown in Fig. 12.56(a). We study the circuits operation in the sampling and integration modes. As shown in Fig. 12.56(b), in the sampling mode S1 and S3 are on and S2 and S4 are off, allowing the voltage across C1 to track
Vin while the op amp and C2 hold the previous value. In the transition to the integration mode, S3 turns off rst, injecting a constant charge onto C1, S1 turns off next, and subsequently S2 and S4 turn on [Fig. 12.56(c)]. The charge stored on C1 is therefore transferred to C2 through the virtual
ground node. differential operation. Moreover, because the left plate of
C1 is driven (Section 12.3.2), the charge injection or absorption of S1 and S2 contributes no error. Also, since node X is a virtual ground, the charge injected or absorbed by S4 is constant and independent of Vin .
Since
S3
turns off rst, it introduces only a constant offset, which can be suppressed by
432
C2 V in S1 C1 P S4 X S2 S3 Vout
(a)
C2 C1 V in Vout C1
C2
Vout
(b)
(c)
Figure 12.56. (a) Parasitic-insensitive integrator, (b) circuit of (a) in sampling mode, (c) circuit of (a) in integration mode. How about the nonlinear junction capacitance of S3 and S4 ? We observe that the voltage across this capacitance goes from near zero in the sampling mode to virtual ground in the integration mode. Since the voltage across the nonlinear capacitance changes by a very small amount, the resulting nonlinearity is negligible.
12.5
Our study of common-mode feedback in Chapter 9 suggested that sensing the output CM level by means of resistors lowers the differential voltage gain of the circuit considerably. We also observed that sensing techniques using MOSFETs that operate as source followers or variable resistors suffer from a limited linear range. Switched-capacitor CMFB networks provide an alternative that avoids both of these difculties (but the circuit must be refreshed periodically.) resistors. Figure 12.57 depicts a simple example, where equal capacitors C1 and C2 reproduce at node In switched-capacitor common-mode feedback, the outputs are sensed by capacitors rather than
X the average of the changes in each output voltage. Thus, if Vout1 and Vout2 experience a, say, positive CM change, then VX and hence ID5 increase, pulling Vout1 and Vout2 down. The output CM level is then equal to VGS 2 plus the voltage across C1 and C2 . How is the voltage across C1 and C2 dened? This is typically carried out when the amplier S1 is on.
Transistors
is in the sampling (or reset) mode and can be accomplished as shown in Fig. 12.58. Here, during and M7 operate as a linear sense circuit because their gate voltages are nominally equal. Thus, the CM level denition, the amplier differential input is zero and switch
M6
circuit settles such that the ouput CM level is equal to VGS 6;7 + VGS 5 . At the end of this mode, S1
433
M3 C1 V out1 M1 M5
M4 C2
VDD Vb V out2
V in
M2
Figure 12.57.
M3
M4
C1 X C2
V in
M1 M5
M2
S1
ID
Figure 12.58. Denition of the voltage across C1 and C2. turns off, leaving a voltage equal to VGS 5;6 across C1 and C2. In the amplication mode, M6 and
M7 may experience a large nonlinearity but they do not impact the performance of the main circuit because S1 is off.
example, the topology shown in Fig. 12.59 may be used. Here, in the reset mode, one plate of C1
M3 S2 V CM S4 M1 M5 C1 C2 M4 S3 S5 M2 V CM VDD Vb
In applications where the output CM level must be dened more accurately than in the above
V in
I REF S1 M6
434
ID3 and ID4 with respect to IREF can guarantee that VGS5 = VGS6 and hence the output CM level is equal to VCM .
differential output [6]. For this reason, part of the tail current of the differential pairs in Figs. 12.58 to the circuit. Problems
, VGS6 . In the amplication mode, S2 and S3 are on and the other switches are off, yielding an output CM level equal to VCM , VGS 6 + VGS 5 . Proper denition of
With large output swings, the speed of the CMFB loop may in fact inuence the settling of the
and C2 is switched to VCM while the other is connected to the gate of M6 . Each capacitor therefore
and 12.59 can be provided by a constant current source so that M5 makes only small adjustments
Unless otherwise stated, in the following problems, use the device data shown in Table 2.1 and assume VDD
=
RF = 1 but the op amp has an output resistance Rout , derive the transfer function Vout s=Vin s. (b) If the op amp is ideal, determine the minimum value of RF that guarantees a gain error of
1% for an input frequency of 1 MHz. resistance Rout .
