difference amplifier

Archiver

Hello all; I'm new here, so thanks in advance for any help you may offer.

Hacking apart a postal scale to get at the pressure sensor, I find that it is
using some
form of (hidden) stress sensor with 4 wires - 2 power, 2 outputs. I need to
use a
difference amplifier to take the ~2mV difference generated by this sensor up to
a 0
-5V reading that I can use for input to a DAC I have hooked up to my Stamp 2.

I'm no expert with op-amps, but it seems to me that this should be relatively
straightforward. Anyone got a good resource they can point me to? Or a circuit
they
can propose? I have been learning with a LM358N and there was a 324N in the
scale
(before you ask, the range of outputs wasn't adequate before I disassembled!)

Much obliged,

hans

Archiver

that's probably a wheatstone bridge. It is in effect four resistive
circuits tied together in a diamond pattern, voltage goes in two
opposing inputs, the other two opposing outputs change dependent on
the stress on the circuits. Do a google search for it, there is a
lot of info out there.


--- In basicstamps@yahoogroups.com, "hrs_mtl" <hans@a...> wrote:
> Hello all; I'm new here, so thanks in advance for any help you may
offer.
>
> Hacking apart a postal scale to get at the pressure sensor, I find
that it is using some
> form of (hidden) stress sensor with 4 wires - 2 power, 2
outputs. I need to use a
> difference amplifier to take the ~2mV difference generated by this
sensor up to a 0
> -5V reading that I can use for input to a DAC I have hooked up to
my Stamp 2.
>
> I'm no expert with op-amps, but it seems to me that this should be
relatively
> straightforward. Anyone got a good resource they can point me
to? Or a circuit they
> can propose? I have been learning with a LM358N and there was a
324N in the scale
> (before you ask, the range of outputs wasn't adequate before I
disassembled!)
>
> Much obliged,
>
> hans

Archiver

Hans,

You are correct: it is a situation that calls for an operational
amplifier.

Op-amps all work the same way. They have two inputs, labeled + and -.
The amplifier works on the difference between these two inputs.
Watch the specification for "common mode rejection", which indicates
how well the op-amp manages to ignore the voltage that both inputs
have in common. For esample, imagine a current shunt that is in the
positive lead from a 9 volt battery. Connecting the - input to one
end of the shunt and the + input to the other lets the op-amp handle
the difference between the two ends, while the "common mode rejection"
tells you how well the amplifier ignores the approximately 9-volts
they have in common.

Above is the simplistic view. To make a working circuit you have to
go a little farther. Speciffically you have to provide a voltage
divider and feedback from the op-amp's output back to the - input.
For example, let's say you use a 1000 ohm resistor between the above
current shunt and the op-amp's - input. You then have to put a
resistor between the - input and the op-amp's output. The ratio of
these two resistors determines the differential gain of the circuit.
(The - input is also known as the "error input" because it amplifies
the very minute voltage at the - input to produce the input. Op-amps
often have gains of a million or more, so the - input has to be
essentially held at zero volts.)

To calculate the feedback resistance, first determine how much gain
you need from the op-amp. In your case, you want the output to be in
the 0 to 5 volt range with a 0 to 2 millivolt input. This gives
5000 mv

---

2 mv

or a gain of around 2500. With an input resistor of 1000 ohms, the
feedback resistor would be around 2500 x 1000 = 2.5 megohms.

With the single-rail op-amps you have to watch out how far negative
you go. This shold not be a problem in your case, as you can make
sure the 2 mv input is positive to the op-amp's - input.

Noise is a critial factor, particularly with such high gains involved.
Keep the leads to the - input of th eop-amp as short and shielded as
possible. Use low-noise resistors. Pick an op-amp with inherently
low noise attributes.

I would be tempted to use two op-amps and split the gain between the
two. That is, say the first op-amp has a gain of 50 and the second a
gain of 50 (50 x 50 = 2500). I would use two 500 ohm resistors in
series between the first op-amp's output and the second op-amp's -
input, with a capacitor connecting the mid-point of the 500 ohm
resistors to ground. (Not sure what the capacitor's value should be,
probably around 0.1 mf)

This isn't easy to describe with words. If you wish I can e-mail you
a diagram.

Russ

--- In basicstamps@yahoogroups.com, "hrs_mtl" <hans@a...> wrote:
> Hello all; I'm new here, so thanks in advance for any help you may
offer.
>
> Hacking apart a postal scale to get at the pressure sensor, I find
that it is using some
> form of (hidden) stress sensor with 4 wires - 2 power, 2 outputs.
I need to use a
> difference amplifier to take the ~2mV difference generated by this
sensor up to a 0
> -5V reading that I can use for input to a DAC I have hooked up to my
Stamp 2.
>
> I'm no expert with op-amps, but it seems to me that this should be
relatively
> straightforward. Anyone got a good resource they can point me to?
Or a circuit they
> can propose? I have been learning with a LM358N and there was a
324N in the scale
> (before you ask, the range of outputs wasn't adequate before I
disassembled!)
>
> Much obliged,
>
> hans