Professional Documents
Culture Documents
BasicOp Amps PDF
BasicOp Amps PDF
GRAF, KA2CWL
Basic Op -Amps
www.americanradiohistory.com
R1 other than to power the amplifier. Many
general -purpose op -amp chips have two
or four separate operational- amplifiers
SUMMING JUNCTION in one package, with common power -
'VIRTUAL GROUND" supply connections. In practice the ideal
amplifier criteria requirements are met
R2
only approximately, but as will be
shown, close enough for most purposes.
+ Practically, an op -amp will have a gain
of 10,000 or more, an input impedance
V out
of megohms, and a 3 dB bandwidth of
several tens of hertz or more. If an
amplifier has a 3 dB bandwidth of 40 Hz
and a gain of 100,000 times, this is a gain
V in
bandwidth product of 4 million hertz, or
4 MHz. (40 x 100,000). It is advanta-
V in = A ein
geous in many feedback applications to
have the gain falling at 6 dB per octave
Vout / Vin = A / (1 +BA)
or 20 dB per decade at frequencies
B = R2 / (R1 + R2) beyond the corner frequency (that fre-
quency at which the amplifier gain has
fallen 3 dB or 70.7 percent of its DC
value).
Since the op -amp is used mainly in
FIG 2
feedback circuits having much lower
BASIC OP AMP VOLTAGE AMPLIFIER
closed -loop gain, these performance fig-
NONINVERTING CONFIGURATION ures are good enough in many cases. In
fact, even a single high -gain (100X)
voltage. Amplifiers are available with may show up as equivalent input signals. common emitter transistor amplifier
low -noise characteristics for those appli- The degree to which the op -amp rejects stage can be treated as an op -amp if
cations where noise must be kept to a this is called the supply -voltage rejection feedback is employed, with surprisingly
minimum. ratio. It is usually better than 60 dB and little error. In many cases, a single tran-
A real -world op -amp has a lot of gain typically 70-80 dB or better. After all, sistor will work almost as well as a more
( >1000X voltage gain) and a fairly high nothing is perfect in life. expensive op -amp device. One example
input impedance (>100K). Generally is a simple audio amplifier stage from
there are two inputs shown, an inverting What's To Gain? which a moderate gain (5 -20X) is
and a non -inverting input, and one out- Op -amp power supply connections required. This will be shown in an
put referenced to ground (but not are sometimes shown in diagrams, espe- example later.
always, differential outputs are some- cially if decoupling capacitors and resis-
times used in certain applications). One tors are necessary, but more often shown Op-Amp Families
of the inputs may be grounded in many elsewhere in the schematic, as they play One of the most popular op -amps of
common applications where a single - no part in the primary circuit function all time is the venerable LM741, its dual
ended signal source is present. This is a
common situation. There are limitations R2
on the DC levels allowable on the inputs f.vn nv
-
1
full output voltage swing between the R1 1
www.americanradiohistory.com
Rí that, while not difficult, can contain sev-
1/W
Note: Iin is almost 12 eral terms and fractional expressions,
zero and is
negligible leading to rather messy algebraic manip-
ulation of these terms. This can confuse,
R1 It
intimidate, and scare away many readers.
It is easy to get wrapped up in the math
SUMMING/
JUNCTION
ein + and then spend way too much time try-
V out ing to figure out what is being done or
what is meant. If you have ever studied
algebra or calculus, you will surely have
been through this. You also lose sight of
12
Vout / Vin =- Rf / Rn for any input
the intended goal and the subject being
R2 n = 1,2,3 ....N
discussed. What was to be a discussion
of circuits turns into a time -wasting
Vout=-{V1(Rf/R1) V2(Rf/R2) V3(Rf/R3) +VN(Rf/RN)F
+ + +
digression, usually a tedious and frus-
trating algebra exercise. This proves lit-
de except that you might be a complete
13
bonehead because you do not immedi-
ately see the "obvious" meaning of these
R3
complex expressions. (Often the authors
needed several hours correctly deriving
them the first time or else just copied
them from elsewhere so as to impress
readers and look like geniuses). This will
FIG 4
BASIC OP AMP SUMMING AMPLIFIER
be avoided. We are going to make some
N - INPUT INVERTING CONFIGURATION simplifying approximations to get rid of
the second and higher order stuff, which
can be studied later after some basics are
covered. Simplified approximations will
RN still yield results accurate to a percent or
IN
so and avoid confusing trivial details
which, while interesting, have dubious
practical consequences for things the
experimenter will get involved with.
