Notes on OPAMP Circuits By Dr.

Venkat Ramaswamy Technical University of Sydney

1. CHARACTERISTICS AND PARAMETERS OF OPERATIONAL AMPLIFIERS

The characteristics of an ideal operational amplifier are described first, and the characteristics and performance limitations of a practical operational amplifier are described next. There is a section on classification of operational amplifiers and some notes on how to select an operational amplifier for an application. 1.1 IDEAL OPERATIONAL AMPLIFIER 1.1.1 Properties of An Ideal Operational Amplifier The characteristics or the properties of an ideal operational amplifier are: i. ii. iii. iv. v. vi. Infinite Open Loop Gain, Infinite Input Impedance, Zero Output Impedance, Infinite Bandwidth, Zero Output Offset, and Zero Noise Contribution.

The opamp, an abbreviation for the operational amplifier, is the most important linear IC. The circuit symbol of an opamp shown in Fig. 1.1. The three terminals are: the non-inverting input terminal, the inverting input terminal and the output terminal. The details of power supply are not shown in a circuit symbol.

1.1.2 Infinite Open Loop Gain From Fig.1.1, it is found that vo = - Ao vi, where `Ao' is known as the open-loop gain of the opamp. Let vo be -10 Volts, and Ao be 105. Then vi is 100 !V. Here

the input voltage is very small compared to the output voltage. If Ao is very large, vi is negligibly small for a finite vo. For the ideal opamp, Ao is taken to be infinite in value. That means, for an ideal opamp vi = 0 for a finite vo. Typical values of Ao range from 20,000 in low-grade consumer audio-range opamps to more than 2,000,000 in premium grade opamps ( typically 200,000 to 300,000). The first property of an ideal opamp: Open Loop Gain Ao = infinity. 1.1.3 Infinite Input Impedance and Zero Output Impedance An ideal opamp has an infinite input impedance and zero output impedance. The sketch in Fig. 1.2 is used to illustrate these properties. From Fig. 1.2, it can be seen that iin is zero if Rin is equal to infinity. The second property of an ideal opamp: Rin = infinity or iin =0.

From Fig. 1.2, we get that

If the output resistance Ro is very small, there is no drop in output voltage due to the output resistance of an opamp. The third property of an ideal opamp: Ro = 0.

1.1.4 Infinite Bandwidth An ideal opamp has an infinite bandwidth. A practical opamp has a limited bandwidth, which falls far short of the ideal value. The variation of gain with frequency has been shown in Fig. 1.3, which is obtained by modelling the opamp with a single dominant pole, whereas the practical opamp may have more than a single pole. The asymptotic log-magnitude plot in Fig. 1.3 can be expressed by a first-order equation shown below.

It is seen that two frequencies, wH and wT, have been marked in the frequency response plot in Fig. 1.3.. Here wT is the frequency at which the gain A(jw) is equal to unity. If A(jwT) is to be equal to unity,

Since Ao is very large, it means that wT = Ao * wH .

1.1.5 Zero Noise Contribution and Zero Output Offset A practical opamp generates noise signals, like any other device, whereas an ideal opamp produces no noise. Premium opamps are available which contribute very low noise to the rest of circuits. These devices are usually called as premium lownoise types. The output offset voltage of any amplifier is the output voltage that exists when it should be zero. In an ideal opamp, this offset voltage is zero. 1.2 PRACTICAL OPERATIONAL AMPLIFIERS

This section describes the properties of practical opamps and relates these characteristics to design of analog electronic circuits. A practical operational amplifier has limitations to its performance. It is necessary to understand these limitations in order to select the correct opamp for an application and design the circuit properly. Like any other semiconductor device, a practical opamp also has a code number. For example, let us take the code LM 741CP. The first two letters, LM here, denote the manufacturer. The next three digits, 741 here, is the type number. 741 is a general-purpose opamp. The letter following the type number, C here, indicates the temperature range. The temperature range codes are: C commercial 0o C to 70o C, I industrial -25o C to 85o C and M military -55o C to 125o C. The last letter indicates the package. Package codes are: D Plastic dual-in-line for surface mounting on a pc board J Ceramic dual-in-line N,P Plastic dual-in-line for insertion into sockets. 1.2.1 Standard Operational Amplifier Parameters Understanding operational amplifier circuits requires knowledge of the parameters given in specification sheets. The list below represents the most commonly needed parameters. Methods of measuring some of these parameters are described later in this lesson.

Open-Loop Voltage Gain. Voltage gain is defined as the ratio of output voltage to an input signal voltage, as shown in Fig. 1.1. The voltage gain is a dimensionless quantity. Large Signal Voltage Gain. This is the ratio of the maximum allowable output voltage swing (usually one to several volts less than V- and V++) to the input signal required to produce a swing of 10 volts (or some other standard). Slew rate. The slew rate is the maximum rate at which the output voltage of an opamp can change and is measured in terms of voltage change per unit of time. It varies from 0.5 V/!s to 35 V/!s. Slew rate is usually measured in the unity gain noninverting amplifier configuration. Common Mode Rejection Ratio. A common mode voltage is one that is presented simultaneously to both inverting and noninverting inputs. In an ideal opamp, the output signal due to the common mode input voltage is zero, but it is nonzero in a practical device. The common mode rejection ratio (CMRR) is the measure of the device's ability to reject common mode signals, and is expressed as the ratio of the differential gain to the common mode gain. The CMRR is usually expressed in decibels, with common devices having ratings between 60 dB to 120 dB. The higher the CMRR is, the better the device is deemed to be. Input Offset Voltage. The dc voltage that must be applied at the input terminal to force the quiescent dc output voltage to zero or other level, if specified, given that the input signal voltage is zero. The output of an ideal opamp is zero when there is no input signal applied to it. Power-supply rejection ratio. The power-supply rejection ratio PSRR is the ratio of the change in input offset voltage to the corresponding change in one power-supply, with all remaining power voltages held constant. The PSRR is also called "power supply insensitivity". Typical values are in !V / V or mV/V. Input Bias Current. The average of the currents into the two input terminals with the output at zero volts. Input Offset Current. The difference between the currents into the two input terminals with the output held at zero. Differential Input Impedance. The resistance between the inverting and the noninverting inputs. This value is typically very high: 1 M" in low-cost bipolar

opamps and over 1012 Ohms in premium BiMOS devices. Common-mode Input Impedance The impedance between the ground and the input terminals, with the input terminals tied together. This is a large value, of the order of several tens of M" or more. Output Impedance. The output resistance is typically less than 100 Ohms. Average Temperature Coefficient of Input Offset Current. The ratio of the change in input offset current to the change in free-air or ambient temperature. This is an average value for the specified range. Average Temperature Coefficient of Input Offset Voltage. The ratio of the change in input offset voltage to the change in free-air or ambient temperature. This is an average value for the specified range. Output offset voltage. The output offset voltage is the voltage at the output terminal with respect to ground when both the input terminals are grounded. Output Short-Circuit Current. The current that flows in the output terminal when the output load resistance external to the amplifier is zero ohms (a short to the common terminal). Channel Separation. This parameter is used on multiple opamp ICs (device in which two or more opamps sharing the same package with common supply terminals). The separation specification describes part of the isolation between the opamps inside the same package. It is measured in decibels. The 747 dual opamp, for example, offers 120 dB of channel separation. From this specification, we may state that a 1 !V change will occur in the output of one of the amplifiers, when the other amplifier output changes by 1 volt. 1.2.2 Minimum and Maximum Parameter Ratings Operational amplifiers, like all electronic components, are subject to maximum ratings. If these ratings are exceeded, the device failure is the normal consequent result. The ratings described below are commonly used. Maximum Supply Voltage. This is the maximum voltage that can be applied to the opamp without damaging it. The opamp uses a positive and a negative DC

