• Qinwen FanEmail author
  • Kofi A. A. Makinwa
  • Johan H. Huijsing
Part of the Analog Circuits and Signal Processing book series (ACSP)


In many measurement situations, the signal of interest is small, possibly in the range of tens of microvolts, and is superimposed on a much larger DC common-mode (CM) signal, possibly in the range of several volts.


Breakdown Voltage Differential Pair Instrumentation Amplifier Versus Sense Input Transistor 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

1.1 Problem

In many measurement situations, the signal of interest is small, possibly in the range of tens of microvolts, and is superimposed on a much larger DC common-mode (CM) signal, possibly in the range of several volts. Coping with such a large CM signal and at the same time accurately measuring such small signals is a big challenge for interface circuits. A good example of such a measurement is in high-side current sensing [1], as shown in Fig. 1.1, where the load current of a battery is monitored by inserting a small sensing resistor R sense in series with the battery. Thus, the current can be determined from the DC voltage drop V sense across the resistor. To minimize its power consumption, R sense is usually very small (hundreds of milliohm) and thus, V sense is also small, typically ranging from tens of microvolts to hundreds of millivolts. This requires a readout amplifier with low offset and low 1/f noise. Moreover, V sense is accompanied by a large CM voltage, which can be as large as 30 V in the case of a laptop battery. This is far beyond the supply voltages of normal CMOS circuitry. Thus, novel circuit techniques to reject this large CM voltage and accurately measure V sense must be found. This problem becomes more challenging as CMOS technology advances, since this has historically been accompanied by a steady decrease in supply voltages.
Fig. 1.1

A simplified schematic of a high-side current-sensing readout circuit

1.2 Traditional Solutions

When a large input CM voltage must be rejected, the best solution is to block it. A first approach involves the use of a magnetic coupling [2, 3, 4, 5]. The basic block diagram [2] of a readout system employing magnetic coupling is shown in Fig. 1.2. It consists of an input and output modulator, a transformer, and a readout amplifier. In this case, the modulators are implemented as choppers, i.e., polarity-reversing switches driven by a digital clock signal with a fixed frequency (f chop) [6, 7, 8, 9]. In [2], the choppers are also driven via an isolating transformer. The input chopper converts the DC differential voltage V sense into a square wave, and the output chopper converts the amplified square wave back to DC. In this way, the differential signal is first modulated to high frequency by the input chopper and so can be coupled to the input of the readout amplifier via a transformer. The DC CM voltage, however, will not be modulated and thus will not be coupled to the readout amplifier. Furthermore, the offset and 1/f noise of the readout amplifier will be up-modulated by the output chopper and so can be filtered out. A big disadvantage of this approach, however, is that transformers are difficult to integrate on chip. Although integrated microtransformers [2] can be realized in some processes, they tend to occupy a lot of chip area.
Fig. 1.2

A schematic of magnetic isolation readout circuit for current-sensing application

Fig. 1.3

Schematic of CM isolation with opto-isolator

A second approach involves isolating the DC CM voltage optically, e.g., with an opto-isolator [10, 11, 12, 13]. Although there are many types of opto-isolator, the most common type simply consists of an LED and a photodiode as shown in Fig. 1.3. The LED is connected to the input signal and converts it into an optical signal, which is then picked up and converted back to an electrical signal by the photodiode and a readout amplifier. In this way, the input CM voltage is completely isolated from the readout amplifier. The main disadvantage of this approach, however, is its lack of accuracy. The signal transfer function between the LED and the photodiode depends on several parameters such as the voltage-to-light transfer function of the LED, the intensity of the light picked up by the photodiode, and the light-to-voltage transfer function of the photodiode. These parameters can be difficult to control and reproduce accurately in large-scale production. Thus, the accuracy of the measurement, especially the system’s overall gain, is not well defined. Moreover, the linearity of the signal is often low, which requires an extra feedback circuit [4].
Fig. 1.4

CM isolation with basic differential pair

A third approach, which is the most commonly used, is to isolate the CM voltage electrically. A simple way to do this is to use the basic differential pair shown in Fig. 1.4. Neglecting circuit non-idealities, the input differential pair acts as an ideal voltage to current converter, which floats between the tail current source and a potential close to ground. Since it is only sensitive to the input differential signal, the CM signal is completely isolated from the rest of the circuit. In reality, however, circuit non-idealities such as mismatch will significantly decrease the measurement accuracy. To improve this, the chopping technique can again be applied as shown in Fig. 1.5. Since the input chopper only modulates the differential input signal, its contribution to the output voltage is separated from that of the CM signal in the frequency domain. The offset and 1/f noise are also up-modulated and so are removed from the base band. However, the maximum CM voltage that can be handled is determined by the differential pair’s supply. So handling a large CM voltage requires a large supply voltage and, consequently, high-voltage input transistors in the input stage. This increases the power consumption considerably [14, 15]. Moreover, the limited output impedance of the input transistors will reduce the circuit’s CM immunity, while their offset and 1/f noise will reduce its DC precision. Finally, the common-mode voltage range (CMVR) of the circuit shown in Fig. 1.5 will not cover both the negative and positive rails [14, 15] (Fig. 1.3).
Fig. 1.5

