Fixing phase reversal with an input resistor

[fryingpan]

I realised that a design of mine may potentially exhibit phase reversal in some borderline cases, since the chosen opamp (NJM4580) can be subject to this (technically the datasheet says that the input common mode range is equal to its output swing, both in its minimum and expected value, but I can't be sure the two are related) and I am employing one as a buffer. The PCB is already being printed and the only solution I may have at my disposal is adding an input resistor to the non-inverting input. How large should this realistically be? I was thinking about 2k should suffice.

[ElectricDruid]

Have you seen this paper about phase reversal?:

https://www.analog.com/media/en/training-seminars/tutorials/MT-036.pdf

TLDR: The resistor is possibly enough, and 1K/2K isn't a bad place to start, but adding schottky protection diodes would cover more different op-amps and provide more protection.

[Rob Strand]

Look at the later Boss input buffers using op amps that add a diode.   You need the input series resistor for the idea to work.

[fryingpan]

The buffer will only see as high a voltage as the preceding opamp can output, and it's the same opamp. I'm not worried about overvoltages and consequent damage to the opamp, I'm only worried about phase reversal.

[Rob Strand]

IIRC, the Boss buffer prevents large positive outputs during phase reversal.   If the input swings negative enough to cause phase reversal the output will raise the input to a point of equilibrium.   Whether it sounds any better is another matter.