2 pF and C2
0:5 pF.
12.2 Suppose in Fig. 12.5(a), the op amp is characterized by a transconductance Gm and an output (a) Determine the transfer function Vout =Vin in this mode.
(b) Plot the waveform at node B if Vin is a 100-MHz sinusoid with a peak amplitude of 1 V, C1 = 1 pF, Gm = 1=100 , and Rout = 20 k. 12.3 In Fig. 12.5(b), node
A is in fact connected to ground through a switch (Fig. 12.4). If the switch introduces a series resistance Ron and the op amp is ideal, calculate the time constant
of the circuit in this mode. What is the total energy dissipated in the switch as the circuit enters the amplication mode and Vout settles to its nal value? 20=0:5 and CH
12.4 The circuit of Fig. 12.9(a) is designed with W=L1 (b) Approximating M1 by a linear resistor equal to
1 pF.
(a) Using Eqs. (12.7) and (12.14), calculate the rime required for Vout to drop to +1 mV.
the time required for Vout to drop to +1 mV and compare the result with that obtained in part (a).
435
12.5 The circuit of Fig. 12.11 cannot be characterized by a single time constant because the
if
(a) Using Eq. (12.19), calculate the time required for Vout to reach 2.1 V. (b) Sketch the transconductance of M1 versus time. 12.6 In the circuit of Fig. 12.8(b),
W=L1 = 20=0:5 and CH = 1 pF. Assume = = 0 and Vin = V0 sin !in t + Vm, where !in = 2 100 MHz). (a) Calculate Ron1 and the phase shift from the input to the output if V0 = Vm = 10 mV. (b) Repeat part (a) if V0 = 10 mV but Vm = 1 V. The variation of the phase shift translates to
distortion. 12.7 Describe an efcient SPICE simulation that yields the plot of 12.16. 12.8 The sampling network of Fig. 12.16 is designed with drop to +1 mV. 12.9 In the circuit of Fig. 12.19, feedthrough. 12.10 The circuit of Fig. 12.60 samples the input on C1 when CK is high and connects C1 and C2
CK V in M1 C1 CK M2 Vout C2
W=L2 = 60=0:5. If Vin = 0 and the initial value of Vout is +3 V, estimate the time required for Vout to
W=L1
20=0:5 and
W=L1
20=0:5 and
CH
error at the output due to charge injection. Compare this error with that resulting from clock
W=L2 and C1 = C2 . (a) If the initial voltages across C1 and C2 are zero and Vin = 2 V, plot Vout versus time for
=
many clock cycles. Neglect charge injection and clock feedthrough. and M2 ? Assume the channel charge of M2 splits equally between C1 and C2. (b) What is the maximum error in Vout due to charge injection and clock feedthrough of M1
436
12.11 For Vin = V0 sin !0 t + V0 , where V0 = 0:5 V and !0 = 2 10 MHz), plot the output waveforms of the circuits shownin Fig. 12.29(b) and 12.30(a). Assume a clock frequency of 50 MHz. 12.12 In Fig. 12.45, S1 turns off t seconds after S2 and S3 turns on t seconds after S1 turns off.
S1-S3 .
Plot the output waveform, taking into account the charge injection and clock feedthough of 2 pF, Cin 0:2 pF and Av
the maximum nominal gain, C1=C2 , that the circuit can provide with a gain error of 1%?