power supply, which are typically 18 V. Maximum Differential Supply Voltage. This is the maximum difference signal that can be applied safely to the opamp power supply terminals. Often this is not the same as the sum of the maximum supply voltage ratings. For example, 741 has 18 V as the maximum power supply voltage, whereas the maximum differential supply voltage is only 30 V. It means that if the positive supply is 18 V, the negative supply can be only -12 V. Power dissipation, Pd. This rating is the maximum power dissipation of the opamp in the normal ambient temperature range. A typical rating is 500 mW. Maximum Power Consumption. The maximum power dissipation, usually under output short circuit conditions, that the device can survive. This rating includes both internal power dissipation as well as device output power requirements. Maximum Input Voltage. This is the maximum voltage that can be applied simultaneously to both inputs. Thus, it is also the maximum common-mode voltage. In most bipolar opamps, the maximum input voltage is nearly equal to the power supply voltage. There is also a maximum input voltage that can be applied to either input when the other input is grounded. Differential Input Voltage. This is the maximum differential-mode voltage that can be applied across the inverting and noninverting inputs.

Maximum Operating Temperature. The maximum temperature is the highest ambient temperature at which the device will operate according to specifications with a specified level of reliability. Minimum Operating Temperature. The lowest temperature at which the device operates within specification. Output Short-Circuit Duration. This is the length of time the opamp will safely sustain a short circuit of the output terminal. Many modern opamps can carry short circuit current indefinitely. Maximum Output Voltage. The maximum output potential of the opamp is related to the DC power supply voltages. Typical for a bipolar opamp with 15 V power supply, the maximum output voltage is typically about 13 V and the

minimum - 13 V. Maximum Output Voltage Swing. This is the maximum output swing that can be obtained without significant distortion(at a given load resistance). Full-power bandwidth. This is the maximum frequency at which a sinusoid whose size is the output voltage range is obtained. 1.2.3 Comparisons and Typical Values Table 1.1 presents a summary of features of an ideal and a typical practical opamp. Table 1.1: Comparison of an ideal and a typical practical opamp Property Open-loop gain Open-loop bandwidth CMRR Input Resistance Output Resistance Input Bias Currents Offset Voltages Offset Currents Slew Rate Drift Ideal Infinite Infinite Infinite Infinite Zero Zero Zero Zero Infinite Zero Practical(Typical) Very high (>10000 ) Dominant pole(!10 Hz) High (> 60 dB) High (>1 M") Low(< 100 ") Low (< 50 nA) Low (< 10 mV) Low (< 50 nA) A few V/!s Low

Table 1.2 shown below presents a summary of the effects of opamp characteristics on a circuit's performance. It is a simplified summary. 1.2.4 Effect of Feedback on Frequency Response The effect of feedback on the frequency response of a system has already been described. Here the effect of feedback is described using the log-magnitude plot. Given that the transfer function of the forward path is specified as:

TABLE 1.2 EFFECTS OF CHARACTERISTICS ON OPAMP APPLICATIONS __________________________________________________________________ OPAMP APPLICATION ___________________________________________ DC amplifier AC amplifier _________________ __________________ Opamp Characteristic that may affect Small Large Small Large performance output output output output __________________________________________________________________ 1. Input bias current Yes Maybe No No 2. Offset current Yes Maybe No No 3. Input offset voltage Yes Maybe No No 4. Drift Yes Maybe No No 5. Frequency Response No No Yes Yes 6. Slew rate No Yes No Yes ________________________________________________________________

If the closed-loop transfer function T(s) of the circuit is

On substituting for A(s) by its expression in equation (1.1), we get that

The plot of frequency response for open loop and closed-loop is shown in Fig. 1.4.

1.3 CLASSIFICATION OF OPAMPS The classification of an opamp can be based on either its function or its family type. The classification based on function is described below. i. General-purpose amplifier. These general purpose opamps are neither special purpose or premium devices. Most of them are internally compensate, so designers trade off bandwidth for inherent stability. A general purpose opamp is the default choice for an application unless a property of another class brings a unique advantage to this application. ii. Instrumentation amplifier. Although an instrumentation amplifier is arguably a special purpose device, it is sufficiently universal to warrant a class of its own. iii. Voltage Comparators. These devices are not true opamps, but are based on opamp circuitry. While all opamps can be used as voltage comparators, the reverse is not true. The special feature of a comparator is the speed at which its output level can change from one level to the other. iv. Low Input Current. The quiescent current needed for these opamps is low. This class of opamps typical uses MOSFET , JFET or superbeta (Darlington) transistors for the input stage instead of npn/pnp bipolar devices. v. Low Noise. generated noise. These devices are usually optimized to reduce internally

10

vi. Low Power. This category of opamp optimizes internal circuitry to reduce power consumption. Many of these devices also operate at very low DC power supply potentials. vii. Low Drift. All DC amplifiers suffer from drift. Devices in this category are internally compensated to minimize drift due to temperature. These devices are typically used in instrumentation circuits where drift is an important concern, especially when handling low level input signals. viii. Wide Bandwidth. The devices in this class are also called as video opamps and have a very high gain-bandwidth level, as high as 100 MHz. Note that 741 has a gain-bandwidth product of about 1 MHz . ix Single DC Supply. These devices are designed to operate from a single DC power supply. x. V. xi. High Voltage. The power supply for these devices can be as high as 44 Multiple devices. Two or quad arrangement in one IC.

The classification based on family type: i. Bipolar opamps, ii. BiFET opamps, iii. JET opamps, iv. CMOS opamps etc. The characteristics of opamps change with the internal architecture also. Some opamps have two-stage architecture, whereas some have three-stage architecture. The purpose of this section is to highlight the facts that it is necessary to select a suitable opamp for the application in hand and that there is a wide choice available. Choosing the right opamp is not simple. Aspects to be considered are: technology,dc performance,ac performance,output drive requirements, supply requirements, quiescent current level, temperature range of operation, nature of input signal, costs etc. Table 1.3 presents a summary of characteristics of a few selected opamps. It is preferable to go through the databooks on linear ICs for selecting the right opamp. 1.3.1 BiFET OPAMP Although the LM741 and other bipolar opamps are still widely used, they are nearly obsolete. Bipolar technology has been replaced by BiFET technology. The term, "BiFET" stands for bipolar-field effect transistor. It is a combination of two technologies, bipolar and junction field-effect, making use of the advantages of

11

each. Bipolar devices are good for power handling and speed whereas field-effect devices have very high input impedances and low power consumption. Most modern general-purpose opamps are now produced with BiFET technology. BiFET opamps generally have enhanced characteristics over bipolar opamps. They have a much greater input impedance, a wider bandwidth, a higher slew rate and larger power output than the corresponding ratings of bipolar opamps. A variety of BiFET opamps are now available: the TL060 low-power, TL070 lownoise and TL080 general purpose from Texas Instruments, LF350 and LF440 series from National semiconductor, the MC34000 and MC35000 series from Motorola etc. The performance of most of them are similar and are normally pinout compatible with 741C. Extremely low bias currents make a BiFET opamp to be more suitable for applications such as an integrator, a sample and hold circuit and filter circuits. But the bias currents double for every 10o C and at high temperature, a BiFET opamp may have a larger bias current than a bipolar opamp! Both BiFET and CMOS have less noise current, which is an important consideration when dealing with sources of high impedance. A BiFET opamp has some disadvantages too. It tends to have a far greater offset voltage than its bipolar counterpart. The offset voltages tend to be unstable too. In addition, a BiFET opamp has poorer CMRR, PSRR and open-loop gain specifications. But some of the recent BiFET do not have these drawbacks. It may please be noted that a BiFET opamp needs a dual power supply. Unlike 741C, TL080 does not have internal compensation and needs an external capacitor of value ranging from 10 pF to 20 pF to be connected between pins 1 and 8. The smaller the capacitor is, the wider the bandwidth is, but the opamp tends to become more underdamped at higher frequencies. 1.3.2 CMOS OPAMPs Although originally considered to be unstable for linear applications, a CMOS opamp is now a real alternative to many bipolar, BiFET and even dielectrically isolated opamps. The major advantage of a CMOS opamp is that it operates well with a single supply. The input common-mode range is more or less the same as the power supply range. A CMOS opamp needs very low supply currents, less than 10 !A

