EPC Notes-Part-4
EPC Notes-Part-4
MODULE 4
NEGATIVE FEEDBACK
The amplifier in which a part of output is sampled and fedback to the input of the amplifier is
called feedback [Link] at the input ,we have two signals:input signal and part of
the output which is fed back to the input.
Both the signals may be inphase or out of [Link] input signal and part of output signal
are in phase,the feedback is called positive [Link] the other hand ,when they are out of
phase ,the feedback is called negative feedback.
Use of positive feedback results in oscillations and negative feedback gives stability and hence
used in amplifiers.
Negative feedback is used to stabilize the amplifier against variations in component values and
parameters of the active devices used in the [Link] feedback reduces
distortion,reduces noise output ,improves frequency response and improves input and output
resistances.
Basic ideas
75
• Since Zout approaches zero,the output side of VCVS is a stiff voltage source to any
practical load resistance.
76
• Figure 4.2 shows a non inverting [Link] opamp has an open loop voltage gain
AVOL,typically 100,000 or more.
• Because of the voltage divider, part of the output voltage is fedback to the inverting
input.
• The feedback fraction B of any VCVS circuit is defined as feedback voltage divided
by the output voltage.
𝑣2
i.e 𝐵 = … .1
𝑣𝑜𝑢𝑡
• The feedback fraction is also called the feedback attenuation factor because it
indicates how much the output voltage is attenuated before the feedback signal
reaches the inverting input.
• The closed loop voltage gain is given by
𝐴𝑣𝑜𝐿
𝐴𝑣(𝐶𝐿) = ….2
1+𝐴𝑣𝑜𝐿𝐵
77
78
Loop gain
• The second term in the denominator of equation 2 AvoLB is called the loop gain because
it is the voltage gain of the forward and feedback paths.
• The loop gain is very important valkue in the design of negative feedback am[plifier.
• In any practical design,the loop gain is made very large. The larger the loop gain,the
better,because it stabilizes the voltage gain and has an enhancing or curative effect on
quantities such as gain stability,distortion,offsets,input impedance and output
impedance.
Ideal closed loop voltage gain
• For VCVS to work well,the loop gain AVOLB must be much greater than unity.
• When the designer satisfies this condition equation 2 becomes:
𝐴𝑣𝑜𝐿 𝐴𝑣𝑜𝐿
𝐴𝑣(𝐶𝐿) = 𝐴𝑣 = ≅
1+ 𝐴𝑣𝑜𝐿𝐵
1 𝐴𝑣𝑜𝐿𝐵
Or 𝐴𝑣 ≅ …..3
𝐵
This ideal equation gives almost exact answers AVOLB≫ [Link] exact closed-loop voltage
gain is slightly less than this ideal closed loop voltage gain.
• The percentage error between the ideal and exact values can be calculated by
(100%)
%𝐸𝑟𝑟𝑜𝑟 = …4
1+𝐴𝑣𝑜𝐿𝐵
For instance,if 1+AVOLB is 1000(60dB),the error is only 0.1 [Link] means that the exact
answer is only 0.1percent less than the ideal answer.
Using the ideal equation
The feedback fraction is
79
𝑣2 𝑅1
𝐵= =
𝑣𝑜𝑢𝑡 𝑅1 + 𝑅𝑓
Ideal open loop gain is
1 𝑅1 + 𝑅𝑓 𝑅𝑓
𝐴𝑣 ≅ = = +1
𝐵 𝑅1 𝑅1
OTHER VCVS EQUATIONS
Negative feedback stabilizes the voltage gain;that is,it almost eliminates the internal opamp
variations and makes the closed loop voltage gain dependent primarily on external resistances.
Negative feedback in VCVS amplifier increases the input impedance, decreases the output
impedance and reduces any nonlinear distortion of the amplified signal.
