2017 ASEE Mid Atlantic Section

Spring Conference: Morgan State University, Baltimore, Maryland Apr 7 Paper ID #20849

A Detailed Lesson on Operational Amplifiers - Negative Feedback

Dr. Nashwa Nabil Elaraby, Pennsylvania State University, Harrisburg, The Capital College
Dr. Elaraby is a faculty at Penn State Harrisburg since August 2014. She received her PhD degree in
Electrical and Computer Engineering from Temple University in 2014. She received her B.Sc and M.Sc.
in Electrical Engineering from Alexandria University, Egypt.
Her research interests include digital logic design using Field Programmable Gate Arrays for massively
parallel data computations, electronic circuit design, and neuronal data processing for Brain Machine
Interface applications.

c American Society for Engineering Education, 2017

 A Detailed Lesson on Operational Amplifiers - Negative Feedback
Nashwa Elaraby, Ph.D., Penn State Harrisburg

Abstract: Operational Amplifiers present one of the important topics in electronic circuits
courses. Although they are widely taught, the model for the operational amplifiers with negative
feedback configuration is incomplete in most of the electronics textbooks1 2 3 4. The closed loop
gain for the non-inverting amplifier is given as AoL/(1+βAoL)1, while the closed loop model for
the inverting amplifier is usually not included. Instead nodal analysis is used to derive the
expression for the inverting amplifier gain. Using different derivation paths and models causes
confusion for the students, as the comparison is not provided from the same point of view. The
constant gain-bandwidth product is mentioned for the closed loop frequency response for the
amplifier, but it is not clearly stated that it only applies to the non-inverting amplifier
configuration1. The missing details usually cause misconception for the students and mismatch
between the lab results and their expectations. The main objective of this paper was to provide
instructors with a detailed lesson on operational amplifiers using negative feedback, that can be
applied in electronic circuit analysis courses for electrical engineering students and electrical
engineering technology students.

I. Introduction:

Operational amplifiers with negative feedback have three modes of operation. The non-inverting
amplifier, the inverting amplifier and the voltage follower. The voltage follower can be
considered a special case of the non-inverting amplifier, but it will be considered separately in
this paper, as it is a commonly used buffer circuit that the students need to closely understand.
The gain for the different amplifiers can be derived using the negative feedback configuration.
The main idea of the negative feedback is to reduce the potential difference between the
inverting and the non-inverting amplifier inputs to obtain a voltage at the output that is confined
within the boundaries of the dc voltages biasing the internal transistors of the operational
amplifier to operate in the forward active mode. The second section of the paper introduces the
methodology used to complete the modeling details of the operational amplifier circuit with
negative feedback. It covers the three negative feedback models and derives the expressions for
the exact calculations of the closed loop gains in terms of the limited open-loop gain of the
operational amplifier. The derivation of the ideal closed loop gains is verified using the model. It
also discusses the open-loop frequency response of the operational amplifier. The methodology
section also provides a transfer function equation for a compensated operational amplifier model,
and shows how it is represented by a Bode plot equation. This step shows the students how to
algebraically prove the roll off rate on the Bode plot. Lab procedures are included to test and
verify the frequency response of a compensated operational amplifier. The closed loop frequency
response is then proved algebraically for both the inverting and the non-inverting configurations,
and lab procedures are included to test them. The third section provides the overall evaluation
results of the applied lesson and the last section provides a discussion on the paper content.

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU

I. Methodology:

(a) Negative Feedback Models:

The negative feedback configuration aims at minimizing the potential difference vd between the
inverting and non-inverting inputs of the operational amplifier, and hence applying it as a linear
amplifier with output values limited between the biasing voltages of the internal circuitry. With
the negative feedback and an ideal open-loop gain AoL of ∞, we can assume that the differential
voltage vd is zero. For an ideal operational amplifier, the input impedance is infinity, which
means that there is no current flowing into the inputs of the operational amplifier. The closed
loop gain for the operational amplifier can hence be derived using simple nodal analysis as
shown in figure 1.