Even five or ten percent accuracy is
good enough in many cases, if you are
not doing instrumentation work.
www.americanradiohistory.com
R2 Gain =(Vout /Vin) =A /(AX [3 +1)
Note: in is small
I
and can be
12
+ A X f3 means the product of these two
neglected for high quantities. Since the order of multiplica-
gain transistors J
V out
High Gain tion does not change the product, A x 3
Transistor
R1 = f3 X A = f3A (realizing the X stands for
i1 _base multiplication we can get rid of it). Also,
{
the order of addition of two quantities
e in v
does not affect the sum. Then the equa-
BIAS
tion appears as:
VOLTAGE
Gain = A /(1 +f3A)
involve complex feedback networks, but back via the feedback network. By sub- plus X approximately equals X with an
the reader should be made aware that stituting the previously mentioned error of around 1/X times 100 percent.
this is not always the case. equalities in the first equation: As an example if X were 10 then 10 -11
In Fig. 2, the output voltage from the approximately with an error of 1 /10 X
op-amp is V out = A X e in. A is the gain A x ein {f3} = Vin -ein 100 percent, or ten percent, which is
of the amplifier (generally 10,000X or obviously true. If X were 100, then (1 +
more). In a practical op -amp circuit If you add like quantities to both sides 100) - 100 with an error of 1/100 X 100
powered by 5 -15 volt supplies, V out of the equation, it still is valid. percent, or 1 percent. Note that in our
will be at most_* 5 to ±15 volts. Therefore, if you add e in to both sides case where A is 10,000 and f3 is 0.01, the
Therefore, e in will be this voltage, V of the equation: product f3A is 100 and 1 + f3A f3A
out divided by the gain of the op-amp within one percent. Therefore, if in any
(10,000 or more). This says that e in is AXein {f3} +ein =Vin case ßA > 1, we can rewrite the equation as:
very, very small, in the millivolt or
microvolt range. However, V in from Noting that e in is common to both Gain =A /(1 + [3A) -A/ (ßA) = 1 /f3
the outside world is the input voltage we terms in the left side of the equation, it
are applying to the circuit and could be can be factored out: (Note that A is common to numerator
a volt or more, such as a line -level audio and denominator and can be cancelled
signal, etc., while e in is very much einx[Ax[3+1}=Vin out.)
smaller. The circuit adjusts itself so that In other words, if the product of the
the ratio of V out to e in equals the gain But e in must equal V out divided by op -amp gain (A) and the feedback factor
of the amplifier, which we will take as A, the gain of the op -amp, so that: ([3) is much larger than one, the value of
10,000. This requires V out to be such b determines the overall gain of the op-
that the portion of V out at the junction (Vout/A){A x (3+ 1}=Vin amp circuit. The product of (3A is called
of feedback network R1 and R2 exactly the open -loop gain. The overall circuit
equals V in minus e in, so the total volt- The effective circuit gain is what we gain with the feedback loop in place is
age difference across the inverting and want, i.e. the ratio of V out to V in. We called the closed -loop gain. The beauty
non -inverting outputs is e in. This are inputting a signal represented by V of this concept is that, given a large
occurs when: in and would like to know the magni- enough value of A, the gain and other
tude of V out that will result. If both parameters of a feedback amplifier or
V out {R2(RI+R2)} = V in -e in sides of the equation are first multiplied any other system employing feedback
by A, then divided by V in, and then can be closely controlled by a network of
But, V out = A X e in, where A is the finally by the entire quantity in brackets components that can be specified to any
gain of the amplifier. Define R2(R1 +R2) {A x f3 + 1), we get an equation that degree of accuracy needed. The value of
as [3, the feedback factor equal to the expresses the ratio of V out to V in as a A, component tolerances, drift, noise,
ratio of R2 to R1 and R2. For example, function of A, the op -amp gain, and b, temperature effects, and all things
if R1 = 9K and R2 = 1K then f3 equals the feedback factor: (Continued on page 41) 23
www.americanradiohistory.com
ALL ABOUT NORTH COUNTRY RADIO:
(continued,from page 23) A HAVEN FOR WIRELESS BUFFS
affecting A become less and less relevant Graf and Sheets are no strangers to the pages of Gernsback. Their educational projects,
to the circuit performance as the value of such as the RF-Field Strength Meter and the MPX2000 FM Transmitter, can be found at
North Country Radio. Established in 1986, this company offers projects related to ama-
ßA increases. teur TV transmitters /receivers, AM and FM transmitters/ receivers, video cameras, and
We do not mean to pull a snow job numerous other subjects. Visit the Web site at www.northcountryradio.com for more
here, but you should spend whatever information.
