Noise Filtering a current clamp circuit

[cupoftea]

Hi,
Please imagine you have been given a crowded PCB layout which comprises the attached Current Clamp circuit. (jpeg and LTspice) The sense resistor is some 5cm away from the opamp, and in the short given time that you have, bringing that nearer to the opamp will not be possible.
As such, I assume that you would want to be sure that as much as possible, the sense resistor gets “diff pair” routed back to the opamp?

Also, the schematic shows six ringed noise mitigation capacitors. Supposing that you can only use 1 of them. Which one would you choose?
Now say you are allowed two of them….which two would you choose?

(BTW, the 1.25V reference is also some 6cm away from the opamp.)

Also, C2 looks like a potential problem....it is pretty well direct from opamp output to rail, which is a no-can-do with many opamps....would you agree?

 

[cupoftea]

Ayk, the cct is a bog standard current clamp , usually seen with NPN and coming off ground instead.

 

[KlausST]

Hi,

Shunt should be kelvin wired.(force / sense)

* C5 and C7 are counter productive and should be omitted anyways.(My opinion)
* C4, C9 can be increased to 1uF
* C8 may be omitted, since it is in parallel to C6 (maybe increase C6)
additionally C3 and C6 are AC coupled via the power supply. Thus they act as if they were in parallel.
Maybe use C3 only. Still the problem is: there is no dedicated R to form a defined (RC) low pass.
and all of them are functionally over ruled by C1. (Thus you may try to omit all C3, C6, C8)
* C2 may be omitted, since it is overruled by C1 (Maybe decrease R3 value)

Klaus