Opamp difference amplifier

[Mosaic]

Folks I need some advice on making a difference amp to perhaps do a 20x amplification on a very low impedance (<1 ohm) voltage on the order of around 20mV to 200mV. The purpose is to read the voltage via a PIC 10bit ADC, using a LM4040 4.096v precision Vref shunt.

I have access to these opamps atm:
MCP 6001, LM258, MC33072, TLE2022, LM311, LM324 and LM239.

I think I can scale the precision Vref to 2.048 or less using a voltage divider switched by a logic FET via a uC pin for better sampling at the lower end of the scale. I have this logic level NFET:
https://www.electro-tech-online.com/custompdfs/2012/06/FDC6401N.pdf

At a glance the MCP6001 or the TLE2022 seem to have the best specs.
Since the input impedance is so low, would I be able to use a single op amp w/o input buffering such as in this tut?
**broken link removed**

thanks!

 

[MrAl]

Hi,

If you are going to measure down to 20mv then you need an op amp with input offset very low or else you have to be able to null the output. Since there are decent op amps that do go down quite low these days and also have minimal drift, you can use one of those.
There is also a question of speed. If you dont need very fast speed (more than maybe 1kHz) then you can use the self adjusting op amps that will provide very very low input offset.

Just as an example, if you have an op amp with 2mv input offset then with a gain of 1 your measurement will be 10 percent off. If it has 0.2mv offset then it will only be off by 1 percent with same gain.

The MCP6V01 is spec'd at something like 2uv input offset. It is self adjusting. So measuring 20mv with a gain of 20 would set the initial accuracy down to around 0.2 percent (that is 2 tenths of 1 percent which is 5 times better than 1 percent). Of course you also need closely matched resistors depending on the required overall accuracy. There are simple calculations we can do in order to investigate the inaccuracies caused by mismatched resistors.

The other thing here is what accuracy you are actually after. If you dont need 2 percent accuracy then it's a simpler problem too.

The fact that the voltage to be measured originates from a low impedance is actually a benefit, not a liability. If it was higher it would pose a problem with the required input impedance of the op amp section. Count yourself lucky for that