Gain stability
The gain stability depends on having a very low percent error between the ideal and the exact
closed loop voltage [Link] smaller the percent error ,the better the [Link] worst case
error of closed loop voltage gain occurs when the open loop voltage gain is minimum.
100%
%𝑀𝑎𝑥𝑖𝑚𝑢𝑚 𝑒𝑟𝑟𝑜𝑟 = …5
1+𝐴𝑣𝑜𝐿𝑚𝑖𝑛𝐵
Where AVOL(min) is the minimum or worst case open loop voltage gain shown on the data
sheet,with a 741C, AVOL(min) =20,000
For instance, if 1+ AVOL(min)B=500
100%
%𝑀𝑎𝑥𝑖𝑚𝑢𝑚 𝑒𝑟𝑟𝑜𝑟 = = 0.2%
500
The closed loop voltage gain of any VCVS amplifier will be within 0.2percent of the ideal
value.
Closed loop input impedance
Figure 4.3a shows a non inverting [Link] exact equation for the closed loop input
impedance of the VCVS amplifier
80
Zin(CL)= (1 + 𝐴𝑣𝑜𝐿𝐵)𝑅𝑖𝑛||RCM….6
Where Rin=the open loop input resistance of the opamp
RCM=the common mode input resistance of the opamp
Rin is the input resistance shown on a data [Link] RCM is ignored because it is large,and
equation 6 is approximated by
Zin(CL)= (1 + 𝐴𝑣𝑜𝐿𝐵)𝑅𝑖𝑛
Since 1 + 𝐴𝑣𝑜𝐿𝐵 is much greater than unity in practical VCVS amplifier,the closed loop
input impedance is extremely [Link] a voltage follower,B is 1 and Zin(CL) would approach
[Link] ulitimate limit on the closed loop input impedance is
Zin(CL) =RCM
The exact value of closed loop input impedance is not [Link] is important is that it is
very large,usually much larger than Rin but less than the ultimate limit of RCM
Closed loop output impedance
The exact equation for the closed loop output impedance is
𝑅𝑜𝑢𝑡
Zout(CL)= …7
1+𝐴𝑣𝑜𝐿𝐵
Where Rout is the open loop output impedance of the opamp shown on the [Link]
output resistance of a 741C is 75Ω.
Since 1+AvOLB is much greater than unity in practical VCVS amplifier,the closed loop output
impedance is less than 1 Ω and may approach zero in a voltage follower.
Non linear distortion
• One more improvement worth mentioning is the effect of negative feedback on distortion.
• In the later stages of amplifier ,nonlinear distortion will occur with large signals becausethe
input/output response of the amplifying devices becomes non [Link] instance,thenon
linear graph of the base emitter diode distorts a large signal by elongating the positive
half cycle and compressing the negative half cycle as shown in fig 4.3(b)
81
82
• Looking back at this circuit ,the two CE stages produce an open loop voltage gain equal
to
AVOL=AV1*AV2
• The output drives the voltage divider formed by rf and [Link] the bottom of re is at
ac ground ,the feedback fraction is approximately:
𝑟𝑒
𝐵≅
𝑟𝑒 + 𝑟𝑓
• The input vin drives the base of the transistor,while the feedback voltage drives the
[Link] error voltage appears across the base emitter diode.
• The closed loop voltage gain is approximately 1/B.
• The input impedance is (1+AVOLB)Rin
• The output impedance is Rout/(1+AVOLB)
• The distortion is THDOL/(1+AVOLB
83
Output voltage
The exact equation for output voltage is
𝐴𝑣𝑜𝐿
Vout=− {𝑖𝑛𝑅𝑓 ∗ ( )}
1+𝐴𝑣𝑜𝐿
An easy way to remember equation 1 is to use the concept of a virtual [Link] inverting
input is a virtual ground to voltage not current. When you visualize a virtual ground on the
inverting input,you can see all of the input current must flow through the feedback resistor.
The magnitude of the output voltage is given by
Vout=-(iinRf)
The circuit is a current to voltage converter.