12

and can operate with a supply voltage as low as 1.4 V, making it ideally suitable for battery-powered applications. In addition, a CMOS opamp has high input impedance and low bias currents. On the other hand, a CMOS tends to have limited supply voltage range. Its offset voltages tend to be higher than those of a bipolar opamp. TABLE 1.3 Typical Performance of Selected Opamp Types Type 741 2-stage
Input offset voltage (mV) Bias Current (nA) Offset Current (nA) Open-loop gain (dB) Input Resistance (M") Slew Rate (V/!s) Unity-gain bandwidth (MHz) Full-power bandwidth (kHz) Settling time(!s) CMRR(dB)

LM 118 3-stage <4

LM 108 Super # <2

AD 611 BiFET <0.5

AD 570K Wideband <5

<5

<500

<250

<2

<0.025

<15

<200

<50

<0.4

<0.01

<15

106 2

100 5

95 100

98 106

100 300

0.5 1

>50 15

0.2 1

13 2

35 35

10

1000

200

600

1.5 80

4 90

1 95

3 80

0.9 100

13

2. BASIC OPAMP APPLICATIONS

2.1 NONINVERTING AMPLIFIER

The basic noninverting amplifier can be represented as shown in Fig. 2.1. Note that a circuit diagram normally does not show the power supply connections explicitly.

2.1.1 Analysis For An Ideal Opamp An ideal opamp has infinite gain. This means that

Thus,

An ideal opamp has infinite input resistance. That is,

We obtain the output voltage as:

14

The gain of the noninverting amplifier is then:

2.1.2 Analysis For an Opamp with a finite gain, Ao Let the opamp have a finite gain. Then the noninverting amplifier can be represented by the equivalent circuit in Fig. 2.2.

From Fig 2.2,

On re-arranging,

15

We can represent the circuit in Fig. 2.2 by a block diagram that represents the feedback that is present in the circuit. From Fig. 2.2, we can state that,

The above equation can be represented by a block diagram as shown in Fig. 2.3. From the block diagram, we get the same expression for the gain of the circuit. It can be seen that if the open loop gain Ao tends to infinity, equation (2.7) reduces to equation (2.6).

Next we analyse the same circuit when the opamp has a finite input resistance, a finite gain and a nonzero output resistance.

2.2 INVERTING AMPLIFIER The inverting amplifier is analysed below using both network theory and feedback
16

theory approach. 2.2.1 Analysis Based On Circuit Theory Analysis is as follows. Apply KCL (Kirchoff's Current Law) at node `a' in Fig. 2.4. Then

For an ideal opamp, ii = 0. Hence is + i2 = 0. Thus the KCL at node 'a' is:

For an ideal opamp, its output resistance is zero. Hence - A ovi = vo. When the gain is infinity, vi is also zero. Therefore, In other words,

When an opamp is considered to be ideal, vi and ii have zero value. If the NI input (noninverting input) is grounded, the inverting input is at zero potential. We can find is & i2 by treating the potential at inverting input terminal to be zero volts. In this condition, the inverting input terminal behaves as if it is grounded and is called as `virtual ground'. When the NI input is not grounded, the inverting input is not at ground potential,it does not behave as if it is grounded and it is no longer called the virtual ground. All that happens is that its potential is the same as that at the NI input. 2.2.2 Analysis Based on Negative Feedback Now the circuit in Fig. 2.4 is to be represented as a system with feedback. From Fig. 2.4, we get that

We can arrive at the result shown above by the use of either superposition theorem
17

or by adding the drop across R1 to Vs. From Fig. 2.4, we get that

where Ao is the gain of the opamp. Here it is appropriate to call it as the opamp's open loop gain. The above two equations can be represented by a block diagram as shown in Fig. 2.5.

For the block diagram in Fig. 2.5, we get the ratio Vo/Vs as

Since the open loop gain tends to be infinite, we get the same ratio for Vo/Vs as obtained earlier. In this case, the feedback that is present in the circuit is negative because the opamp has a negative gain. An opamp has a negative gain when vi is measured as shown in Fig. 2.4. 2.2.3 Applications and Extensions to Inverting Amplifier An inverter is a basic application using opamp. described by the equation An opamp's output can be

where

With v+ = 0, and v- = 0, vcm = 0. Even though the common-mode gain Acm of

18

opamp may not be zero(common-mode gain is zero for ideal opamp), its contribution to output is almost zero in the case of an inverting amplifier, because vcm = 0. The circuit of an inverting amplifier can be extended to more than one input. A circuit with two inputs is shown in Fig. 2.6. For the circuit in Fig. 2.6,

If v1 and v2 be of opposite signs, the circuit in Fig. 2.6 can be used as a proportional controller. Even though the inverting amplifier is a reliable and useful circuit,it is not suitable if the source vs has a large source resistance. Normally the value of source resistance is not known precisely. With a source resistance Rs, the output of circuit in Fig. 2,4 is given by

It can be seen that the output may be unreliable in this case. For such applications, we need a circuit with a very high input resistance. A non-inverting amplifier built with an opamp is highly suitable for this purpose. 2.3 DIFFERENCE AMPLIFIER

The circuit of a difference amplifier is shown in Fig. 2.7. Here we find out its output assuming that the opamp is ideal. It is easy to get the output using the superposition theorem. When we apply this theorem, we consider one input at a time. With v1 = 0, we can find output due to v2. Due to v2 only: Using the result obtained for the non-inverting amplifier, we get

19

Given v2,

Then,

Due to v1 only: Here we get the output from the result obtained for the inverting amplifier.

Due to both v2 and v1 :

If R1 = R3 and R2 = R4, then

The above equation explains why this circuit is known as the difference amplifier. The difference amplifier circuit is used for measuring the difference between two sources. Such circuits, or improved difference amplifier circuits are used widely in instrumentation. Configuration using three opamps, with two opamps as buffer and the third as the difference amplifier, is used more often.

20

3 AN INTRODUCTION TO FEEDBACK IN AMPLIFIERS

3.0 OBJECTIVES To show how to identify negative and positive feedback in a To identify applications suitable for positive or negative feedback. To outline the effects of feedback. To stress the need for negative feedback in amplifiers by outlining the advantages of using feedback. To show the block diagram of the four feedback topologies.

(i) circuit. (ii) (iii) (iv) (v) 3.1

IDENTIFYING THE NATURE OF FEEDBACK IN A CIRCUIT

A system with feedback is usually represented by a block diagram as shown in Fig. 3.1(a). To identify the nature of feedback, (i) (ii) (iii) assume that the input is grounded, break the loop as shown in Fig. 3.1b, and find the ratio of V2 / V1 .
21

This ratio of V2 / V1 is known as the loop gain. If the loopgain is negative, then the system has negative feedback and the system has positive feedback if the loop gain is positive. Quiz Problem 3.1 Identify the nature of feedback for the block diagrams shown in Fig. 3.2.