Virtually at vin
vd  0
v o  v in
A cL  1

Kirchoff ' s Current Law :

iR2 i R1  i R 2
Virtually at vin
0  v in v in  v o

R1 R2
 R 
iR1 v o  1  2   v in
 R1 
 R 
A cL  1  2 
 R1 

Kirchoff ' s Current Law :

iR2 i R1  i R 2
Virtually at ground
v in  0 0  v o

R1 R2
 R2 
iR1 v o      v in
 R1 
 R2 
A cL    
 R 1 

Fig. 1 The negative feedback configurations of an ideal operational amplifier: the voltage follower, the non-
inverting operational amplifier and the inverting operational amplifier. The derivations of the closed loop gains
assume an infinite open-loop gain AoL and an infinite input impedance of the ideal operational amplifier.

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU

For a non-ideal operational amplifier with finite open-loop gain the following models in fig. 2
can be applied to derive the closed loop gain. The addition operator signs are determined based
on the connection to the inverting or non-inverting terminals. The negative feedback connection
strives to cancel out the input signal contribution to minimize the differential voltage vd. For the
non-inverting amplifier the input is connected to the non-inverting input while the positive
feedback signal is fed into the inverting input of the operational amplifier1. For the inverting
amplifier the feedback signal and a fraction of the input signal are fed into the inverting input,
but as the output is negative, the inverting input adds the absolute value of the feedback signal to
the negative input signal. The α and β are voltage dividers found by applying the superposition
theorem for vin and vo respectively.

A oL
A cL 
1  A oL 

A oL
A cL 
1  A oL 
where
 R1 
   
 R1  R 2 

A oL
A cL 
1  A oL 
where
 R2 
   
 R1  R 2 
 R1 
   
 R1  R 2 

Fig. 2 The negative feedback model of (a) the voltage follower, (b) the non-inverting operational amplifier and (c)
the inverting operational amplifier.

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU

In the table 1, a thorough derivation of the closed loop gain expressions is presented.

The voltage follower The non-inverting amplifier The inverting amplifier

v o  A oL  v d v o  A oL  v d v o  A oL  v d
v o  A oL  v in  v f  v o  A oL  v in  v f  v o  A oL   v in  v f 
v o  A oL  v in  vo  v o  A oL  v in  v o  v o  A oL   v in  v o 
v o 1  A oL   A oL  v in v o 1  A oL   A oL  v in v o 1  A oL   A oL  v in
vo A oL vo A oL vo  A oL
  
v in 1  A oL  v in 1  A oL  v in 1  A oL 

Table 1: The derivation of the closed loop gain of the different negative feedback configurations using their models.

To verify the models used for the negative feedback configurations the open-loop gain can be set
to infinity for the ideal operational amplifier, and the derived closed loop expressions will be
compared to the ones obtained using nodal analysis. In the derivation steps we neglect the 1 term
with respect to βAoL assuming that with the infinite values of AoL, the term βAoL >> 1.

The voltage The non-inverting The inverting

follower amplifier amplifier

 R   R 
The ideal AcL A cL  1 A cL  1  2  A cL    2 
 R1   R1 
A oL A oL A oL
A cL  A cL  A cL 
The closed loop 1  A oL  1  A oL  1  A oL 
gain based on where where
the negative
 R1   R2   R1 
feedback model            
 R1  R 2   R1  R 2   R1  R 2 
A oL A oL A oL
A cL  A cL  A cL 
Verifying the 1  A oL  1  A oL  1  A oL 
closed loop gain  A oL  
A oL A oL 1
model A cL  1 A cL   A cL  
A oL A oL  A oL 
R1  R 2   R2   R1  R 2 
      
R1  R1  R 2   R1 
 R   R 
 1  2     2 
 R1   R1 

Table 2: Comparison of the negative feedback expressions obtained using the negative feedback model and the
nodal analysis assuming ideal operational amplifier open-loop gain and input impedance.