1000. What is
12.14 In Problem 12.13, what is the maximum nominal gain if 12.15 The integrator of Fig. 12.54 is designed with C1
MHz. Neglecting charge injection and clock feedthrough, sketch the output if the input is a estimate the output amplitude.
10-MHz sinusoid with a peak amplitude of 0.5 V. Approximating C1 ; S1, and S2 by a resistor,
12.16 Consider the switched-capacitor amplier depicted in Fig. 12.61, where the common-mode
VDD Vb V CM C3 M1 C2 V CM M2 C4 M3 M4
C1 V in
I SS
Figure 12.61.
C3 = C4 = 0:5 pF, and the output CM level is 1.5 V. Neglect the transistor capacitances.
(a) What is the maximum allowable output voltage swing in the amplication mode?
W=L1,4
50=0:5; ISS
1 mA,
C1
C2
2 pF,
437
(b) Determine the gain error of the amplier. (c) What is the small-signal time constant in the amplication mode? 12.17 Repeat Problem 12.16 if the gate-source capacitance of M1 and M2 is not neglected. 12.18 A differential circuit incorporating a well-designed common-mode feedback network exhibits the open-loop input-output characteristic shown in Fig. 12.62(a). In some circuits, however, the characteristic appears as in Fig. 12.62(b). Explain how this effect occurs.
Vout1 Vout1
Vout2
0 (a)
Vout2 V in
0 (b)
V in
Figure 12.62.
12.19 In the common-mode feedback network of Fig. 12.58, assume transistors, ID5 = 1 mA, and ID6;7 common-mode level.
=
W=L
10=0:5.
12.21 Suppose in the common-mode feedback network of Fig. 12.58, S1 injects a charge of q onto the gate of
M5 .
change due to this error? 12.22 In the circuit of Fig. 12.63, each op amp is represented by a Norton equivalent and characterized by Gm and Rout . The output currents of two op amps are summed at node Y [7]. (The circuit is shown in the amplication mode.) Note that the main amplier and the auxiliary a proportional current into node Y . Assume Gm Rout (a) Calculate the gain error of the circuit. (b) Repeat part (a) if the auxiliary and error ampliers are eliminated and compare the results. amplier are identical and the error amplier senses the voltage variation at node X and injects 1.
References
Main Amplifier C2
438
C1 V in Gm R out Y Vout
Gm
Error R out Amplifier
C2 C1 X Gm R out
Auxiliary Amplifier
Figure 12.63.
References
[1] G. Wegmann, E. A. Vittoz, and F. Rahali, Charge Injection in Analog MOS Switches, IEEE J. Solid-State Circuits, vol. SC-22, pp. 1091-1097, Dec. 1987. [2] B. J. Sheu and C. Hu, Switch-Induced Error Voltage on a Switched Capacitor, IEEE J. Solid-State Circuits, vol. SC-19, pp. 519-525, April 1984. [3] R. Gregorian and G. C. Temes, Analog MOS Integrated Circuits for Signal Processing, John Wiley and Sons, New York, 1986. [4] J. H. Fischer, Noise Sources and Calculation Techniques for Switched Capacitor Filters, IEEE J. Solid-State Circuits, vol. 17, pp. 742-752, Aug. 1982. [5] B. S. Song, M. F. Tompsett, and K. R. Lakshmikumar, A 12-Bit 1-Msample/s CapacitorAveraging Pipelined A/D Converter, IEEE J. Solid-State Circuits, vol. SC-23, pp. 1324-1333, Dec. 1988. [6] B. Razavi, Principles of Data Conversion System Design, IEEE Press, New York, 1995. [7] P. C. Yu and H.-S. Lee, A High-Swing 2-V CMOS Op Amp with Replica-Amp Gain Enhancement, IEEE J. Solid-State Circuits, vol. 28, pp. 1265-1272, Dec. 1993.