Answer: (a) Positive , (b) positive, (c) negative, and (d) positive. _______________________________________________________________ 3.2 EFFECT OF FEEDBACK ON PERFORMANCE

Feedback in a system can be either positive or negative. Negative feedback improves the linearity of a system and is hence employed in a linear system such as an amplifier. On the other hand, positive feedback tends to produce a two-state output in a system and is used in circuits such as square-wave oscillators, comparators, Schmitt triggers etc. However, it may be noted that every circuit with positive feedback need not have a two-state output. For example, a sinewave oscillator circuit has positive feedback without having a two-state output. The Laplace transform of

is

22

It can be seen that V(s) has a conjugate pair of poles on the imaginary axis of splane. A physical system tends to have poles on the left-hand side of s-plane. Positive feedback tends to shift some of the poles of the system towards the right-side of s-plane. Due to positive feedback, a system can have conjugate poles on the imaginary axis and such a system oscillates. In sinewave oscillators, positive feedback must be closely controlled to maintain oscillations and waveform purity. On the other hand, in a square-wave oscillator or a comparator, the effect of positive feedback at cross-over points is to increase the speed of transition from one level to the other and prevent unwanted spikes at changeover points. In terms of control theory terminology applied to linear systems, an amplifier in general represents a stable system, whereas a sinewave oscillator is a marginally stable system and a square-wave or a Schmitt trigger circuit is an `unstable' system. It is necessary to know what stability means. Where `stability' is desired as in the case of an amplifier, negative feedback is used. Where two-state/digital output is desired, positive feedback is used. It is worth remembering that the nature of feedback can change with frequency. Feedback may change from negative to positive as the frequency varies and the system may be unstable at high frequencies. Gain of the system also affects stability. Variation in component values and characteristics of devices due to operating conditions such as temperature, voltage, or current, can bring about instability despite negative feedback. Ageing leads to variation in component values and can hence affect stability. Negative feedback is widely used in an amplifier design because it produces several benefits. These benefits are: i. permits feasible ii. iii. Negative feedback stabilizes the gain of an amplifier despite the parameter changes in the active devices due to supply voltage variation, temperature change, or ageing. Negative feedback a wider range for parameter variations than what would be without feedback. Negative feedback allows the designer to modify the input and output impedances of a circuit in any desired fashion. It reduces distortion in the output of an amplifier. These distortions arise due to nonlinear gain characteristic of devices used. Negative feedback causes the gain of the amplifier to be determined by the feedback network and thus reduces distortion. Negative feedback can increase the bandwidth. It can reduce the effect of noise.

iv. v.

These benefits are obtained by sacrificing the gain of a system. Another aspect
23

to be borne in mind is that unless the feedback circuit is properly designed, the system tends to be unstable. Mathematical analysis to support the above statements can be found in section 3.4. In a circuit with negative feedback, the gain of the circuit with feedback depends almost only on the feedback network used if the loop gain is sufficiently large. It is often possible to build the feedback circuit by using only passive components. Since the passive components tend to have a stable value, the circuit performance is then independent of the parameters of the active device used, and the circuit performance becomes more reliable. The four basic feedback configurations are specified according to the nature of the input signal/input circuit and output signal/feedback arrangement. Since both the input and the output signals can be either voltage or current, there are four configurations as described in section 3.1. 3.3. FEEDBACK TOPOLOGIES

The four feedback topologies are: (i) (ii) (iii) (iv) Series-shunt topology Shunt-series topology, Series-series topology and Shunt-shunt topology.

The block diagrams for these topologies are shown in Fig. 3.3. It is seen that each topological configuration is described by two terms. The first term refers to the input stage and the second to the output. If the two signals are connected
24

in series, the term 'series' is used. It is logical to connect two voltage signals in series. For series-shunt topology, the source signal vs and the feedback signal vf are connected in series at the input and the difference ve is applied to the voltage amplifier. On the other hand if two signals are connected in parallel, the term 'shunt' is used. It is logical to connect two current sources in parallel. If the amplifier in the forward path is either a current or a transresistance amplifier, it needs a current signal at its input, which in turn is obtained as the difference of the source signal Is and the feedback current If. In either of these cases, the first term to be used is 'shunt'. The second term that describes a feedback topology is related to its output. If the output signal is a voltage and a feedback signal is to be derived from it, then the feedback network must be connected across the output terminals. It is appropriate to use the term 'shunt' here. On the other hand, if a feedback signal is to be obtained from the output current, the feedback network must be connected such that the output current flows through it to produce a feedback signal and the feedback network is then connected in 'series'. At this stage, identifying the feedback topology of a circuit may appear to be simple. Given a circuit with feedback, it turns out that it is not so straightforward to identify its topology. This aspect will be described in greater length after the study of the four topological configurations. 3.4 ANALYSIS

3.4.1 Gain Sensitivity The block diagram of a system with feedback is shown in Fig. 3.1, where G is the gain of the amplifier in the forward path. In most practical situations, gain G of the amplifier in the forward path is not well defined. For example, if we consider BJT devices with the same type number, the current gain can vary from one device to another, by as much as 50%. In addition, the gain of the forward amplifier is dependent on temperature, and other operating conditions. If an amplifier is used without feedback, the output is more sensitive to the changes in the gain of the amplifier. For example, let Vo be the output voltage and Vs be the input voltage of an amplifier with gain G. Then

If the gain of the amplifier changes by !G, the output changes by (!G * Vs).
25

The fractional change in output is then (!G) / G . Thus for an open-loop amplifier,

With a closed-loop system, the output is not as sensitive to changes in gain G. Ideally, variations in G should not affect output at all. From Fig. 2.1, the gain T of the closed-loop system is:

Then

From equations (3.3) and (3.4), we get that

For example, let G = 100, "G = 5 and H = 0.1. From equation (3.2), the fractional change in the output of the amplifier operating in open-loop is 0.05. On the other hand, the fractional change in the output of the circuit operating in closed-loop is only (0.05/11). It is seen that the output is less sensitive to changes in the gain of the amplifier in the forward path. We define the sensitivity function as follows:

3.4.2 The Effect of Feedback on Nonlinear Distortion Feedback reduces the nonlinear distortion in output. Let an amplifier be operating in open loop, as shown in Fig. 3.4a. Here it is assumed that essentially all the nonlinear distortion is produced by the output stage of the amplifier, which is typically the case. From Fig. 3.4a, we can state that

26

where Vd represents the nonlinear content in output. As illustrated in Fig. 3.4b, if the same amount of nonlinear distortion is produced at the output stage in an amplifier with feedback, the resultant output is obtained to be

For the same amount of distortion Vd introduced, the circuit with feedback has less distortion in its output. From equations (3.7) and (3.8),

3.4.3 The Effect of feedback on Frequency Response

If the gain of the amplifier in the forward path be A and the feedback factor be #, the closed-loop gain T is then A/(1 + A#). Gain A, instead of being be a constant. can be a function of frequency, with poles and zeros. Let A be
27

defined as follows.

where Ao is the mid-frequency gain, wL is the low-frequency pole and wH is the high-frequency pole. Then the frequency response is as shown in Fig. 3.5. Since the gain is a function of frequency, the closed-loop gain T also becomes a function of frequency. At low frequencies, where w << wH, we can approximate A(s) as A0 s A( s) = (311 . ) ( s + wL ) Since

we get the closed-loop gain at low frequencies by substituting for A(s) from equation (3.11) .