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU

(b) Open-Loop Frequency Response of a Compensated Operational Amplifier:
For this section the lab procedures for measuring the open-loop frequency response and the unity
gain bandwidth of the LM741 operational amplifier will be included. The main objective of this
lab is to let students realize that the operational amplifier does not have a constant gain for all
frequencies, instead it is behaving like a low pass filter with a dominant pole, governed by the
following equation:

A oL 0
A oL f  
 
1  j f 
 f pD 

where A oL 0  The low frequency open - loop gain
f pD  the dominant pole frequency

At lower frequencies, it will be very hard to measure the gain, as the operational will reach
saturation at the output for very low amplitudes of input voltages due to the high gain. As the
frequency increases, the gain decreases, and it will be possible to obtain an undistorted output
voltage signal. The students will measure the gain and phase shift at different frequencies
between 10KHz and 500KHz. If any non-symmetrical clipping is observed on the output signal,
positive or negative DC offset voltage in the range of a few mV can be added to the input signal
to cancel out the offset voltage causing the clipping. The range of frequencies examined is much
higher than the dominant pole frequency (~10Hz), hence the magnitude and the phase equations
can be approximated to the following formulas. It is important to measure the gain based on
undistorted signals.

A oL 0  A oL 0  A oL 0 
A oL f     f   Phase angle
2 2  f 
 f   f     f 
1      f pD  f   tan 1    90
 f pD   f pD     f pD 
     
 f 
A oL f  dB  20 log A oL 0  20 log  
 f pD  This term sets the roll-off rate at -20dB/decade
 
on the Bode plot applying logarithmic scale for
the frequency

The students then plot | AoL(f) |dB versus the frequency using a logarithmic scale for the
frequency. On the graph, the data points are extrapolated to find the frequency fT at which the
gain is unity or 0dB. To get the dominant pole frequency, the graph is extrapolated to the
frequency at which the gain reaches the maximum value of the LM741 gain of 107dB.

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU

The following MATLAB code can be used to plot the data points on the Bode plot and
extrapolate it to obtain fT and fpD.

Unity gain
frequency

Dominant pole cut-off frequency

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU

(c) Closed-Loop Frequency Response:

The operational amplifier with negative feedback also behaves as a low-pass filter. To find its
bandwidth, the feedback models presented in section II will be applied. The closed-loop gain
equations derived will be implemented by substituting for the open-loop gain AoL by the function
AoL(f), and the comparing the equation to the general low-pass filter transfer function.
The voltage follower:
 
 
 oL
A 0  
 f 
1 j 
A oL f   f  A oL 0 A oL 0
 
A cL f  
pD
  
1  A oL f    1  j f  A 0  
  1  A oL 0  1  j
f 
 A oL f  
oL
f pD 
 f pD 1  A oL 0 
1 
f 
 1 j 
 f pD 

A cL 0  A oL 0
A cL f   where A cL (0)  B.W. voltage follower  f pD 1  A oL 0
f 1  A oL 0
1 j
f pD 1  A oL 0

The non-inverting Amplifier:

 
 
 A oL 0 
 f 
1 j 
A oL f   f pD  A oL 0 A oL 0

A cL f     
1  A oL f     A oL 0 1  A 0  1  j 
f
  1 j f 
 A f   f pD oL  f 1  A 0  
1     oL   pD oL 
f
1 j 
 f 
 pD 
A cL 0 A oL 0
A cL f   where A cL (0)  B.W. non -inverting  f pD 1  A oL 0
f 1  A oL 0
1 j
f pD 1  A oL 0

For the voltage follower and the non-inverting amplifier the Gain-Bandwidth product is a
constant value:

A cL 0  B.W.  A oL 0  f pD

 f pD   B.W. 
A cL 0  A oL 0   
 AcL 0 dB  A oL 0 dB - 20  log  
 B.W.   f pD 
 

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU

The inverting Amplifier:

 
 
    A oL  0  
 f 
 1  j 
 A oL f   f pD     A oL 0     A oL 0 
A cL f     
1  A oL f    1  j f  A 0  
        1  j f 
 A oL f  
oL 1 A 0
1 
f pD oL 
 f pD 1  A oL 0  

f 
 1  j 
 f pD 
   A cL 0     A oL 0 
A cL f   where A cL (0)  B.W. non -inverting  f pD 1  A oL 0 
1 j
f 1  A oL 0 
f pD 1  A oL 0 

For the voltage follower and the non-inverting amplifier the Gain-Bandwidth product is a
constant value:

A cL 0   B.W.    A oL 0   f pD

The gain-bandwidth product is not a constant value. It depends on the forward attenuation α,
which varies for every closed loop gain.

