OPA2192: Measure high current with OPA2192

Hi Kevin,

The simplest way to reduce noise is to limit the bandwidth as much as possible. The more bandwidth you can throw away, the less total noise you will see at the output. In a non-inverting configuration the noise gain will only roll off to a minimum of 0dB. I assume you are operating in high gain, so this makes sure the higher frequency noise components don't see the same gain as your signal, but if you want to actually attenuate the noise then you will also need a low pass filter at the output. You must be careful in designing this though as you can easily cause the op amp to become unstable.

It's also possible to parallel multiple stages to average the noise. The noise is effectively reduced by the square root of the number of channels. However it is often not practical to do this.

If you are unable to reduce the noise to a satisfactory level, then you will want to find a lower noise device. Something like the OPA211 or OPA209 would work well.

Generally when you are working with such high current levels, parasitic resistances on your ground plane start to become an issue and when doing a low-side measurement you may find that "ground" is not actually the same everywhere on your board. For this reason it is generally better to do a differential measurement with high currents so that you ensure you are only measuring the voltage across the shunt and not including any additional ground bounce.

There is a TI-design for 200A measurement that also includes isolation. You can find it here: www.ti.com/.../tidubv1a.pdf.
You could also use an integrated instrumentation amplifier instead, like the INA188 or INA828.