If Rf=1KΩ then an input of 1 mA produces an output of [Link] Rf=10KΩ the same input current
produces an output of 10V.
ICVS input and output impedances
In Figure 4.5 ,the exact equation for closed loop input and output impedances are
𝑅𝑓
Zin(CL)= … .1
1+𝐴𝑣𝑜𝐿𝐵
𝑅𝑜𝑢𝑡
Zout(CL)= ….2
1+𝐴𝑣𝑜𝐿𝐵
In both equations,the large denominator will reduce the impedance to a very low value.
The inverting amplifier
The inverting amplifier of figure 4.6 has a closed loop voltage gain of
𝑅𝑓
Av=− …3
𝑅1
This type of amplifier uses ICVS negative [Link] of the virtual ground on the
inverting input,the input current equals
𝑉𝑖𝑛
iin=
𝑅1
84
With a VCIS amplifier, an input voltage controls an output current. Because of the heavy
negative feedback in this kind of amplifier, the input voltage is converter to a precise value of
output current.
Figure 4.7 shows a transconductance amplifier. It is similar to a VCVS amplifier, except that
RL is the load resistor as well as the feedback resistor. In other words, the active output is not
the voltage across R1+RL;rather it is the current through RL This output current is stabilized;
that is ,a specific value of input voltage produces a precise value of output current.
In figure 4.6,the exact equation for output current is
𝑉𝑖𝑛
in practical circuit, the second term in the denominator is much smaller than the first,and the
equation simplifies to
𝑉𝑖𝑛
iout=
𝑅1
when you visualize a virtual short between the input terminals of figure 4.7,the inverting input
is bootstrapped to the noninverting [Link] all the input voltage appears across [Link]
current through this resistor is
i1=𝑉𝑖𝑛
𝑅1
since the input side of figure 4.6 is same as the input side of a VCVS amplifier,the approximate
equation for the closed loop input impedance of a VCIS amplifier is
Zin(CL) =(1+AvoLB)Rin
Where Rin is the input resistance of the [Link] stabilized output current sees a closed loop
output impedance of :
Zout=(1+AvoL)R1
In both equations a large AvoLincreases both impedances toward infinity,exactly what we want
for a VCIS [Link] circuit is almost perfect voltage to current converter because it has
very high input and output impedances.
THE ICIS AMPLIFIER
An ICIS circuit amplifies the input [Link] of the heavy negative feedback,the ICIS
amplifier tends to act like a perfect current [Link] has very low input impedance and a
very high output impedance.
Ai=𝐴𝑣𝑜𝐿(𝑅1+𝑅2)…1
𝑅𝐿+𝐴𝑣𝑜𝐿𝑅1
Usually the second term in the denominator is much larger than the first and the above equation
simplifies to
𝑅2
Ai ≅ 1 + ( )…2
𝑅1
The equation for the closed loop input impedance of an ICIS amplifier is
86
𝑅2
Zin(CL) = ..3
1+𝐴𝑣𝑜𝐿𝐵
ACTIVE FILTERS
IDEAL RESPONSES
The frequency response of a filter is the graph of its voltage gain versus [Link] are
five types of filters:low pass,high pass,bandpass,bandstop,and all-pass.
Low pass filter
Figure 4.9a shows the ideal frequency response of a low pass filter.A low pass filter passes all
frequencies from zero to the cutoff frequency and blocks all frequencies above the cut off
frequency.
With a low pass filter ,the frequencies between zero and the cut off frequency is called the
[Link] frequencies above the cutoff frequency are called the [Link] roll -off
region between the passband and the stopband is called the [Link] ideal low pass filter
has zero attenuation (signal loss) in the passband,infinite attenuation in the stopband.