Lab procedures:

The lab procedures for the closed loop frequency gain are divided into two main sections: the
closed loop gain for the non-inverting amplifier and the closed loop gain for the inverting
amplifier. Each section is subdivided into two parts, one showing that the operational amplifier
with the negative feedback is behaving as a low-pass filter, and then calculating the GBW
product for different gain values.

In the discussion section of the lab report, students compare the frequency responses of the two
configuration of negative feedback. They compare the GBW products. They also compare their
frequency responses to that of the open-loop configuration.

They plot the closed-loop low frequency gain versus the frequency on a logarithmic scale and
explain the roll-off rate using the constant GBW product relation for the non-inverting amplifier.
They show that this relation does not hold for the inverting amplifier.

The set of labs explaining the different negative feedback models and the frequency response,
help students build a concrete foundation for the understanding of the operational amplifier
application as a linear amplifier in the different modes.

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU

Part 1: Closed-loop Frequency Response of a Non-Inverting Amplifier:

(1) Connect the circuit shown in figure. Verify the gain as the ratio of Vout/Vin using a low
frequency sinusoidal signal of an appropriate amplitude. (avoid clipping, but remember
that the larger the signal the more accurate your recordings will be)

(2) What is the feedback attenuation β?

(3) Measure the gain at different frequencies between 100Hz and 1MHz, and plot the closed-
loop frequency response. Plot the graph (Take at least 20 readings). Measure the
bandwidth by gradually raising the frequency until the gain drops to 0.707 times its
nominal value (The nominal value AcL(0) is the value of the gain at very low
frequencies). The cut-off frequency needs to be one of you recordings.

(4) For any non-inverting amplifier circuit, the gain of the amplifier is
A oL
A CL  .
1  A oL 
Sketch the model for the non-inverting amplifier, and prove the equation for the closed-
loop gain. At high values of A oL the closed loop gain can be approximated to be 1/β.
Prove that, and show how it applies to your circuit results.

(5) Using the transfer function of the operational amplifier

A oL (0)
A oL (f ) 
 
1  j f 
 f pD 
 

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU

 prove that the bandwidth for the non-inverting amplifier is f c  f PD 1  AoL(0) .

(6) Prove that the Gain-Bandwidth Product (GBW) for the non-inverting amplifier is a
constant value equal to GBW  f pD  A oL (0) .The GBW value also represents the Unity
gain Band Width f T , which is the bandwidth at a closed loop gain of 1 (Voltage
follower).

(7) Use different values of Rf to obtain closed-loop gains from 20dB to 50dB with a step size
of 5dB. For each value of the low-frequency closed-loop gain, chose an appropriate
feedback resistor value, and verify gain by calculating the ratio of Vout/Vin., using a low
frequency sinusoidal signal of an appropriate amplitude (avoid clipping at the output) For
each gain value, measure the bandwidth by gradually raising the frequency until the gain
drops to 0.707 times its nominal value.

(8) Use your measurements to create a graph of the gain (in dB) on the y-axis versus the
bandwidth (logarithmically scaled) in Hz on the x-axis.

f c  A cL  f pD  A oL(0)
 f pD 
A cL:  A oL(0)   
 f 
 c 
 f pD 
 
20 log A cL:   20 log A oL(0)  20 log 
 f 

 c 
 f 
A cL dB  A oL(0)  20 log  c 
dB  f pD 
 
(9) On your graph, extrapolate your data points to the location where gain would be unity
(zero dB). At this point, the unity gain bandwidth will be equal to the value of the GBW.
Then extrapolate your graph to the point where the gain would be 220,000 (107dB). At
that point, the bandwidth should be that of the 741 in an open-loop configuration.

(10) Connect a voltage follower and increase the frequency until the gain becomes 0.707 to
find the unity gain BW fT. What might be your limitation for the gain measurement at
the high frequency?

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU

Part 2: Closed-loop Frequency Response of an Inverting Amplifier:

(1) Connect the circuit shown in figure. Verify the gain as the ratio of Vout/Vin using a low
frequency sinusoidal signal of an appropriate amplitude. (avoid clipping, while the larger
the signal the more accurate will be your recordings)

(2) What is the feedback attenuation β and the multiplier α?

(3) Measure the gain at different frequencies between 100Hz and 1MHz, and plot the closed-
loop frequency response. Plot the graph (Take at least 20 readings). Measure the
bandwidth by gradually raising the frequency until the gain drops to 0.707 times its
nominal value (The nominal value is the value of the gain at very low frequency). The
cut-off frequency needs to be one of your recordings.

