ECE3204 Microelectronic Circuits II D2013

Lab 2

Unity Gain Buffer

Op-Amp Nonidealities: Slew Rate, Output Current Limit
Complementary Emitter Follower Power Amplifier
Frequency Dependence: Low-Pass Filter, Integrator
Objective
One purpose of this lab is to investigate the use of an op-amp to provide buffering: voltage gain
is low (unity), but power gain is high since the op-amp provides much more output current than
it draws from the input source. Along the way we will encounter some nonideal performance
features: slew rate and output current limiting. To overcome the output current limitation, we
will investigate the use of external emitter follower transistors to provide even more current gain
when driving very low impedance loads. Although the emitter follower by itself distorts the
signal severely, this effect can be removed by closing the op-amp feedback loop around the
distorting stage.
Finally, we will also examine op-amp circuits with frequency-dependent gain: the active lowpass filter and the integrator.
Prelab
P1.

A 5V peak, 1kHz sinusoidal source has a 100k output impedance. This source must
drive a 1k load with no loss of signal.
Try the direct approach of Figure 2.1:
MODEL OF SOURCE

MODEL OF SOURCE

Rs
100k

vo

100k
+

vs

Rs

vo

RL
1k

vs
-

Figure 2.1

YOUR
DESIGN
HERE!

RL
1k

Figure 2.2

P1.1

How well will this work? What is the "gain" from vs to vo?

P1.2

Design a circuit to meet the requirement of vo = vs, as shown in Figure 2.2. RL and RS
remain as in Figure 2.1. You may assume 15V supplies are available.

P2.
P2.1

Slew Rate Limiting.

The gain-bandwidth product ("Bandwidth" in the data sheet) for the LM741 op-amp is
about 1.0 MHz. Based on this figure, what should be the closed loop 3-dB bandwidth of
the follower circuit shown in Figure 2.3?

P2.2

Suppose the source in Figure 2.3 is the drive waveform to an ultrasonic transducer: a
100kHz sine wave with a 5V peak amplitude:
vs = (5V)sin[2(100kHz)t]
Based on the expected 3-dB bandwidth from 2.1, and assuming that the op-amp operates
in its linear region, what should the output amplitude be? Sketch this waveform.

P2.3

What is the maximum slope (dV/dt) of the output waveform you sketched in 2.2? How
does this compare with the slew rate limit of 0.5 V/sec for the LM741 op-amp? How
would this affect the waveform you would see in the lab?

P2.4

At a peak amplitude of 5V, what is the highest frequency sine wave the LM741 could
process without distortion due to slew rate limiting?

* C1, C2: "BYPASS

CAPACITORS"
IMPORTANT FOR HIGH
FREQUENCY (1 MHz
PERFORMANCE

+15V
C1*
0.1F
3

7
6

+5V

vS

-5V

C2*
0.1F

f = 100kHz

-15V

Figure 2.3

Vout

P3.

Complementary Emitter Follower ("Push-Pull") Power Amplifier

Suppose we have a requirement to drive a 1V peak audio signal into a speaker. The
audio signal can be represented as a 1kHz sine wave and the speaker as an 8 load
resistance, as shown in Figure 2.4.

P3.1

Suppose we try to use the LM741 in an op-amp follower as shown in Figure 2.4. What is
the peak current required by the load from the op-amp output? How does this compare
with the LM741 maximum output current of approximately 25mA? What would you
expect for the actual output voltage if you built this circuit in the lab?
+15V
C1*
0.1F
3

7
6

+1V

vS

--1V

C2*
0.1F

f = 1kHz

-15V

Fig. 2.4

Vout
RL
8

P4.

Integrator
We have seen that the gain of the inverting configuration can be written as -Z2/Z1, where
Z1 and Z2 are impedances which are functions of complex frequency s. This allows us to
make "active filters:" op-amp circuits with a gain that varies with frequency.
Z2

C
R2

vI

R1

2
6
3

vI

Vo

R1

2
6
3

LM741

Figure 2.5

Vo

LM741

Figure 2.6

Figure 2.5 shows a simple low-pass active filter.

P4.1

Write an expression for the transfer function H(s)=vo/vi in terms of R1, R2, C, and
complex frequency s. Then, let s=j and sketch a Bode plot of the magnitude |H(j)| vs.
. What is the 3dB bandwidth of this low-pass filter?

P4.2

At low frequencies: what does vo/vi approach as s 0?

P4.3

At high frequencies: what does vo/vi approach as s ?

P4.4

At what frequency does the behavior "cross over" from the low frequency limit to the
high frequency limit?
Figure 2.6 shows an integrator.

P4.5

Write a frequency domain expression for H(s)=vo/vi in terms of R1, C, and complex
frequency s. Let s=j and plot the magnitude |vo/vi| vs. .

P4.6

Over what range of frequencies does the lowpass filter transfer function from 4.1 "look
like" that of the ideal integrator?

P4.7

Write a time domain expression for the integrator output in terms of R1, C, and time t.

Lab
1.

Construct the circuit of Figure 2.1 with a 5V, 1kHz sinusoidal source.

1.1

Verify the "gain" from vs to vo for the direct approach that you calculated in prelab sec.
P1.1. How well does this approach meet the requirement vs = vo?

1.2

Verify your design for a circuit to meet the requirement of vo = vs, as shown in Figure
2.2.

2.

Slew Rate Limiting.

Construct the circuit of Figure 2.3. Adjust the signal generator voltage vs until the opamp input is a 100kHz sine wave of amplitude 5V peak.

