Son of Screamer cuts-out when room light is switched

[PRR]

> can the inputs have their own slew-rate limits

They could.

But input signal levels are far-far lower than output signal levels (it IS an "amplifier"). Using similar technology throughout, it "seems reasonable" that input slew is not much less than output slew, while input signal is much-much less, so output slew dominates.

Also we know some specifics. The 'O7x inputs are J-FET which generally will not slew, or not near as bad as BJT devices. Also for "most" opamps the slew limit is really set in the 2nd stage, and really by loop compensation concerns.

> I wonder why TI is explicitly

Utter speculation. The 'O7x part is very old. Un-compensated amps like '301 still wandered the landscape. A '301 can slew 50V/uS when compensated for closed-loop gain of 100, but only 1V/uS when compensated for unity gain. Obviously the hi-gain slew was featured prominently on the sheet, the unity-gain slew in the small print. The faster pre-compensated amps liked to mention the unity-gain slew loudly, to entice designers of fast low-gain stages.

All of this became moot decades ago, but datasheets never keep up with the times, except to list new packages, disclaim life-support use, or tighten max-abuse specs that turned out to be too generous.

[Transmogrifox]

I have to add that all high frequency stuff and datasheet guarantees hold no promises when you're working on a breadboard.   PCB layout becomes a real issue for all of these things that have slew rates of 16V/us, and so on.  Even your through-hole capacitors are a resonant LC network at 10 MHz to 30 MHz, not to mention the inductance and parasitic capacitance related to breadboard traces probably create >8th order networks on each node.

If you're having strange problems on a breadboard, then try to perf board the circuit with a tidy layout keeping wire lengths and lead lengths to a minimum.  You may never understand the weird impedance characteristic that made your op amp unstable -- not to mention if you have fluorescent lights, those will couple 50 kHz noise directly into your circuit.  It may not be audible, but it most certainly will mess up what you hear in the audio band.