time is needed to understand these con-
cepts, as they are the heart of the theory. Ordering Information: PO Box 53, Wykagyl Station, New Rochelle, NY 10804 -0053:
Once understood, op-amp circuits will Voice: 914- 235 -6611; Fax: 914-576-6051; e -mail: Rgraf30832 @aol.com.
be a breeze to work with.
Engineering and Technical Support
PO Box 200, Hartford, NY 12838
Virtual Ground Voice /Fax: 518- 854 -9280
In a practical op -amp circuit e in is e -mail: support@northcountryradio.com
very small, since the value of A is at
least several thousand. Since e in is that
voltage appearing across the input of
the op -amp (see Fig. 3), if one input zero (Kirchoff's current law, the law of between all the input sources. This circuit
terminal of the op -amp is connected to continuity-and plain common sense), makes an excellent audio mixer with virtual-
ground or has zero signal on it, the it follows that the positive current flow- ly no cross-talk effects. By varying the values
other input will also be very close to ing in R1 must be cancelled by a cur- of the input resistors, Rl RN, different
ground. Note again that e in is at most rent flowing in R2, except for a tiny gains can be obtained for the various inputs.
a few millivolts in practical circuits. current flowing into the op -amp, which As far as AC signals are concerned, a
Under all signal levels this will be true, is zero for all practical purposes. The high -gain single -tran-
provided the op -amp is not driven into only way this can happen is if V out sistor amplifier
saturation or another region where the equals Vin (R2 /R1). circuit can ap-
gain falls to a low value. This gives rise Note that there is an inversion in phase, proximate the
to the term virtual ground, since the since the currents must cancel. The voltage behavior of an
op -amp input is always very close in gain is simply the ratio of R2 to Rl. The op -amp in these
voltage to ground. The input terminal two resistors set the gain. If multiple inputs circuits if the
in many applications is the inverting are desired, extra input resistors and input collector is con-
input, with the non -inverting input sources can be added as in Fig. 4. The out- sidered the out-
grounded or connected to a source of put voltage is given as: put, the base
zero signal. Additionally, the amplifier the inverting in-
itself has a high -input impedance, often Vout= -[Vn x Rf/R1+Vin2 x Rf
1 put, and the
measured in megohms. The input cur- /R2+Vin3 x Rf/R3+VinN x Rf/ emitter the non -
rent to the op -amp itself is negligible RN] inverting input. Naturally DC biasing
and zero for all practical purposes. arrangements are needed and there are
Therefore, in Fig. 3, the input current I This is called a summing amplifier (see DC level considerations, but the princi-
in R1, equal to V in /R1, has to equal Fig. 4) and the junction of all the resistors at ples of feedback still apply. (see Fig. 5).
the feedback current in R2, equaling V the input is called the summing junction. Several op -amp circuits will be dis-
out/R2. Since these currents entering Since the input of the amplifier is a virtual cussed in the next part of this article. p
and leaving any junction must equal ground, there is almost complete isolation
41
www.americanradiohistory.com
WILLIAM SHEETS, K2MQJ, and RUDOLF F. GRAF, KA2CWL
The op-amp should be rated to work with the supply volt- SHORT ANTENNA
OR PROBE
ages you will use. Anything from 5 to 12 volts will be okay. You
STATIC
will need two supplies of equal voltage, one delivering a posi- DETECTOR
www.americanradiohistory.com
A = TLO81 or
similar the thousands of megohms -available with this circuit. By
using a voltage -divider network and appropriate switching,
1 uF 50V
I in you can make a high- impedance voltmeter that will have input
impedance of hundreds of megohms or more. By using CMOS
1 MEG
AC INPUT 330 OHMS
R3 type op-amps, you can construct a simple electrometer able to
R1
VOLTAGE read currents as low as I trillionth of an ampere (or 1 picoam-
TO BE
MEASURED
-0.
