Physics 160 Lecture 16

Transcription

Physics 160 Lecture 16
Physics 160
Lecture 16
R. Johnson
May 20, 2015
Behavior, Limitations of Real Op-amps
•
•
•
•
•
•
•
•
•
•
Input offset voltage
Input current, and input-current mismatch
– Design to minimize voltage offsets from this current
– Input offset current
See the data sheet for the
Temperature dependences
LF411 (posted on the course
Slew rate
web page)
page).
– e.g. correcting crossover distortion
Frequency range (gen. purpose op amps are not for high frequencies)
Voltage gain and phase shift
– Stability!
CMRR and Common-mode input range
Input impedance with negative feedback is extremely high
Output impedance with negative feedback is very low
– What is really relevant is the limitation on output current drive
Output range (how close can the output get to the supply rails?)
– Single-supply op amps; can the output slew all the way to zero?
May 20, 2015
Physics 160
2
Offset Voltage and Input Current
•
Offset voltage
– 2 mV max. for the 411
• This is the effective mismatch at the input. At the output it gets
multiplied
lti li d b
by th
the gain!
i !
– Trimming
+
V+
U1
3
7
V1
15
OS2
2
-
V-
OUT
OS1
6
1
R2
4
LM741
5
V2
– Temperature and time drifts
•
10k
15
I
Input
t currents
t (esp.
(
for
f BJTs)
BJT )
– Balancing resistors on inputs
– Input offset current
•
A/C Amp: reduce DC gain to unity to minimize offset effects
May 20, 2015
Physics 160
3
Slew Rate
•
411 can slew outputs up to 15 V/s with no capacitive load
– For a 1 V sine wave output, for example, this limits the frequency to
2.4 MHz to avoid “harmonic”
harmonic distortion.
– Recall the limits on using an op-amp to cure crossover distortion
2.4  10 6 Hz  2  1V  15  10 6 V/s
4
Q1
3
+
V+
V1
15Vdc
OUT
V3
1k
-
LM324
V
11
2
Q2N3904
1
V-
R1
VOFF = 0
VAMPL = .2
FREQ = 10000
U1A
Q2N3906
Q2
R2
10k
V2
15Vdc
May 20, 2015
Physics 160
Above ~100 Hz the
op-amp cannot slew
fast enough to cure
the cross-over
di t ti
distortion.
4
OP-AMP
OP
AMP STABILITY
May 20, 2015
Physics 160
5
Gain and Phase Shift
Open loop gain of 100,000
Gain (dB
B)
Dominant Pole
LM301A Op-Amp
(uncompensated)
Second Pole
Unity gain
Frequency
May 20, 2015
Physics 160
6
Stability Criterion
•
If the loop gain is >1 when the phase shift around the negative
feedback loop hits 180 degrees, then the circuit will oscillate.
– Your negative feedback becomes positive feedback at high
frequency!
– You probably won’t see the oscillation in a PSpice transient
simulation without doing extra work to introduce real-life effects,
plus some stimulus to initiate the oscillation. Don’t be fooled by
trusting your computer simulation too much!
A
B
•
Remember, the loop gain is AB, where A is the open-loop gain,
and B is the gain of the negative feedback.
– The follower has B=1, so that is usually the worst-case scenario for
stability (inverting amp usually has B<1)
• (A loop with B>1 is even more dangerous, hence the caution
about
b t putting
tti a capacitor
it across th
the ttransistor
i t iin th
the log-amp
l
feedback.)
May 20, 2015
Physics 160
7
Phase
LM301A Op-Amp
(uncompensated)
Gain ((dB)
The open-loop gain is >0 dB at 180
degrees phase shift, so this op-amp
will surely oscillate if used as a
follower without compensation.
p
This doesn’t include additional
phase shifts that might exist in the
feedback network!