CM isolation with chopper differential pair

Another approach to isolate the input CM voltage is to use a so-called flying capacitor [16] to sample and hold the input signal, as shown in Fig. 1.6. The CM voltage at the input of the succeeding amplifier is set by a feedback resistor network and so it can be realized with low-voltage circuitry. However, the kT/C noise associated with the sample-and-hold action of the flying capacitor increases the total input-referred noise. Moreover, continuous-time operation is not possible. Last but not least, the input switches must once more be able to handle the large CM signal.
Fig. 1.6

Schematic of CM isolation with the flying capacitor

From the above introduction, it seems that none of these approaches is very satisfactory. Thus, a new approach is required.

1.3 A Promising Solution: Capacitively Coupled Chopper Amplifier

An intuitively appealing solution to the problem of CM isolation is the use of capacitive coupling. Capacitors are widely available in most standard CMOS process and exhibit a natural ability to block DC signals without any extra power consumption. Thus, a capacitively coupled amplifier will perfectly reject DC CM voltages, as long as they are smaller than the breakdown voltage of the coupling capacitors (Fig. 1.7). Although the breakdown voltages of on-chip capacitors is usually less than 100 V, this is still sufficient for many applications. However, it is also obvious that the DC input signal is also blocked. One solution is to use the chopping technique described in the previous sections to up-modulate the input signal. This has the added advantage of suppressing the offset and low 1/f noise of the input stage.
Fig. 1.7

Schematic of a capacitively coupled amplifier

However, the concept of capacitively coupled chopper amplifiers is not new! As early as 1940, the classic capacitively coupled chopper amplifier shown in Fig. 1.8 was invented [6]. The input signal is up-modulated by the input chopper, amplified, and finally de-modulated by the output chopper. However, the CM voltage is also modulated and so no CM isolation can be obtained. The differential structure, as shown in Fig. 1.9, changes the story completely. The input chopper modulates the DC differential signal to high frequencies, which can then travel through the input capacitors. The DC CM signal, however, is blocked. As a result, the input CM level of the succeeding amplifier can be fixed arbitrarily via biasing resistors R 1,2 and so the amplifier can be implemented with low-voltage circuitry. At the amplifier’s output, an output chopper demodulates the signal back to the base band. The offset and 1/f noise of this amplifier, however, are blocked/filtered by the output capacitors. To prevent its output from saturating, the gain of the (transconductance) amplifier is limited by the output resistors R out1,2. However, the amplifier cannot be used as an operational amplifier (opamp) due to its low gain or an instrumentation amplifier (IA) due to its inaccurate gain (G m1*R out1,2).
Fig. 1.8

Schematic of a classic single-ended capacitively coupled chopper amplifier

Fig. 1.9

Schematic of a classic fully differential capacitively coupled chopper amplifier

In 2007, the first capacitively coupled chopper instrumentation amplifier (CCIA) was described by Timothy Denison [17]. It is shown in Fig. 1.10. This work represents a great improvement on Fig. 1.9 topology. By eliminating the need for R out1,2 and C out1,2, the open-loop gain of the amplifier can be quite large. A chopped capacitor feedback path ensures that the gain of the IA is accurately defined as C in1,2/C fb1,2. Since the amplifier was intended for biomedical applications, a DC servo loop was employed (an SC integrator and C hp1,2) to give it a high-pass characteristic. The up-modulated offset and 1/f noise of G m1 was suppressed by a second stage (not shown) which acts as a low-pass filter. Although not designed for high input CM voltages, this CCIA demonstrated the feasibility of realizing on-chip capacitively coupled precision IAs. In this thesis, the design of capacitively coupled chopper IAs and opamps that can handle large input CM voltages, in some cases even larger than their supply voltages, will be explored.
Fig. 1.10

Block diagram of the first CCIA: from Timothy Denison [18]

1.4 Challenging Issues

Although the use of capacitively coupled chopper amplifiers seems to be a promising way of obtaining both wide CMVR and high DC precision, there are still many issues to be solved. The first issue is the robust implementation of the input chopper, which must handle the high CM voltage present before the input capacitors. Otherwise, the maximum input CM will be limited by the input chopper and not by the breakdown voltage of the input capacitors.