11=VIn1330 pere, if you prefer).
- rduk
If the input of the amplifier is connected to a short (3 -inch)
I
\
.V
REFERENCE
VOLTAGE NONINVERTING
A = TL081
The Voltage Follower INPUT , or similar
lower impedance load with no loss in voltage. The gain here is Output - Positive if V input O REF. VOLTAGE
all current gain. Since the feedback factor is unity, the output INPUT VOLTAGE =
V input
voltage will equal the input voltage minus ein Since ei is very FIG 5
small and the gain of the amplifier is high (10,000), the output BASIC OP AMP COMPARATOR
voltage will equal the input voltage within 0.01 percent.
This circuit can be used as a simple, high- impedance meter must equal the input current through R1. This equilibrium
amplifier. If a DVM is connected to the op -amp output and a forces the op -amp to produce sufficient voltage to overcome
single 1.5 -volt AA cell is connected to the input through a 22- the 0.6-volt diode drop, irrespective of the input level.
meg resistor (highest commonly available resistor value), the Therefore, the meter will read linearly. It is easy to make an
resistor will have practically no effect on the voltage reading of AC voltmeter with a full -scale deflection of 100 millivolts or
the battery. This demonstrates the high -input impedance -in less with this circuit with a perfectly linear scale. This AC-
voltmeter circuit works well and will be fairly accurate in the
A = TLO81 or audio -frequency range and into the low -frequency RF range
similar (100 kHz) or higher with fast diodes and a wideband op -amp.
1 uF 60V
A voltage follower can be used as a peak detector (Fig. 4) to
give the peak voltage of a waveform. The signal is applied to
AC INPUT
1 MEG
the noninverting input as shown. Capacitor Cl will charge to
330 OHMS
VOLTAGE
TO BE
the peak voltage of the input signal. Since the diode is in series
MEASURED with the amplifier output, it will compensate for the diode
drop. The output voltage will equal the peak value of the input
voltage.
E out=
www.americanradiohistory.com
REFERENCE Now the capacitor discharges towards
VOLTAGE NONIN VERT IN
INPUTS the negative supply rail. This sequence
Vr
will continue until the voltage at the
ourPUr H- inverting input reaches the new reference
voltage. At this point the comparator
Al to A4 Is TL084
INVERTING
INPUTS
or similari switches to high (positive) output. The
section op amp
cycle is repeated. A square -wave output
results. By using diodes and two separate
feedback resistors, you can make dissimi-
lar charge and discharge paths, allowing
0.15 Vr two different time constants. This setup
allows generation of a variable duty-cycle
OUTPUT U - waveform. A potentiometer can be used
along with, or in place of, R1 and R2 to
REFERENCER
DIVIDER
adjust this duty cycle, as shown in Fig. 8.
