# difference amplifier input

Discussion in 'Electronic Basics' started by kell, Dec 17, 2006.

1. ### kellGuest

I was reading up on op amp difference amplifiers
one thing I haven't seen addressed explicitly is the impedance of your
voltage signals.
Suppose your input resistors are 100 and feedback resistors 1 Meg, for
a gain of 10,000.
Now let's say you are using a 100k pot to set one of the voltages going
into the difference amplifier. Would this high impedance source affect
the difference amplifier, like adding resistance to the input resistor,
or does it matter?

2. ### JamieGuest

IF Op-amps were ideal, it would not effect it with the acception of the
feed back resistor.
if you look up the spec's of the op-amp, you'll find the gain of the
op-amp and impedance. These figures can be used to get the exact gain
over all.
The gain is so high on most op-amps that these factors are almost
non negotiable.

3. ### Phil AllisonGuest

"kell"

** The voltage you work with is the one that really exists at a particular
point in a circuit.

Where a voltage or signal has a significant series resistance that must
always be taken into account.

** A difference amplifier responds to the voltage difference between the two
inputs.

In the example you gave, the input imposes a 200 ohm load on that
difference. If this seriously alters the signals you want to difference,
then that particular circuit is no use.

When high impedance sources are involved, the solution is to add two more
op-amps, wired as simple voltage followers, to function as input buffers
for the difference amp's inputs.

Use FET op-amps for this and the effective input impedance can be G ohms.

........ Phil

4. ### ChrisGuest

Mr. Allison's advice is great. The setup he's talking about, with the
voltage followers preceeding the difference amplifier, is generically
called an instrumentation amplifier. It's good to use an IA when you
have to worry about higher or uneven input impedances. Here's a
reference:

http://www.national.com/ms/LB/LB-1.pdf

Also, I guess I should mention that your example is a little
unrealistic. Gains of greater than 100 have to be done very carefully,
and a gain factor of 1000 is pretty much the realistic limit for even
the best instrumentation amp setups. You end up amplifying offsets and
noise more than your signal. Also, your maximum input frequency goes
way down.

Good luck with your studies
Chris

5. ### EeyoreGuest

An op-amp with feedback isn't a classic difference ( differential ) amplifier.

Graham

6. ### kellGuest

I have to set up the difference amp so that the output swings to the
limit when the input differential reaches just a few millivolts. Can I
do this, given the caveat you express about excessive gain? Frequency
isn't a problem, it's very slow.

7. ### ChrisGuest

OK, Kell. You might want to just try it yourself, and see what
happens. Nothing will blow up, and the Electronics Gods will still
smile on the world.

I basically found out most of what I know about pulling millivolt-level
signals out of the grass when I tried to do it for real. It's a good
education.

Cheers
Chris

8. ### kellGuest

Maybe I was thinking too hard. A comparator ought to do. I only need
to keep the behavior of a physical system within certain bounds, not
get an accurate measurement of it. Instrumentation amplifiers can wait
for another day.