The input impedance of the amplifier is also an issue. Since the input capacitors are constantly switched between V in+ and V in−, they are constantly being charged and discharged, which requires a certain amount of input current. The resulting impedance depends on the input capacitance and the chopping frequency, and typically ranges from hundreds of kilo-ohms to tens of megaohms, which may not be high enough for some applications. Thus, techniques to boost the input impedance are required.

Since the input DC CM voltage is blocked, the DC CM level of the amplifier’s input stage (G m1 in Fig. 1.9) must be fixed by something else. This can be done by using biasing resistors (R 1,2 in Fig. 1.9). However, as will be shown later, these resistors introduce noise, and so must be rather large (hundreds of megaohms). Thus, another challenge is that of realizing such large resistors in an area-efficient and relatively accurate manner.

A fourth issue of capacitively coupled chopper amplifiers is the up-modulated offset and 1/f noise of the amplifier. Without the AC-coupling output capacitors shown in Fig. 1.9, the amplifier’s offset will be up-modulated by the output chopper and result in ripple. This is usually not acceptable and, thus, must be suppressed effectively. A low-pass filter (Fig. 1.10) is the most straightforward way but it often involves the use of large passive components. In recent years, more effective ripple-reduction techniques [18, 19, 20, 21] have been developed. However, care must be taken to minimize the extra area and power required, especially in low power and cost sensitive applications.

Finally, when a ripple-reduction technique is applied, it is very likely to introduce a notch at the chopping frequency in the amplifier’s transfer function. This notch limits the amplifier’s useable bandwidth and, moreover, results in a step response that will exhibit a certain amount of undesirable ringing. Thus, techniques of dealing with such transfer function notches must be devised.

1.5 Organization

In Chap.  2, the chopping technique, which has been frequently mentioned in the above, is introduced in detail. Several other commonly used dynamic offset cancelation techniques are also introduced. Like chopping, these techniques can also be used to achieve low offset and 1/f noise. As mentioned above, the up-modulated offset and 1/f noise due to chopping results in ripple, which must be sufficiently suppressed. Thus, ripple-reduction techniques will also be discussed.

In Chap.  3, the proposed capacitively coupled chopper topology for both opamps and IAs will be described. Its operating principles will be discussed in detail as well as its strengths and weaknesses.

As mentioned earlier, input capacitive coupling is not the only prerequisite to obtaining wide input CMVR. Thus, specially designed input choppers with wide CMVR are proposed, which will be presented in Chap.  4.

In Chaps.  5, two opamp prototypes employing the capacitively coupled chopper topology are presented. First, an overview of the state of the art is given. Later, the design considerations of the prototypes on both system level and transistor level are described. A single-path capacitively coupled chopper opamp is presented first, followed by a multipath capacitively coupled chopper opamp that improves the high-frequency performance and step response of the single-path opamp. Measurement results of each opamp will be given.

In Chaps.  6 and  7, two IA prototypes employing the capacitively coupled chopper topology will be proposed. The first IA features a ±30 V input CMVR and high power efficiency and is designed for HV current-sensing applications. The second IA is wireless sensor node, where power efficiency and small chip area are critical. It can be operated in two modes: a DC mode for DC and low frequency signals sensing; and an AC mode for biomedical signals such as ECG sensing. As in Chap.  5, a state-of-the-art overview will be given at the beginning of each chapter. The systematic and transistor level designs will be presented as well as experimental results.

The thesis ends with conclusions and future work, which are presented in Chap.  8. Two capacitively coupled chopper analog-to-digital converters (ADC) are proposed as future work, which can directly digitize a signal source with high CM voltage. The original contributions of the author are listed at the end of the chapters.