Frequency
May 20, 2015
Physics 160
8
Compensated Op-Amp
Deliberately
D
lib t l enhance
h
the Miller effect to move
the dominant pole to
lower frequency
R
C
Miller effect at this gain
stage produces the
dominant pole
May 20, 2015
Physics 160
9
External Compensation
Advantage
Ad
t
off
external
compensation: you
don’t need to
compensate as
much if B<1, so you
can have higher
frequency
performance.
Cs=30 p
pF
According to this formula from the data sheet, to make a
follower with R1>>R2
follower,
R1>>R2, C1 needs to be at least equal to Cs.
Cs
May 20, 2015
Physics 160
10
Simulation of Follower Instability
Insignificant load,
load
like a ×10 scope
probe.
+
V+
V
U5
V1
OUT
C1
VS
2
-
V-
15Vdc
5
6
1
V
8
LM301A
4
V1 = 0
V2 = 0.2
TD = 0.1u
TR = 10n
TF = 10n
PW = 10u
PER = 10m
B
C2
3
7
Square wave input
CL
0.1pF
RL
10Meg
CC
1p
V2
R2
15Vdc
10Meg
Should be 30pF for compensation at unity gain (B=1).
May 20, 2015
Physics 160
11
Simulation of Follower Instability
This sortt off ringing
Thi
i i iin th
the
simulation should be a warning
that in real life it is probably
going
g
g to oscillate wildly!
y
May 20, 2015
Physics 160
12
Simulation of Follower Instability
With capacitive load CL on output increased to 1nF!
May 20, 2015
Physics 160
13
Gain (dB)
Phase
LM301a with 30pF Single-Pole Compensation
Rule of thumb: add enough capacitance
such that the loop gain drops to 0 dB at or
below the 3dB frequency of the 2nd pole.
I have p
plotted the forward g
gain,, but
remember that for a follower, the
loop gain equals the forward gain.
Frequency
May 20, 2015
Physics 160
14
Finite Gain Effects and Bandwidth
D tto internal
Due
i t
l compensation.
ti
For a single pole, the gain falls as 1
over the frequency.
The bandwidth (f3dB) times
the gain (G) will always equal
a constant gain-bandwidthproduct (fT).
)
May 20, 2015
Physics 160
15
Pure Differentiator
R1
100k
V2
+
V+
U1
3
7
15Vdc
OUT
2
V4
May 20, 2015
B1
5
6
1
V
LF411
4
1Vac
1V
0Vdc
Rises 6dB/octave!
Remember that this
circuit is unstable at
high frequency. 15Vdc
.01uF
-
V-
C2
Voltage gain = jRC
B2
This is a good example of
instability that occurs when
negative feedback becomes
positive feedback at high
frequency.
V1
0
Physics 160
16
Spice AC-sim: gain rising at 6dB/octave
(20 dB/decade)
dB/d
d )
90 phase shift from principal pole
411 open-loop gain
Gain
n (dB)
The curves would cross
at about 35kHz. At that
point the gain must turn
over, but also, there will
then be an additional
90 p
phase shift from the
feedback network.
May 20, 2015
Frequency
Physics 160
17
Phase Shifts Around the Loop
This low-pass filter gives a
90 degree phase shift!
= 270 at low f
= 360 at high f (>35 kHz) R1
100k
Break the loop here.
C2
May 20, 2015
B1
3
+
V+
OUT
0
U1
B2
1
6
5
= 180 at low f
= 270
270 at high f (>35 kHz)
7
The source
looks like a low
impedance to
ground.
V-
=0
LF411
2
-
4
.01uF
0
Physics 160
18
Phase (d
(deg)
Spice AC-sim up to high frequency:
Gain
n (dB)
Resonance at
about 35 kHz!
