| |
Operational amplifiers, commonly known as op-amps, got their name from modules used in analogue
computers to perform "operations" such as adding and multiplying. Now they
are integrated circuits for application as general amplifiers. They
seem easy to use, but the availability of so many types suggests that
their use can require significant knowledge and skill.
The power supply for an op-amp is normally bipolar, with voltages above
and below ground, called +V and -V. Most common op-amps can stand up to
36V, or ±18V. It is convenient for our experiments to use ±12V, which is
usually available from multi-output power supplies. You can always arrange
a bipolar supply from two ordinary supplies. Ground in this case is merely
a voltage between the supply rails, as they are called, of no special
significance. Op-amps have no ground terminal, since this reference is
unnecessary. If you have trouble remembering polarity, have lots of
op-amps around, since they are instantly destroyed by any mistake.
The pinouts of the most common op-amp is shown on the left. This is the
usual DIP package, as seen from the top, with pins numbered from upper
left, down one side and up the other to upper right. The 741 is a
bipolar op-amp, long a standard.
The connections marked + and - are the inputs to the op-amp,
and the connection from the point of the triangle is the output. The
output can go from a value near +V to a value near -V. When the ouput is
near one of these limits and can go no farther, it is said to be
saturated. You can short-circuit an output if you want, since it is
internally protected against too much current. On the other hand, the
output will handle only up to about 20mA at best. Op-amps are not for
power applications, but can drive a power amplifier (usually transistors)
if power is needed. The output is proportional to the difference in
voltage v+ - v- between the two inputs, where v+ is the voltage at the +
or non-inverting input, and v- the voltage at the - or inverting input.
The voltage gain of the amplifier is perhaps 100,000 or 100 dB at low
frequencies. With such a gain, the voltage between the inputs must be very
small if the output voltage is not to be at saturation. This amounts to a
rule: the voltages at the inputs are equal when a circuit is working
properly.
In order to make the voltages at the inputs equal to each other, it is
necessary to arrange this by feedback. All op-amp circuits use feedback,
and the properties of the circuit are determined by the feedback, not by
the properties of the op-amp. It's best to study op-amp circuits with no
reliance on feedback theory, and to use the results to understand and
appreciate feedback theory instead. Then one can come back with greater
knowledge to handle more difficult cases.
The common-mode input signal is the average of the potentials of the two
input connections. Since they are usually at the same voltage, this
voltage is the common-mode input voltage. The op-amp ignores the
common-mode input, and determines its output only by the difference
signal. Nevertheless, it is important to look at the common-mode input
voltage and see that it does not leave its permissible range. The
common-mode range of an op-amp is less than from +V to -V, and the op-amp
usually does something unpleasant when the range is exceeded. Some early
op-amps had a very limited common-mode range.
That the inputs are usually at the same voltage does not mean that they
can be connected to each other. If you do this, the output usually
saturates. The voltages must be held equal by the active participation of
the output, acting through the feedback network. The inputs also carry a
small dc bias or leakage current that must have a route to the power
supply. With bipolar op-amps, this current is actually the base bias
current for the input transistors, and sometimes has to be considered in
the circuit design. In ordinary circuit analysis, the bias currents
can be neglected, and it can be assumed that the inputs carry no current.
Don't forget that this is only approximate!
The most important factor hidden from the casual user of op-amps is the
question of stability. Stability is always important with high-gain
amplifiers, and when feedback is applied. The feedback loop can become the
route for a signal to be fed back to the input in the proper phase to
cause oscillation, called instability. Without some care, feedback always
results in instability, which is always fatal. The oscillation can occur
either at a higher or a lower frequency than that for which the circuit is
designed, usually higher (like the feedback with a microphone and
speaker). With the ordinary op-amps, stability is guaranteed by making the
gain fall off at 20 dB per decade of frequency, beginning at about 10 Hz,
so that the gain of the amplifier falls to unity at around 1 MHz. Unless
you have capacitors in unfortunate places, this guarantees that the
circuits you put together will be stable, no matter what you do. What you
pay for this is a severe restriction on the bandwidth of op-amp circuits,
and overcoming it is advanced work. |
|
|