  1. 1.
    A. Mehta, “Understand low-side vs. high-side current sensing,”
  2. 2.
    F. Rothan, H. Lhermet, B. Zongo etc, “A ± 1.5 % nonlinearity 0.1-to-100A shunt current sensor based on a 6 kV isolated micro-transformer for electrical vehicles and home automation,” ISSCC Dig. Tech. Papers, pp. 112-113, Feb., 2011.Google Scholar
  3. 3.
    M. Munzer, W. Ademmer, B. Strzalkowski, and K. Kaschani, “Insulated signal transfer in a half bridge driver IC based on a coreless transformer technology,” Int. Conf. Power Electronics and Drive Systems, vol. 1, pp. 93-96, Nov., 2003.Google Scholar
  4. 4.
    K. Kaschani, B. Strzalkowski, M. Münzer and W. Ademmer, “Coreless transformer a new technology for driver ICs,” PCIM 2003, Nürnberg.Google Scholar
  5. 5.
    S. Kaeriyama, S. Uchida, M. Furumiya, M. Okada, T. Maeda, and M. Mizuno, A 2.5 kV isolation 35 kV/us CMR 250 mbps digital isolator in standard CMOS with a small transformer driving technique,” IEEE J. Solid-State Circuits, vol. 47, no. 2, pp. 435-443, Feb., 2012.Google Scholar
  6. 6.
    J. H. Huijsing, Operational Amplifiers: Theory and Design Second Edition, New York: Springer, 2011.Google Scholar
  7. 7.
    C. C. Enz, and G. C. Temes, “Circuit techniques for reducing the effects of Op-Amp imperfections: autozeroing, correlated double sampling, and chopper stabilization,” Proceedings of IEEE, vol. 84, no. 11, pp. 1584-1614, Nov., 1996.Google Scholar
  8. 8.
    C. Menolfi, and Q. Huang, “A fully integrated, untrimmed CMOS instrumentation amplifier with submicrovolt offset”, IEEE J. Solid-State Circuits, vol. 34, no. 3, pp. 415-420, Mar., 1999.Google Scholar
  9. 9.
    K.C. Hsieh, P.R. Gray, D. Senderowicz, D.G. Messerschmitt, “A low-noise chopper-stabilized differential switched-capacitor filtering technique”, IEEE J. Solid-State Circuits, vol. 16, no. 6, pp. 708-715, Dec., 1981.Google Scholar
  10. 10.
    Wikipedia, “Opto-isolator,”
  11. 11.
    H.L. Skolnik, “Design considerations for linear optically coupled isolation amplifiers,” IEEE J. Solid-State Circuits, vol.17, no. 6, pp. 1094-1101, Dec., 1982.Google Scholar
  12. 12.
    W. Olschewski, “Optical coupling extends isolation-amplifier utility,” Electronics, Aug., 1976.Google Scholar
  13. 13.
    1S0100 Product Data Sheet, Burr-Brown Res. Corp., Tucson, AZ, Mar., 1982.Google Scholar
  14. 14.
    J. F. Witte, J. H. Huijsing, and K. A. A. Makinwa, “A current-feedback instrumentation amplifier with 5 µV offset for bidirectional high-side current-sensing,” IEEE J. Solid-State Circuits, vol. 43, no. 12, pp. 2769-2775, Dec., 2008.Google Scholar
  15. 15.
    V. Schaffer, M. F. Snoeij, M. V. Ivanov, and D. T. Trifonov, “A 36 V programmable instrumentation amplifier with sub-20 µV offset and a CMRR in excess of 120 dB at all gain settings,” IEEE J. Solid-State Circuits, vol. 44, no. 7, pp. 2036-2046, July, 2009.Google Scholar
  16. 16.
    LTC1043, Linear Technology,, Jan., 1984.
  17. 17.
    T. Denison, K. Consoer,W. Santa, et al., “A 2 µW 100nV/√Hz chopper stabilized instrumentation amplifier for chronic measurement of neural field potentials,” IEEE JSSC, vol. 42, no. 12, pp. 2934–2945, Dec., 2007.Google Scholar
  18. 18.
    R. Burt, and J.A. Zhang, “Micropower chopper-stabilized operational amplifier using a SC notch filter with synchronous integration inside the continuous-time signal path,” IEEE J. Solid-State Circuits, vol. 41, no. 12, pp. 2729-2736, Dec., 2006.Google Scholar
  19. 19.
    M. Belloni, E. Bonizzoni, A. Fornasari, F. Maloberti, “A micropower chopper-correlated double-sampling amplifier with 2 µV standard deviation offset and 37nV/√Hz input noise density,” IEEE ISSCC Dig. Tech. papers, pp. 76-77, Feb., 2010.Google Scholar
  20. 20.
    R.Wu, K.A.A. Makinwa, J.H. Huijsing, “A chopper current-feedback instrumentation amplifier with a 1 mHz 1/f noise corner and an AC-coupled ripple-reduction loop,” IEEE J. Solid-State Circuits, vol.44, no. 12, pp. 3232–3243, Dec., 2009.Google Scholar
  21. 21.
    Y. Kusuda, “Auto correction feedback for ripple suppression in a chopper amplifier,” IEEE J. Solid-State Circuits, vol.45, No. 8, pp. 1436-1445, Aug., 2010.Google Scholar

Copyright information

© Springer International Publishing AG 2017

Authors and Affiliations

  • Qinwen Fan
    • 1
    Email author
  • Kofi A. A. Makinwa
    • 2
  • Johan H. Huijsing
    • 3
  1. 1.Mellanox TechnologiesDelfgauwThe Netherlands
  2. 2.Delft University of TechnologyDelftThe Netherlands
  3. 3.SchipluidenThe Netherlands

Personalised recommendations