2.1

Observe the op-amp output voltage vout. How does it compare with the output you
expect if the op-amp is in its linear region?

2.2

Reduce the generator frequency until the vout sine wave is undistorted. How does this
frequency compare with your calculation from prelab P2.4?
Note how much lower this frequency is than the maximum frequency limit imposed by
the gain-bandwidth product. For large output signals and moderate gains, op-amp
frequency performance is often limited by the maximum slew rate, not the gainbandwidth product!

2.3

To measure the effect of slew rate on the rise and fall times, adjust the signal generator so
that vs is a 10 kHz square wave with a 5V peak-to-peak amplitude. Now measure the
rise time at vout. Using the bandwidth predicted in prelab P2.1, how well does this rise
time conform to the BW x tr = 0.35 relationship? (Hint: it should be way off!). This
should illustrate that BW x tr = 0.35 holds only for linear systems, since it was derived
assuming an exponential step response. When something nonlinear happens (such as
slewing or clipping) that distorts the exponential response of a linear system, all bets are
off.

2.4

Measure the slew rate in both directions (positive going and negative going). How do
your results compare with the data sheet specification of 0.5 V/sec?

3.

Complementary Emitter Follower "Push-Pull" Power Amplifier

Construct the circuit of Figure 2.4. To simulate the 8 speaker load resistance, use a
parallel combination of 10||51. The input should be a 1kHz sine wave of peak
amplitude 1V.

3.1

Observe the output waveform. How does it compare with the input vs? How does it
compare with what you expected from prelab P3.1?

3.2

Based on your observation of the output waveform, what are the output current limits
(source and sink) of your LM741 op-amp? Compare with the data sheet specification of
approximately 25mA.

Remembering that the bipolar transistor acts as a current amplifier, you build the circuit
in Figure 2.7. The BJTs operate as emitter followers providing most of the load current
demanded by RL. We will examine the behavior of the circuit when the op-amp feedback
is taken from different points P and Q in the emitter follower stage.
Note: be very careful in wiring this circuit. Double-check your connections; pay
attention to the pin assignment diagram in Figure 2.7. Turn off power before making any
wiring changes. All resistors (except RL) are for protection of the op-amp and/or
transistors.

2N3904, 2N3906

+15V
33

vS

2N3904

3
6

Q
100
2N3906

FEEDBACK

33
-15V
Figure 2.7
6

Vout
RL
8

3.3

Using three channels of your oscilloscope, look at the input vS, the op-amp output at
point P, and the overall output vout at point Q. Use the same scale on all three (consider
0.5V/div) and align the 0V reference levels on the oscilloscope display to help you
visualize the voltages in the circuit relative to each other.
With feedback to the op-amp inverting input connected to point P, observe vout when vs
is the 1V peak, 1 kHz sine wave. Record the waveforms at the op-amp output (point P)
and vout. Is there any distortion at vout? Is it different from the distortion observed in
part 3.1? Explain.

3.4

Turn off the power, and change the feedback connection to point Q. Again, observe and
record the waveforms at the op-amp output (point P) and vout. How has circuit operation
changed? Explain.

Optional: use a small speaker in place of RL to listen to the waveforms in each case. Is there
any audible difference?

4.

Integrator
Construct the low-pass active filter of Figure 2.5, with R1 = 1k, R2 = 100k, and C =
1000pF. The 1000pF capacitor has a value code of "102".

4.1

Measure the 3dB bandwidth, using an input sine wave of amplitude 50mV peak. How
does this compare with the bandwidth predicted for these values of R1, R2, and C from
your analysis in prelab P4.1?

4.2

Set the input to be a square wave of frequency 100Hz and amplitude 100mV. Is the
circuit acting more like a lowpass filter or more like an integrator?

4.3

Gradually increase the square wave frequency to 10kHz, keeping the amplitude 100mV.
At what frequency does the behavior "cross over" from the low frequency limit to the
high frequency limit?

4.4

At 10kHz, is the circuit acting more like a lowpass filter or more like an integrator? How
does the output compare to what you would expect from an ideal integrator, based on the
result of your prelab P4.7?

4.5

In an effort to make the circuit act like an integrator at all frequencies, remove resistor R2
to obtain the circuit of Figure 2.6. What happens to the op-amp output?
The problem is that a real op-amp has small DC errors at its inputs. Even though the
errors are small, an integrator will faithfully integrate them until the op-amp output
saturates. The dominant error may be either the input offset voltage or the input bias
current:

Another way of understanding the problem is stated as "an integrator has infinite gain at
DC" - which makes sense since the integrator transfer function is H(j) = 1/jRC and
DC corresponds to =0. This means that even though the op-amp DC errors are small
(of order mV, nA) -- no matter how small they are, it's just a matter of time until the
integral of the error exceeds the op-amp's output range.
Methods of dealing with this problem depend on the application. If the circuit needs to
function as an integrator only above a certain frequency, then the topology of Figure 2.5
can be used. If the integrator must work down to DC, then the integrating capacitor in
the op-amp feedback of Figure 2.6 can be periodically reset (shorted) before the op-amp
saturates. Another option is to enclose the integrator in a larger negative feedback loop to
compensate for op-amp DC errors with an external correction.

Lab Writeup
Organize your lab writeup in sections.
Be sure to (at a minimum) answer any questions posed in this lab handout. Additionally, if any
other insights come to you in the course of your analyzing and thinking about your data, discuss
those as well.
Feel free to use screen shots of the oscilloscope to illustrate your measurements.
See the Sample Lab Writeup for general tips on writeup presentation style.