Hello,
my problem is similiar to this:
https://e2e.ti.com/support/amplifiers/etc_amplifiers__other_linear/f/18/t/167338
But in my case it already occurs at 10kHz (and very low output amplitude).
At 100kHz it's worse (so also GBW seems to play a role).
My latest flattened circuitry now is noninverting with unity gain at +/-12V symmetric supply with recommended blocking capacitors. The load is 2 ohm, the peak current therefore +/-50mA only. The 40dB 'safety margin' is fullfilled at 10kHz, unity gain as Thomas recommended in the other thread. To me it looks like a too late conducting element which results in a larger voltage difference amplified by Aol, too late oposite regulation thus creating overshoot/ringing. It's strange that the distortion always occurs in negative output only. Even at 100kHz the positive half looks well. I can't explain it by myself...
Regards
Andreas