(4) For an inverting amplifier circuit, the gain of the amplifier is

 A oL
A CL  .
1  A oL 
Sketch the model for the non-inverting amplifier, and prove the equation for the closed-
loop gain. At high values of A oL the closed loop gain can be approximated to be -α/β.
Prove that, and show how it applies to your circuit results.

(5) Using the transfer function of the operational amplifier

A oL (0)
A oL (f ) 
 
1  j f 
 f pD 
 
prove that the bandwidth for the non-inverting amplifier is f c  f PD 1  AoL(0) .

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU

 (6) Prove that the Gain-Bandwidth Product (GBW) for the inverting amplifier is equal to
GBW    f pD  A oL (0) .

(7) Use different values of Rf to obtain closed-loop gains from 20dB to 50dB with a step size
of 5dB. For each value of the low-frequency closed-loop gain, chose an appropriate
feedback resistor value, and verify gain by calculating the ratio of Vout/Vin., using a low
frequency sinusoidal signal of an appropriate amplitude (avoid clipping at the output) For
each gain value, measure the bandwidth by gradually raising the frequency until the gain
drops to 0.707 times its nominal value.

(8) Use your measurements to create a graph of the gain (in dB) on the y-axis versus the
bandwidth (logarithmically scaled) in Hz on the x-axis.

f c  A cL    f pD  A oL ( 0)
A cL:  f pD 
 A oL ( 0)   

  fc 
 f pD 
A 
 
20 log  cL:   20 log A oL ( 0)  20 log  

    fc 
 f 
A 
 
20 log  cL:   20 log A oL ( 0)  20 log  c 
    f PD 

(9) The GBW for the non-inverting amplifier is a constant value. What about the inverting
amplifier? Explain.

III. Results:

I have taught circuits courses covering the operational amplifier application as linear amplifier
for both Electrical Engineering and Electrical Engineering Technology student for at least three
years both in the spring and fall semesters. With the detailed modeling of the operational
amplifier circuits and emphasizing the difference between the inverting and non-inverting
configurations, I had much better results for the exams and quizzes. The lab experiments go
faster and are successfully completed with more than 95% of success rate compared to 60% in
the past when we relied on the incomplete theoretical explanation and models of the textbooks.
I have asked students whether it was easier to study from the textbook or from the class notes,
and they answered that the detailed comparison of the different negative feedback configurations
provided in the class notes, made it easier for them to remember their differences and
similarities. Before adding the detailed lesson to my class notes, many students were applying
the constant GBW product for both the inverting and the non-inverting amplifier configurations,

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU

and they were frustrated when the lab results were not aligned with their calculations. With the
detailed explanation the rate of these instances has significantly decreased.

IV. Conclusion:

Modeling the negative feedback configurations of operational amplifiers and the frequency
response of the amplifier circuits is a topic that needs clarification in most of the textbooks. A
detailed comparison of the three configurations: Non-inverting amplifier, inverting amplifier and
voltage followers are needed to give the students a concrete theoretical basis to understand the
circuit behavior. The main objective of this paper was to provide instructors with a detailed
lesson on operational amplifiers, that can be applied in electronic circuit analysis courses for
electrical engineering students and electrical engineering technology students. It presents a
thorough lesson with lab procedures on the negative feedback models as well as their frequency
responses, which has improved student's understanding of the operational amplifier behavior and
modeling.

References:

1. Donald A. Neaman, Microelectronics: Circuit Analysis and Design, 4th Edition McGraw-Hill 2010,
ISBN 978-0-07-338064-3

2. Basic Operational Amplifiers and Linear Integrated Circuits, 2nd edition, Floyd and Buchla, Prentice Hall, 1999

3. Adel S. Sedra and Kenneth C. Smith Microelectronic Circuits, Oxford Series in Electrical & Computer
Engineering, 6th Edition

4. Richard C. Jaeger, Travis N Blalock, Microelectronic Circuit Design, 4th edition, McGraw-Hill 2011,
ISBN 978-0-07-338045-2

Spring 2017 Mid-Atlantic ASEE Conference, April 7-8, 2017 MSU