We get the expression shown above by assuming that the high frequency pole would not the affect the performance at low frequencies and hence it can be neglected. At high frequencies, we neglect both the zero at origin and the low frequency pole. Then at high frequencies, the closed-loop gain can be stated to be

Combining the two equations (3.13) and (3.14) , we get that

28

The frequency response of closed-loop system is also shown in Fig. 3.5. It is seen that the range of frequency response has increased significantly. In Fig. 3.5,

3.5

SUMMARY

This chapter has described briefly the four types of feedback topologies and has highlighted the effect of feedback i. on the sensitivity of amplifier to changes in the gain of the amplifier in the forward path, ii. on the nonlinear distortion and iii. on the frequency response.
29

4 SERIES-SHUNT TOPOLOGY
4.1 IDEAL SERIES-SHUNT NETWORK

The block diagram of a circuit with series-shunt feedback configuration is shown in Fig. 4.1. It is seen that the amplifier in the forward path is an ideal voltage amplifier. An ideal voltage amplifier has an infinite input resistance and a zero output resistance. It can be seen that the feedback network senses the output voltage vo and produces a feedback voltage signal, vf. For this feedback network, vo is the input voltage signal and vf is its output signal and hence the feedback network is functionally only a voltage amplifier with a low gain, called here as the feedback ratio !. The block representation is thus drawn based on the assumption that the voltage amplifier and the feedback network are ideal. This implies that: i. ii. iii. iv. the input resistance of voltage amplifier is infinity, the output resistance of voltage amplifier is zero, the input resistance of feedback network is infinity and the output resistance of feedback network is zero.

4.2

NONIDEAL NETWORK

When a voltage amplifier is nonideal, it has a large input resistance and a low output resistance. Since the feedback network functions as a voltage amplifier, its equivalent circuit or model is the same as that of a voltage amplifier and in addition, it should have a large input resistance and a low output resistance, like a voltage amplifier. That the feedback network has a large input resistance means that it draws negligible current from the output port of the voltage amplifier in the forward path and that it does not load the voltage amplifier. That the feedback network has a low output resistance means that the error that can result from its non-zero output resistance is less and often negligible.
30

Even though the feedback network functions as a voltage amplifier, it must be borne in mind that it is usually a passive network and that we usually associate the word 'amplifier' with a circuit that contains an active device. The equivalent circuit of the series-shunt topology with these nonideal parameters is shown in Fig. 4.2. It is worth remembering that a practical circuit will have other limitations such as limited frequency response, drift in parameters due to ageing, ambient temperature , but these aspects have been ignored here.

4.3

EFFECTS OF FEEDBACK

The principal aim of using feedback is to make the gain of the closed-loop system to be dependent on the feedback ratio ! and to free it from its dependence on the gain of the amplifier in the forward path. As outlined earlier, the closed-loop system has larger bandwidth. In addition, feedback has the effects stated below for a series-shunt topology. (i) The input resistance of the closed-loop system is far greater than the input resistance Ri of the voltage amplifier. (ii) The output resistance of the closed-loop system is far lower than the output resistance Ro of the voltage amplifier.
31

4.4

RESULTS OF ANALYSIS OF THE CIRCUIT WITH FEEDBACK

The circuit in Fig. 4.2 can be reduced to the block diagram shown in Fig. 4.3. From Fig.4.3, the closed-loop gain T is expressed to be:

where A is the gain of the forward path. It is shown in section 4.5.1 that the gain A of the forward path is:

The value of A is not normally much less than G, the gain of the voltage amplifier. The input resistance Rx of the closed-loop network is obtained to be:

It is seen that Rx is much larger than Ri, the input resistance of the voltage amplifier since the magnitude of loop gain is much greater than unity. The derivation of expression (4.3) for Rx is shown in section 4.5.2. The output resistance Ry of the closed-loop network is obtained to be:

32

It is seen that Ry is much smaller than Ro. The derivation of expression (3.4) for Ry is shown in section 4.5.3. 4.5 DERIVATIONS

4.5.1 Obtaining an expression for Gain of the forward path

The circuit for obtaining the gain A of the forward path is shown in Fig. 4.4. This gain is the gain that would exist were there no feedback. This means that the dependent source representing the feedback ratio in Fig. 4.2 is to be bypassed. The bypass arrangement should ensure that there is a path for current Is shown in Fig. 4.2 , and that the dependent source !Vo does not affect Is. This condition is met if the dependent source !Vo is replaced by a short-circuit. The circuit that results by replacing !Vo by a short-circuit is shown in Fig. 4.4. We define A as

From Fig. 4.4, we get that

While the dependent source !Vo is replaced by a short-circuit, its output resistance R3 is left in the circuit. (Why ?) 4.5.2 Input Resistance with Feedback The input resistance is the resistance seen by the source at its terminals. If the
33

source, as shown in Fig. 4.2, has an internal resistance Rs which is normally referred to as the source resistance, then the input resistance Rx is defined to be

where Vx is the voltage at the terminals of the source and Is is the current supplied by the source. Please note that the source voltage Vs and the voltage Vx at its terminals are not one and the same. Normally Vx would be slightly less than Vs in magnitude. The input resistance Rx is obtained to be:

where Rif is the input resistance of the circuit used for obtaining the forward gain A. It is fairly easy to remember the above expression. Draw the circuit for obtaining the forward gain, as shown in Fig. 4.4. Then Rif is the input resistance of that circuit which includes Rs too. This resistance gets magnified (1+A!) times and the value of the source resistance Rs is subtracted from the product Rif(1+A!) to obtain Rx, the input resistance with feedback. For a circuit with negative feedback, the real part of loop gain (-A!) is negative, which means that the product (A!) has a positive real part. It means that if A has a positive real part, then ! also has a positive real part and that if A has a negative real part, ! also has a negative real part. The derivation of an expression for Rx is as follows. From Fig. 4.2, we can state that

Since Rx is just the ratio (Vx/Is), we can get an expression for Rx if we can express Ve and Vo in terms of Is. From Fig. 4.2, we find that

In addition, we have that

Here the load on the output port is Rp, the parallel value of load resistance
34

RL,and the input resistance R2 of the feedback network. Substitute for Ve and Vo in equation (4.8) from equations (4.9) and (4.10) and then obtain the ratio of Vx /Is.

Usually Rp is far greater than Ro. Then

In addition, the approximation that Rx ! Ri(1 + G!) is usually valid. It may be seen from Fig. 4.2 that R3 is the output resistance of the feedback network. The feedback network functions as a voltage amplifier, but its output resistance need not be negligible since a passive network is used in the feedback path. On the other hand, the input resistance Ri of the voltage amplifier is large and in reality, it by itself should be much greater than R3. In addition, the term (1 + G!) is also much greater than unity. Hence

It is seen that negative feedback has increased the input resistance seen by the source. Equation (4.11) expresses the input resistance in terms of the voltage gain G of the voltage amplifier. An alternative expression for Rx can be obtained in terms of the forward gain A, the feedback factor ! and the other resistors. From equation (4.5), we get that

Therefore, based on equation (4.14), equation (4.11) becomes

A more compact expression can be obtained as follows.