87
88
For instance,if the cutoff frequencies are 450 and 460KHz,the bandwidth is”
𝐵𝑊 = 460𝐾𝐻𝑧 − 450𝐾𝐻𝑧 = 10𝐾𝐻𝑧
As another example,if the cutoff frequencies are 300 and 3300Hz,the bandwidth is
𝑓𝑜 = √𝑓1𝑓2
For instance,telephone companies use a bandpaqss filter with cutoff frequencies of 300 and
3300Hz to separate phone [Link] center frequency of these filters is
𝑓0 = √(300𝐻𝑧)(3300𝐻𝑧) = 995𝐻𝑧
The Q of a band pass filter is defined as the center frequency divided by the bandwidth
𝑓𝑜
𝑄=
𝐵𝑊
For instance ,if fo=200KHz and BW =40KHz then Q=5.
If Q is less than 1 ,the bandpass filter is called a wideband [Link] Q is greater than 1,the filter
is called narrow band filter.
• For example ,a filter with cutoff frequencies of 95 and 105KHz has a bandwidth of
[Link] is a narrow band because Q is approximately 10.
• A filter with cutoff frequencies of 300 and 3300Hz has a center frequency of
approximately 1000Hz and a bandwidth of [Link] is wideband because Q is
approximately 0.333.
Bandstop filter
Figure 4.9d shows the ideal frequency response of a bandstop [Link] type of filter passes
all frequencies from zero up to the lower cutoff [Link] it blocks all the frequencies
between the lower and upper cutoff [Link] passes all frequencies above the upper cutoff
frequency.
With a bandstop filter,the stopband is all the frequencies between the lower and upper cutoff
[Link] frequencies below the lower cutoff frequency and above the upper cut off
frequency are the passband.
The definitions for bandwidth,narrowband and center frequency are the same as [Link]
bandstop filter is sometimes called a notch filter because it notches out or removes all
frequencies in the stopband.
89
90
When the frequency increases above the cutoff frequency,the capacitive reactance decreases
and reduces the noninverting input [Link] the R1C1 lag circuit is outside the feedback
loop,the output voltage rolls [Link] the frequency approaches infinity ,the capacitor becomes a
short and there is zero input voltage.
Figure 4.10b shows another noninverting first order low pass [Link] has 2 additional
[Link] voltage gain is given by
𝑅2
𝐴𝑣 = 1 + ( )…3
𝑅1
Above the cutoff frequency,the lag circuit reduces the noninverting input [Link] R3C1
lag circuit is outside the feedback loop,the output voltage rolls off at a rate of 20dB per decade.
Figure 4.10c shows an inverting frist-order low-pass filter and its [Link] low frequencies
,the capacitor appears to be open and the circuit acts like an inverting amplifier with a voltage
gain of
𝑅2
𝐴𝑣 = − …5
𝑅1
As the frequency increases ,the capacitive reactance decreases and reduces the impedance of
the feedback [Link] implies less voltage [Link] the frequency approaches infinity,the
capacitor becomes short and there is no voltage [Link] shown in figure 4.10c,the cutoff
frequency is given by
1
𝑓𝑐 = … .6
2𝜋𝑅2𝐶1
91
Figure 4.10 first order low-pass stage (a) noninverting unity gain (b) noninverting with
voltage gain (c) inverting with voltage gain.
High-pass stage
Figure 4.11 a shows the first order high-pass active [Link] voltage gain is
Av=1…1
The 3dB cutoff frequency is given by
1
𝑓𝑐 = …2
2𝜋𝑅1𝐶1
When the frequency decreases below the cutoff frequency,the capacitive reactance increases
and reduces the noninverting input [Link] the R1 C1 circuit is outside the feedback
loop,the output voltage rolls off ,As the frequency approasches zero,the capacitor becomes an
open and there ios zero input voltage.
Figure 4.11 b shows another noninverting first order high pass filter,The voltage gain above
the cutoff frequency is given by
𝑅2
𝐴𝑣 = 1 + ( ) … 3
𝑅1
Below the cutoff frequency,the RC circuit reduces the noninverting input [Link] the
R3C1 lag circuit is outside the feedback loop,the output voltage rolls off at arate of 20dB per
decade.
