I'm trying to measure the current through a sense resistor (0.1R) in both directions.
I've found a few circuits using opamp + transistor or mosfet to translate the sense voltage onto an output resistor. This allows a simple way to drop the voltage down to a reasonable level to go into an ADC. Essentially the same as whats inside a high-side current shunt monitor like
INA168.
Example with fet here:
https://www.analog.com/en/technical-articles/upower-precision-dual-op-amp-combines-advantages-bipolar-cmos-amplifiers.html
I can get the fet simulations to work OK with standard parts (micro-cap), but the built up circuit is not behaving the same way. Both outputs are at high levels, say 5V (about 0.3mA across a 15k resistor).
This is the actual circuit:

Opamp:
MC34072Offset voltage: 0.5mV typ, up to 3mV
Input offset current: 6nA typ, up to 300nA
I would use two high side current sense amplifiers, but best I found is something like INA282:
- Can only handle -14V common mode range, so needs protection if its going to be reversed
- Bandwidth isn't great (10kHz)
I could give in and switch to a precision difference amplifier. The
easysmu uses LT1991 for example, but I'm not completely certain how that is sensing in both directions, as Vref is set at 3.3V and not mid rail.
In my case its kind of nice to have the dual outputs (one for + current, one for -) as it doesn't cut ADC resolution in half.
Anything obvious I'm missing here? I realize the opamp is not an ideal part, but I'm not sure yet if the issue is entirely from that.
edit: I realize the FET feedback is going to the wrong input terminal in the schematic, will see if that resolves the issue tomorrow and update. Oddly in simulation taking it to the negative input pin seemed to work better.