LOGIC DIGITAL
R - 100 ohms GATE OUTPUT
TO BUS
to 1011
typically
ARRAY
Generating A Sine Wave
It is possible to generate a sine wave
0.5 V r using a circuit known as a Wein Bridge.
At a frequency f = 1 /(2rtRC), the network
OUTPUT T- shown in Fig. 9 will have a transfer func-
tion of 1/3 with zero phaseshift between
input and output-permitting its use as a
frequency- determining network. The
Wein bridge is connected between the
output and noninverting input as shown
in Fig. 9, allowing positive feedback and
0.25 Vr oscillation.
However, the amplifier would generate
Al OUTPUT 1-
a poor waveform, since limiting the out-
put can only be accomplished by driving
INPUT VOLTAGE
the amplifier to its positive and negative
Outputs= Negative it V input < NVr
V Input Outputs Positive It V input NVr
= limits, resulting in severe clipping of any
Fraction of reference voltage from reLdivider
N =
generated sine wave. Another feedback
FIG 6
network is used to introduce negative
BASIC FLASH A -D CONVERTER feedback. A resistive divider with a divi-
sion ratio of slightly more than 1/3 is
used. This ratio reduces the gain to a lit-
sand, this transition is very sharp. Voltages less than this tle over 3, which is enough to sustain oscillation. Limiting
amount will cause the output to go in a negative direction. would still be obtained by clipping in the output, although the
Output can be fed to an LED indicator or logic circuit as an waveform would be somewhat improved.
indicator or used to drive another circuit. Using a voltage- dependent resistor for R4 allows automatic
Several op -amps can be connected to a resistive divider and gain control. Resistor R4 is selected to have a resistance that
can have their inputs connected to a common input. (See Fig. increases with applied voltage. A thermistor can be used for
6). The outputs can be fed to a system of logic gates that will this purpose, but a more common approach is to use an ordi-
produce a binary pattern that is a function of how many com- nary tungsten- filament lamp. This kind of lamp has the exact
parators are ON or OFF. The output can be made to be a characteristic we need. As more voltage is applied across the
binary value representing the number of ON or OFF com- lamp, the filament heats up, its resistance increases, and the
parators. The idea is to make a "flash" analog to digital (A -D) negative feedback increases, and lowers the gain of the op-
converter, since the output is an instantaneous function of amp. This sequence tends to reduce the amplitude of oscilla-
input. The flash A -D converter is useful for digitizing fast tion to a level that will not drive the op -amp into limiting.
waveforms and is widely used in digitization of video signals. Very pure sine waves can be generated in this manner, and less
By summing the comparator outputs, a staircase wave can be than 1 percent distortion is easy to achieve. By making R1 and
generated from a ramp- waveform input. R2 a ganged potentiometer, you can obtain variable frequency
By using both positive and negative feedback, it is possible operation. This circuit was widely used in the vacuum -tube
to make oscillators with different output waveforms. Figure 7 days, and a 120 -volt, 3 -watt tungsten lamp was used for R4.
shows a square -wave oscillator. Capacitor Cl charges toward For an op -amp version, one of the 5 -volt, 10 mA subminiature
the positive supply rail through Rl. After it reaches the refer- lamps will work well. The lamp is typically operated at 10 to
ence voltage derived from R2 and R3, the comparator output 20 percent of rated voltage, and the filament should barely
goes low. This result also changes the reference voltage to a glow. Resistor R3 is a pot to adjust the amplitude of oscillation
lower (more negative) level, which forces the comparator to a at that level that yields satisfactory operation.
negative output. 23
www.americanradiohistory.com
se
R3
bandwidth B and center frequency f and
10K
1Nv gain A:
A= TLO81 or
similar
POSITIVE
w=27rf Q=f/B a=1/Q H=
+
R2
FEEDBACK alAl andQ>V(A/2)
10K
NAAf-- 3
R1=1/(HxwxCl)
T1 T2
Req = 1/Q (C1 + C2)w
oE
1CYCLF
www.americanradiohistory.com
123 = 83A R3B
R3A
100
R3B
SERIA
Serial LCDs work great with BASIC Stamps' and
R4
A = TLO81 or
LAMP similar
other microcontrollers. One -wire interface simple
5V 10 rnA
serial protocol low cost h gh quality in stock
BPI -216N
2x16 text LCD
Low Iptr, ease
2400/9600 bps
serial interface
_Taark/V.I.
$45 (non -backlit)
R1
TYPICAL VALUES
WEIN
BRIDGE R=160K SGX-120L
NETWORK R2 POSITIVE C=.001 uF Mini graphics LCD
FEEDBACK f = 1000 Hz
FOR THIS CIRCUIT.
R1
Electronic Projects For
INPUT 16K
HOME SECURITY
I OUTPUT
There is no substitute for experience, and undoubtedly you for delivery. Address
will come up with some circuits of your own that can be tai- City State ZIP
MA10
lored for your applications. P L a
www.americanradiohistory.com