Figure 4.11c shows another first order high-pass filter and its [Link] high frequencies,the
circuit acts like an inverting amplifier with a voltage gain of
𝑋𝑐2 𝐶2
𝐴𝑣 = − = − …5
𝑋𝑐1 𝐶1
92
As the frequency decreases,the capacitive reactance increases and eventually reduce the input
signal and the [Link] implies less voltage [Link] the frequency approaches zero,the
capacitor becomes open and there is no input [Link] 3dB cutoff frequency is given by
1
`𝑓𝑐 = …6
2𝜋𝑅1𝐶2
Figure 4.11 first order high-pass stages (a) noninverting unity gain (b) noninverting
with voltage gain (c) inverting with voltage gain
VCVS UNITY-GAIN SECOND ORDER LOW-PASS FILTERS
Second order or two pole stages are the most important because they are easy to build and
analyze. Higher order filters are usually made by cascading second order stages.
93
This section discusses the Shallen -key low-pass filters(named after the inventors) These filters
are also called as VCVS filters because the op-amp is used as a voltage controlled voltage
[Link] low-pass circuits can implement three of the basic approximations: Butterworth
,Chebyshev and Bessel.
Circuit implementation
Figure 4.12 shows a Shallen-key second order low-pass [Link] that the two resistors have
the same value.,but the capacitors are different. There is a lag circuit on the noninverting
input,but this time, there is feedback path through a second capacitor [Link] low frequencies
,both capacitors appear to be open,and the circuit has a unity gain because the opamp is
connected as a voltage follower.
Figure 4.12 second order VCVS stage for butterworth and Bessel
As the frequency increases ,the impedance of C1 decreases and the noninverting input voltage
[Link] the same time,capacitor C2 is feeding back a signal that is inphase with the input
[Link] the feedback signal adds to the source signal ,the feedback is [Link] a result
,the decrease in the non inverting input voltage caused by C1 is not as large as it woild be
without the positive [Link] larger C2 is with respect to C1,the more the positive
feedback.
Pole frequency
As shown in the figure 4.12
𝐶2
𝑄 = 0.5√ ….1
𝐶1
And
1
𝑓𝑝 = ….2
2𝜋𝑅√𝐶1𝐶2
The pole frequency (fp) is a special frequency used in the design of active filters.
In more complicated circuits, the pole frequency is given by
1
𝑓𝑝 = …3
2𝜋√𝑅1𝑅2𝐶1𝐶2
94
95
Figure 4.14 shows sallen second order low pass filter. Both resistors and both capacitors have
the same value. Hence the circuit is called a Sallen -key -equal-component filter. The circuit
has a midband voltage gain of
𝑹𝟐
𝑨𝒗 = 𝟏 +
𝑹𝟏
The operation of the circuit is similar to that of the Sallen-Key unity gain filter, except for the
effect of the voltage gain. Since the voltage gain can produce more positive feedback through
the feedback capacitor,the Q of the stage becomes a function of voltage gain and is given by
1
𝑄=
3 − 𝐴𝑣
Because Av can be no smaller than unity, the minimum Q is [Link] Av increases from 1 to
3,Q varies from 0.5 to infinity. Therefore the allowable range of Av is between 1 and [Link] we
try to run the circuit with Av greater than 3,it ill break into oscillations because the positive
feedback is too large. A butterworth filter has Q=0.707 and Kc=1.A Bessel filter has Q=0.577
and Kc=0.786.
For instance if the pole frequency is 2.5KHz and Kc=1.3,the cutoff frequency for the high
pass filter is
2.5𝐾𝐻𝑧
𝑓𝑐 = = 1.92𝐾𝐻𝑧
1.3
Figure 4.15 b shows the Sallen -Key equal component high -pass filter and its [Link]
the equations are the same as for a low pass [Link] positions od the resistors and
capacitors have been reversed.