May 20, 2015
Frequency
Physics 160
19
Transient Simulation with Square-Wave Input
R1
100k
V2
+
V+
U1
3
7
15Vdc
OUT
2
V5
B1
5
6
1
V
LF411
4
V1 = 0
V2 = 0.1
TD = 50n
TR = 1n
TF = 1n
PW = 0.5m
PER = 1m
-
V-
C2
.01uF
B2
V1
15Vdc
May 20, 2015
0
Physics 160
20
Transient Simulation with Square-Wave Input
Input
34 kHz oscillation
O tp t
Output
May 20, 2015
Physics 160
21
Add a 1k Resistor in Series with the Cap
R1
100k
V2
2
V5
May 20, 2015
B1
6
1
V
LF411
This resistor
keeps
p the cap
p
from shifting the
phase by a full
90 when going
around the loop.
V1
15Vdc
-
5
4
V1 = 0
V2 = 0.1
TD = 50n
TR = 1n
TF = 1n
PW = 0.5m
PER = 1m
.01uF 1k
B2
OUT
R2
V-
C2
+
V+
V
U1
3
7
15Vdc
0
Physics 160
22
Phase (deg)
Low-Pass Filter with Extra Resistor
Gain (dB)
G
Never gets
g
all the way
to a 90
phase shift.
in
out
R3
100k
V
V1
1Vac
0Vdc
R2
1k
Frequency
C1
0.01uF
Looks like a simple voltage
divider at high frequency
May 20, 2015
Physics 160
23
With 1k Resistor in Series with Cap
Spice simulation of the
differentiator with a
square wave input.
i
t
Now we see just the derivative of
the square wave (no oscillation).
May 20, 2015
Physics 160
24
No more
resonance
Gain
n (dB)
Phase (d
deg)
With 1k Resistor in Series with Cap
Frequency
May 20, 2015
Physics 160
25
Lab Differentiator Circuit
C1
0.1n
Added cap here
gives even a little
more phase margin
R1
100k
V2
2
V3
.01uF
V
-
B1
5
6
1
V
LF411
4
V1 = 0
V2 = .1
1
TD = 0
TR = .5m
TF = .5m
PW = 0.5m
PER = 2m
1k
B2
OUT
C2
V-
R2
+
V+
+
U1
3
7
15Vdc
Ordering of R2 and C2
doesn’t matter.
V1
15Vdc
May 20, 2015
0
Physics 160
26
Differentiator without feedback capacitor
The g
gain turns over at
about 35 kHz, as
expected.
May 20, 2015
Physics 160
27
Differentiator with feedback capacitor
Turn-over lowered from
35 kHz to 14 kHz
May 20, 2015
Physics 160
28
Input and Output Impedance
•
Differential input impedance
– For a non-inverting amp, it is boosted by a factor of (1+AB) to near
infinity due to negative voltage feedback. Thus it becomes an
insignificant parameter. (Remember: AB=“loop gain”.)
•
Output impedance: 40 ohms for the 411 without feedback
– For a real application the output impedance depends on the
feedback network (low for V source; high for I source)
– With negative voltage feedback the output impedance drops by a
factor of 1/(1+AB) to such a low value as to become insignificant.
The maximum current drive of the output stage is then the
much more relevant parameter.
•
Maximum output current and output swing
– 411
11 can go ffrom VEE+2V
2 to VCC2V
2 with a 1 kohm load
– Single-supply op-amps (e.g. LM358) can swing from ground to
close to the positive rail (but typically they can only reach ground if
they don
don’tt have to sink too much current
current, e
e.g.
g <50 A).
A)
May 20, 2015
Physics 160
29
Op-Amp Example (Single Supply)
Details of the current references and mirrors are not shown.
High Gain
DC-coupled
diff
differential
ti l
amp.
IIncreases Miller
Mill effect
ff t
to kill gain at high
frequency (for stability)
Short-circuit protection
Pushpull
output
t t
LM324
Darlington
2nd stage amp
(common
(commonemitter)
2 emitter
f ll
followers
May 20, 2015
Physics 160
Helps to pull
the output all
the way to
ground.
30
Input and Output Impedance
•
Input
impedance
without
feedback.