35

It can be seen that Rif is the input resistance of the circuit in Fig. 4.4 which has been used for obtaining the gain of the forward path. 4.5.3 Output Resistance with feedback

An ideal voltage amplifier has zero output resistance. In practice, it has a small and finite output resistance, which is denoted as Ro in Fig. 4.2. The output resistance is the resistance of the circuit seen by the load. That is, we can replace the load resistance by a voltage source. The resistance as seen by this source with the input source Vs removed is usually defined as the output resistance. By definition, output resistance Ry is:

For the sake of convenience, the output resistance is defined here to be (Vo/Io) with Vs = 0, where the load resistor is replaced by a source,say Vo. The circuit that results is shown in Fig. 3.5 and the expression obtained for Ry is:

36

It is easy to remember the above equation. It is known that the output resistance gets reduced to feedback. It means that the output conductance gets magnified due to feedback. Obtain the output conductance of the circuit used for getting the forward gain. With RL in circuit, the output conductance (1/Rof) of circuit in Fig. 4.4 is

The above expression is obtained with the input source Vs replaced by a shortcircuit. When Vs has zero value, the dependent source GVe also has zero value and can hence be replaced by a short-circuit. To obtain the output conductance of the circuit with feedback, multiply (1/Rof) by (1 + A!) and remove from the product the conductance 1/RL due to load resistor. The expression for Ro can be derived as follows. From Fig. 4.5,

The ratio of Vo/ Io gives Ro. Then if can express Ve in terms of Vo, an expression for Ro can be obtained. From Fig. (3.5), we get that

Replace Ve in equation (4.20) by its expression in equation (4.21). On rearranging, we get that

Usually Ri >>(Rs + R3). Hence (Ri +Rs + R3) is nearly equal to Ri. Also R2, the input resistance of the feedback network tends to be far greater than Ro/(1 + G!). Thus

37

It is seen that the output resistance with feedback is much smaller than the output resistance of the voltage amplifier, since G! 1. It is preferable to find an alternative expression for output resistance Ry in terms of the forward path gain A. From equation (4.5), we get that

From equations (4.22) and (4.24), we get that

Since Rp = (R2 || RL), the above equation can be stated as:

We can get a more compact expression by using equation (4.19). Then

4.5.4 Closed-Loop Gain Next an expression for the closed-loop gain is obtained. Let the closed-loop gain be T. The expression for T shown in equation (3.1) is obtained as follows. From Fig. 3.2,

Substitute for Is in the expression for Ve. Then

38

On using equation (3.2) and re-arranging, we get the expression (4.1) for closed-loop gain. 4.6 SUMMARY

Given a circuit which has the series-shunt topology, first obtain its forward gain A, and then obtain the input resistance Rif and the output resistance Rof the circuit drawn for obtaining the gain A. From the values of A and !, we can get the closed-loop gain T. With a practical circuit, the analysis is slightly more complicated. The given circuit has to be reduced to the form in Fig. 4.2 before we can apply the results derived in this lesson.

39

5 SHUNT-SERIES TOPOLOGY
5.1 IDEAL NETWORK

The block diagram of a circuit with shunt-series feedback topology is shown in Fig. 5.1. It can be seen that the error current Ie , obtained as the difference between Is and If, is amplified by the current amplifier. The feedback signal If is obtained from the load current Io.

The ideal network with this topology can be represented by an equivalent circuit as shown in Fig. 5.2. This topology is the dual of series-shunt topology considered in the previous section.This topology uses a current amplifier in the forward path.
40

The feedback network also has a current signal as its input and output signal. Ideally, i. the input resistance of a current amplifier is zero, ii. the output resistance of current amplifier is infinity, iii. the input resistance of feedback network is zero and iv. the output resistance of feedback network is infinity. Since the feedback network provides a current signal derived from another current signal, it behaves essentially as a current amplifier and it also has the same ideal parameters. Again it is worth reiterating that the feedback may have only passive components and no active device. Hence its gain, shown as the feedback ratio !, is less than unity. In practice, the source, the current amplifier and the feedback network are not ideal and the equivalent circuit representing the nonideal shuntseries topology is shown in Fig. 5.3.

5.3

EFFECTS OF FEEDBACK

Due to feedback, the closed-loop circuit has reduced sensitivity to changes in gain of the current amplifier, increased bandwidth and less distortion. In addition, the feedback leads to: i. ii. 5.4 a large reduction in input impedance and a large increase in output impedance.

RESULTS OF ANALYSIS OF THE CIRCUIT WITH FEEDBACK

The closed-loop gain of the circuit in Fig. 5.3 is obtained to be :

41

where A is the gain of the forward path. The value of A is obtained to be:

The value of A is not usually very much less than G, the gain of the current amplifier. The input resistance of the circuit with feedback is obtained to be:

From equation (4.3), it can be seen that Rx is much smaller than Ri. The output resistance of the circuit with feedback is found to be equal to:

It can be seen from equation (4.4) that Ry Ro. 5.5 DERIVATIONS

42

5.5.1 Obtaining an expression for the gain of the forward path As shown in equation (4.2), the gain A of the forward path is obtained subject to the condition that If = 0. It means that the dependent current source !Io is replaced by an open-circuit. The circuit that results is shown in Fig. 4.4. From Fig. 4.4, we get that

5.5.2 Input Resistance with Feedback An expression for input resistance Rx with feedback is shown in equation (5.2). That the input resistance Rx gets reduced means that the input conductance increases. This input conductance, 1/Rx can be expressed as follows. Let 1/Rif be the input conductance of the circuit used for obtaining A. From Fig. 5.4,

Then

Now an expression for the input resistance of the circuit in Fig. 5.3 is shown below. By definition,

where Vs is the voltage across the current source Is . Note that Ix is not the same as Is. From Fig. 4.3,

Therefore,
43

Usually, the output resistance of a dependent current source, Ro here, is large. We can state that Ro >> R2 and that Ro >> RL. In addition, R3 >> Ri /(1+G!). Hence

Equation (5.9) can be expressed using A instead of G, where A is the forward current gain of the network ignoring the effects of feedback. From equation (5.2), we get that

From equations (5.9) and (5.11), we get that

5.5.3 Output Resistance with Feedback A current amplifier has a large output resistance and due to feedback, the closedloop circuit will have a much larger output resistance. The resistance of the circuit as seen from the load terminals is the output resistance. The expression for the output resistance in equation (5.4) can easily be remembered this way. Let Rof be the resistance of the path for flow of current Io in the circuit drawn for obtaining current gain of the forward path. In Fig. 4.3, if Is be a steady value, GIe is a dependent current source and to circuit external to it, it appears to have a resistance of Ro, its output resistance. If Rof be the resistance of the path for Io, then from Fig. 5.3,

44

Usually the output resistance is defined with the input source removed. To get a circuit that can be used for evaluating the output resistance, replace the load resistor RL by a voltage source and the input current source is replaced by an open circuit. That is, the output impedance is evaluated with Is = 0. From Fig. 5.5,

Since Ro R2 and (1/Ri) (1/R3 + 1/Rs), we can state that

45

Equation (4.15) can be expressed in terms of A instead of G. From equation (4.2), we get that

From equations (4.15) and (4.17), we get that

It is seen that the effect of feedback is to change the input and output resistance of current amplifier such that they tend to reach their ideal value. An expression for closed-loop current gain can be obtained now. 5.5.4 Closed-Loop Gain Closed-loop gain T is defined as

We can derive an expression for T from Fig. 5.3. From Fig. 5.3,

and

46

From equations (5.2), (5.18), (5.19) and (5.20), we get that

5.6

SUMMARY

Given a circuit which has shunt-series topology, we first obtain the gain of its forward path, called A. Then we can get expressions for its input resistance, output resistance and closed-loop gain.