96
Figure 4.15 second order VCVS high pass stages: (a) unity gain (b) voltage gain greater
than unity
MFB BANDPASS FILTERS
A bandpass filter has a center frequency and a [Link] basic equations for a bandpass
response:
𝐵𝑊 = 𝑓2 − 𝑓1
𝑓𝑜 = √𝑓1𝑓2
𝑓𝑜
𝑄=
𝐵𝑊
When Q is less than 1,the filter has a wideband [Link] this case a bandpass filter is
usually built by cascading a low pass stage with a high pass [Link] Q is greater than
1,the filter has a narrow band response.
Wideband filters
Suppose we want to build bandpass filter with a lower cutoff frequency of 300Hz and an
upper cutoff frequency of [Link] center frequency of the filter is
97
𝑓𝑜 995𝐻𝑧
𝑄= = = 0.332
𝐵 3𝐾𝐻𝑧
𝑊
Since Q is less than 1 ,we can use cascaded low-pass and high-pass stages,as shown in figure
[Link] high -pass filter has a cutoff frequency of 300Hz and the low pass filter has acut off
frequency of [Link] the two decibels are added ,we get a bandpass response with
cutoff frequencies of 300Hz and 3.3KHz.
Figure 4.16 wideband filter uses cascade of low pass and high pass stages
Narrowband Filters
When Q is greater than 1 ,we can use the multiple feedback (MFB) filter shown in figure
[Link],the input signal goes to the inverting input rather than the noninverting [Link]
,the circuit has two feedback path,one through a capacitor and another through a resistor.
At low frequencies,the capacitors appear to be [Link] the input signal cannot reach the
opamp,and the output is [Link] high frequencies ,the capacitors appear to be [Link] this
case the voltage gain is zero because ethe fedback capacitor has zero [Link] the
low and high extremes in frequency,there is a band of frequencies where the circuit acts lika an
inverting amplifier.
The voltage gain at the center frequency is given by
𝑅2
𝐴𝑣 = − ..1
2𝑅1
This is almost identical to the voltage gain of an inverting amplifier,except for the factor of 2
in the denominator. The Q of the circuit is given by
𝑅2
𝑄 = (0.5)√ …2
𝑅1
Which is equivalent to
𝑄 = 0.707√−𝐴𝑣…3
When using equation 3,the absolute value of the voltage gain will be [Link] instance,if Av=-
100
98
𝑄 = 0.707√100 = 7.07
Equation 3 tells us that greater the voltage gain,the higher the Q
The center frequency is given by
1
𝑓𝑜 = …4
2𝜋√𝑅1𝑅2𝐶1𝐶2
𝑄 = (0.5)√((𝑅2)/(𝑅1||𝑅3)….6
1
𝑓𝑜 = …7
2𝜋𝐶√𝑅1||𝑅3)𝑅2
The circuit has the advantage of higher input impedance because R1 can be made higher fir a
given Q.
99
1
𝑓𝑜 = …3
2𝜋𝐶√2𝑅1(𝑅1||𝑅3))
100
Figure 4.19 MFB stage with variable center frequency and constant bandwidth
Bandstop filters
Figure 4.20 shows a Sallen-key second -order notch filter and its analysis [Link] low
frequencies,all capacitors are [Link] a result ,all the input signal reaches the noninverting
[Link] circuit has a passband voltage gain of:
𝑅2
𝐴𝑣 = 1 + ( )…1
𝑅1
At vert high frequencies,the capacitors are [Link] all the input signal reaches the non
inverting input.
Between the low and high extremes in frequency,there us a center frequency given by
1
𝑓𝑜 =
2𝜋𝑅𝐶
The Q of the circuit is given by
𝑄= 0.5
2 − 𝐴𝑣
The voltage gain of a Sallen ket notch filter must be less than 2 to avoid oscillations.
101