Non-inverting amplifier (voltage, or series, feedback)
– Zin gets increased by (1+AB)
– Zout gets decreased by 1/(1+AB)
V
A
G  out 
Vin 1  AB
Vin  BVout
Vin
I in 

Ri
Ri  (1  AB)
Zin
Differential
gain = A
Output
impedance
without
feedback.
Suppose Vin  0 and apply V at output.
Amp sees 0  BV across its 2 inputs.
V  ( ABV )
1  AB
I out 
V 
R0
R0 1/Z
out
May 20, 2015
Physics 160
31
Input and Output Impedance
•
Inverting amp (shunt feedback)
Zin reduced by 1/(1+A)
Iin
R1
2
V-
In
4
R2
-
Out
8
+
1
V+
OUT
3
V V
V  (1  A)
I in  in out  in
R2
R2
Impedance into the amp input itself is
so high that we can ignore it here. It is
in parallel with R2/(1+A), which is tiny.
T
Transresistance
it
Amp
A
0
2
-
+
1
Out
V+
OUT
3
Z in  R1 
V-
R1
R2
 R1
1 A
8
In
4
R2
0
May 20, 2015
The output impedance of the inverting
amp is the same as for the noninverting amp. In both cases the
voltage is sampled at the output and
fed back.
Physics 160
32
Output Impedance
•
Current source (current sampling)
The feedback is
using R3 to
sample the output
current.
15.00V
1.020mA
R1
R3
470
2.7k
3
V1
12 25V
12.25V
-28.05pA
28 05pA
+
7
12.25V
V+
V
U6
B2
OUT
2
-28.05pA
-
B1
M1
6
8.629V
VP1310
1
LF411
Vout
4
9.064mA
V-
15Vdc
5
R2
8.793V
5.862mA
RL
12k
1.5k
0V
The op-amp output
uses the MOSFET
gm to control the
current.
V2
15Vdc
2.170mA
-15.00V
When Vout changes from 0 to 11.5V, the current
changes from 5.86209625 mA to 5.86209775 mA.
Z out 
May 20, 2015
11.5
1.5  10
Physics 160
9
 7.7 G !!
33
Some additional circuits studied in the lab:
ACTIVE RECTIFIERS
May 20, 2015
Physics 160
34
Active ½ Wave Rectifier
U1
3
+
7
Negative feedback exists
here only when the op-amp
output swings positive!
V+
V2
15Vdc
The input does not
need to exceed a
diode drop
p to g
get a
non-zero output.
B2
2
-
VV
OUT
B1
5
6
1
V
LF411
V3
4
VOFF = 0
VAMPL = 0.5
FREQ = 100
D1
D1N914
V
R2
10k
This is very similar to how
we removed
d th
the di
diode
d
drops of the push-pull
output stage by using
g
feedback.
negative
V1
15Vdc
May 20, 2015
Physics 160
35
Active Rectifier Transient Simulation
Op-Amp Output
100 Hz
Negative
feedback
No feedback
Rectifier Output
Glitch
May 20, 2015
Physics 160
36
Active Rectifier Transient Simulation
Op-Amp Output
10 kHz
Rectifier Output
May 20, 2015
Physics 160
37
Improved Active ½ Wave Rectifier
R1
10k
V2
V
15Vdc
3
+
7
B1
5
D1N914
6
1
LF411
4
V3
-
V-
2
VOFF = 0
VAMPL = .5
FREQ = 10k
B2
OUT
R2
10k
D1
V+
V
U2
D2
D1N914
V1
15Vdc
15Vd
May 20, 2015
Physics 160
D2 provides a feedback path
when the op-amp output
swings negative, keeping
the op amp from trying to
swing all the way to VEE.
38
Improved Active Rectifier
Op-Amp Output
10 kHz
Note how the op
amp no longer
has to swing
g
down to 15V, so
the requirements
on slew rate are
much relaxed.
relaxed
Rectifier Output
Note the
inversion with
respect to the
previous circuit.
May 20, 2015
Physics 160
39