47

6. SOME EXAMPLES
An operational amplifier is a versatile device, because its properties are close to the ideal values. This lesson outlines some applications of operational amplifiers. 6.1 INTEGRATOR

6.1.1 Circuit Operation The circuit of an integrator is shown in Fig.6.1. The switch across capacitor is opened up at t= 0. Then vC(0) = 0 V. With the opamp being treated as ideal, we get the current through R as

Also,

vC (t ) =

1 1 i. dt + v C ( 0) = v S . dt ! C RC !
48

since vC(0) = 0 V. From Fig. 10.1, we get that vo(t) = - vC(t). Figure 10.2 sketches the input-output relationship of an ideal opamp. In terms of Laplace transforms,

6.1.2 Effect of input bias current

Given an opamp with npn transistors at the input stage, there would be some input current into both the inverting and non-inverting terminals. This current is known as the input bias current. The effect of this current can be explained with the help of circuit in Fig. 6.3. Let the current flowing into the inverting terminal be i- . This current flows through capacitor and the output of opamp would reach its highest positive value some time after switching on the circuit. In order to reduce the effect of bias current, a resistor can be connected across C.

49

For the circuit in Fig. 6.4, the transfer function is given by

At frequencies below the corner frequency given by (1/R2C), we can approximate the transfer function as:

At frequencies above the corner frequency, the transfer function is:

The Bode plot of the transfer function for this circuit is shown in Fig. 6.5. An opamp with FETs at its input stage can behave differently and this discussion may not be relevant to BIFET opamps.
50

It is to be noted that the value of resistor across C is to be much greater than R1. Then the circuit starts operating as an integrator from a relatively low frequency. The integrator circuit is a basic building block and has many applications, such as in filters and waveshaping circuits.. It can also be used as an integral controller in a closed loop system. It was one of the principal blocks in the analogue computer. Now that we have studied what an integrator is, the next step is to find out how a differentiator can be built and analysed. Unlike an integrator, the use of differentiator is not common. 6.2 DIFFERENTIATOR

A circuit that works well as a differentiator under ideal conditions is shown in Fig. 6.6a. For this circuit, the transfer function is:

This circuit is not suitable for practical use because gain of the circuit increases with frequency. A noise signal contains high frequencies mostly. Hence noise-tosignal ratio in its output is too high. This drawback is remedied by the addition of a resistor in series with the capacitor as shown in Fig. 6.6b. Frequency response of this circuit is shown in Fig. 6.6c. For the circuit in Fig. 6.6b,the transfer function is:

51

At frequencies below (1/R2C), the circuit behaves as a differentiator. Above this frequency,this circuit has a fixed gain. It is to be noted that R1 >> R2.

6.3 OPAMP CIRCUITS USING DIODES AND ZENERS 6.3.1 Precision Diode

Figure 6.7 shows the circuit of a precision diode. When Vs is positive,

It is seen that Vo is less than Vs by (V2/Ao) only, where Ao is the open loop gain of opamp. If only a diode be used, the output will be less by a diode drop which is approximately 0.7 V, but here the output is only 1 or 2 mV less than Vs. When Vs is negative, the diode is reverse-biased, and the output is at zero volts. 6.3.2 Full-wave Rectifier Circuit Figure 6.8 shows a full-wave rectifier circuit using opamps. This circuit is known as a precision rectifier circuit. From Fig. 6.8, it is seen that

When Vs is positive, the inverting input terminal of first opamp tends to be

52

positive and its output becomes negative. Then Vx = - Vs and V1 = - Vx - 0.7, assuming that the voltage drop across diode D1 in conduction is 0.7 V with D2 remaining reverse-biased. Under these conditions, Vo = Vs.

Let Vs be negative. The inverting input terminal of first opamp tends to be negative and its output becomes positive. Then V1 = 0.7 V, the diode drop and Vx = 0 V. Now D2 conducts and D1 is reverse-biased. Under these conditions, Vo = Vs. Hence we have that Vo = |Vs |. If diodes D1 and D2 are reverse-connected, Vo = - |Vs |. 6.4 WORKED EXAMPLES Indicate how each circuit in Fig. 6.9 works.

WE 6.4.1 Solution:

Circuit (a) : Current amplifier or current-controlled current source.

Circuit (b) : Transconductance amplifier / voltage-controlled current source

53

Circuit (c) :Transresistance amplifier / current-controlled voltage source

_________________________________________________________________ WE 6.4.2: Figure 6.10 shows an application for difference amplifier. Resistor Rf can be a transducer the quiescent value of which is R2. The transducer here could be a strain gauge, the resistance of which changes about R2 as it is bent or twisted. It can be a thermistor or a magnetic field-dependent resistor. Find the output of circuit in Fig. 6.10 when Rf = R2 + !R. Solution: From Fig. 6.10,

54

______________________________________________________________ WE 6.4.3 Get an expression for output of circuit in Fig. 6.11.

Solution: From Fig. 6.11,

Eliminating Va, we get that

55

______________________________________________________________ WE 6.4.4: Find the CMRR of the circuit in Fig. 6.12.

Solution: For the circuit in Fig. 6.12,

We can express the output as:

By comparing the coefficients of V1 and V2, we get that Adm = 1/1.01 and Acm = 0.02/1.01. Then CMRR = Adm/Acm = 50. Comment: This problem can be re-phrased as follows: A difference amplifier is constructed with four resistors of same nominal value. If these resistors have a tolerance of 1%, find out the worst-case CMRR of this circuit. __________________________________________________________________
56

EXAMPLE 6.4.5: Obtain the transfer function of the circuit shown below.

SOLUTION: From the circuit in Fig. 6.13, we get that

We know V1 in terms of Vo. Then I2 can be expressed in terms of Vo. We can express I1 and I4 in terms of Vo. We then get
57

For example, let

Y1 =
Then

1 1 1 , Y3 = , Y4 = , Y = sC 2 and Y5 = sC5 R1 R3 R4 2

With the components as specified, the network functions as second-order low-pass filter. The above equation can be presented as follows.

58

Then

For the circuit in Fig. 6.15, the transfer function is obtained as:

Vo (s ) = VS (s )

s2 s 2 + s( C1 + C3 + C 4 ).

C1 C4 1 1 + C 3 C 4 R 5 C 3C 4 R 2 R 5
59

The above transfer function represents a second-order high-pass filter. Let

Then

A Matlab program is used to obtain the response. The program is presented below. % Second-order High Pass Filter clear; Ao=10; df=0.6; for n=1:175; af(n)=(1.05^n)/100; num(n)=Ao*(af(n)*af(n)); den1(n)=1-(af(n)*af(n)); den2(n)=2*df*af(n); mag(n)=num(n)/(den1(n)+j*den2(n)); phase(n)=180/pi*angle(mag(n)); end; subplot(2,1,1) loglog(af,abs(mag)) ylabel('Magnitude') grid on; subplot(2,1,2) semilogx(af,phase) ylabel('Phase angle') xlabel('normalized angular frequency') grid on; The response obtained have been presented next.

60

10

Magnitude

10

10

-2

10

-4

10 0 Phase angle -50 -100 -150

-2

10

-1

10

10

10

-200 -2 10

10

-1

10 10 norm alized angular frequency

10

__________________________________________________________________

EXAMPLE 6.4.6: The opamp used for the non- inverting amplifier has an open-loop gain with a single pole. It is defined as:

61

A( jw ) =

Ao jw 1+ wH

Obtain the transfer function for the inverting amplifier and sketch its frequency response. SOLUTION: The gain for the non-inverting amplifier has been obtained as:

Now the above expression becomes:

Vo = VS 1+

1+ 1+ j

R2 R1

w wH " R % $1 + 2 ' Ao # R1 &

" R % If ( A o ) >> $ 1 + 2 ' , then R1 & #

R2 R 1+ 2 Vo R1 R1 = = , w T = A ow H VS R2 % R2 % w " w " 1+ j ' $1 + ' 1+ j $1 + A ow H # R1 & wT # R1 & 1+

62

The frequency response of the non-inverting amplifier is sketched in Fig. 6.17.

_________________________________________________________________

63

7ACTIVE FILTERS
The commonly used types of filter responses are : Butterworth and Chebyshev. The Butterworth Filter response is characterized by a flat frequency upto the critical frequency, followed by a smooth roll-off of 20 db per decade per pole. The phase shift, however, varies nonlinearly with frequency. This means that different frequency components will experience different time delays as they pass through the filter. This will cause distortion and ringing on square and pulse waves. For pulse waves it is better to use Bessel filter response, which has a linear variation of phase with frequency . Bessel filters have a somewhat slower initial roll-off than Butterworth filters and consequently poorer for linear applications such as audio circuits. The Chebyshev filter response has a faster initial roll-off than a Butterworth filter, allowing the design of a much sharper cutoff filter with the same number of poles. The price of this faster roll-off is increased nonlinear phase shift and ripples in the amplitude response of the filter passband. The filter function H(s) can be expressed as:

H ( s) =
7.1

A( s ) B( s )

( 71 .)

BUTTERWORTH POLYNOMIALS

The first-order normalized Butterworth polynomial for the denominator B(s) can be expressed to be:
B( s ) = ( s + 1)

The corresponding normalized transfer function is:

H ( s) = 1 s +1 ( 7.2 )

The above transfer function can be realized by the circuit shown in Fig. 7.1

64

For the circuit in Fig. 7.1, the transfer function is H ( s ) =

1 sRC + 1

( 7.3)

In order to obtain the normalized transfer function, we can let R = 1 and C = 1 F. To design a first-order low-pass filter with a cut-off frequency at 1 kHz,

RC =

1 50 . Let R = 10 k. Then C = 2 1000

nF

The normalized Butterworth polynomial B(s) is corresponding transfer function is:

(s

+ 2s + 1 .

The

H ( s) =

1 ( 7.4) 2s + 1

The 2-pole filter circuit is shown in Fig. 7.2.

65

For the circuit in Fig. 7.2, the transfer function is:

H ( s) = 1 ( 7.5) s R C1C2 + sRC2 + 1
2 2

To obtain the normalized transfer function, let

Rn = 1 , C1n = 2 F and C2 n =

1 F 2

The suffix n indicates that these are normalized values. To design a unity-gain low pass filter with a cut-off frequency at 1000 Hz, the components can be selected as follows. For design, two scale factors have to be defined. Let frequency scaling constant be Kf.
K f = wc = 2 f c = 2000

Divide the capacitor values by this scaling constant. The capacitor and the resistor values still do not lie in the acceptable range for an opamp circuit. Let us define another constant, known as the impedance scaling constant, Km. Let Km = 10000. Then each of the resistors has the value of 10 k. Also

C1 = C2 =

C1n = 22.5 nF , KmK f C2 n = 1125 . nF . Km K f

Choose the nearest commercially available values. The Butterworth polynomial for a 3-pole filter is: B(s) = (s + 1) (s2 + s + 1) =s3 + 2s2 + 2s + 1. The corresponding transfer function is:
66

H ( s) =

1 ( 7.5) s + 2s + 2s + 1
3 2

The 3-pole filter circuit is shown in Fig. 7.3.

The transfer function for the circuit in Fig. 3 is:

H ( s) = s + (2T1 T3 + 2T2 T3 )s 2 + (3T3 + T2 )s + 1
3

( 7.6)

where
T1 = RC1 , T2 = RC2 , T3 = RC3 .

To normalize, let
R = 1 , C1 = 3546 . F , C2 = 1392 . F , C3 = 0.2024 F . ( 7.7)

Example: Design a unity-gain Butterworth filter with a critical frequency of 3000 Hz and a roll-off of 60 db per decade. Solution: The frequency scaling constant is: K f = 2f C = 2 3000 = 18,849.6 rad / s After applying the frequency scaling constant, we get that
R = 1 , C1 = 18812 . F , C2 = 73848 . F , C3 = 10.738 F .

To make C1 be equal to 10 nF, the amplitude scale factor is 18,812. With this
67

scale factor,
R = 18 k, C1 = 10 nF , C2 = 3926 . nF , C3 = 560 pF .

_______________________________________________________________ 4-th order Butterworth polynomial B(s) = (s2 + 0.765s + 1).(s2 + 1.848s + 1) The corresponding normalized transfer function is:

H ( s) =

1 s 2 + 0.765s + 1 . s 2 + 1848 . s +1

)(

( 7.8 )

The circuit for 4-pole filter is presented below in Fig. 7.4

68

The transfer function for the above circuit is:

H ( s) =

1 s 2 T1T2 + sT2 + 1 . s 2 T3T4 + sT4 + 1

)(

( 7.9)

where

T1 = RC1 , T2 = RC2 , T3 = RC3 , T4 = RC4 .

The transfer function can be normalized if:
R = 1 , C1 = 1082 . F , C2 = 0.9241 F , C3 = 2.614 F , C4 = 0.3825 F

( 710 . )

Example Design a unity-gain low-pass Butterworth filter with a critical frequency of 15 kHz. The attenuation should at least 300 at 20,000 kHz.

Solution: In Fig. 7.5, let fc = critical frequency, fd = decade frequency, 10 fc, fy = another frequency.
69

Let a = attenuation at frequency, fd b = attenuation at frequency, fy Then

f fc a = d b f y fc
For the given problem, b = 300, fc = 15 kHz, fd = 150 kHz, fy = 20 kHz. Then a = 8100 In terms of dB, attenutation in dB per decade = 20 log (8100) = 78.17 dB. It means that a roll-off of at elast 80 dB decade is required. A 4-pole active filer will suffice. For a 4-pole filter, we have that
R = 1 , C1 = 1082 . F , C2 = 0.9241 F , C3 = 2.614 F , C4 = 0.3825 F

We have that
K f = 2f C = 2 15000 = 94,247.8 rad / s

Then we get that

R = 1 , C1 = 11,480 F , C2 = 9,805 F , C3 = 27,725 F , C4 = 4,0585 F

Let the amplitude scale factor be 11480. Then

R = 1148 . k, C1 = 10 nF , C2 = 8.541 nF , C3 = 2415 . nF , C4 = 3535 . nF

Choose the nearest values. _______________________________________________________________

70

CHEBYSHEV FILTERS For a Chebyshev Filter

H ( jw ) =
2

( H 0 )2
2! w $ 1 + 2 Cn # & " wC %

! w$ where Cn # & is " wC % ! ! w$ ! w $$ w Cn # & = cos# n.cos 1 # & & for 0 1 wC " wC % " wC % % " ! ! w$ ! w $$ w Cn # & = cos# n.cosh 1 # & & for >1 wC " wC % " wC % % "

The normalized polynomials for low pass filter are defined as follows.

s +1 s 2 + 1425 . s + 1516 . ( s + 0.626) s 2 + 0.845s + 0.356

(s

+ 0.351s + 1064 s 2 + 0.845s + 0.356 .

)(

CONCLUSION High-pass filters, band-pass filters and notch filters can also be realized using the well-known filter configurations. _______________________________________________________________ Reference: 1. Microelectronics, Millman and Grabel, Second Edition. MacGraw-Hill. 2. Linear Integrated Circuits, Winzer,Saunders College Publishing 3. Active Filter Design, Alan Waters, Macmillan Education _______________________________________________________________
71