adaptable soft power limit with backoff (mostly S60, useful for very-low-ohm tip like JBCs on USB PD)

[jakobi]

I would like to request a "soft power limit backoff on recent PD/QC resets":

Defaulting to off, or VALUE in WATTs. This feature should catch very recent PD resets due too-fast-switching or too-high-currents, and in this case lower the "soft power limit" in multiple steps down to say a lower sanity boundary of 20W.

The temporary soft limit should be kept until the tip is changed. A boot up without seeing any "recent reset" should change the value back to off (or the already existing hard power limit, which should take precedence if lower :)).

Alternative implementations: boolean to control backoff, and add the behaviour to the current power limit code, with a hard limit in flash and a current in ram.

Advantage: When switching to "friendly tips", power limit is off, when switching to tips where pwm is "too aggressive for the capricious charger", we get a reset or two, and the tip heats up with a temporary lower power limit. So no fiddling with the current power limit settings for proplematic tips.

Note that the notion of "recent" might be problematic, and sometimes the reset is immediate so that the tip is still cool to the touch, thus a hot tip is not sufficient for 'recent' (and unplugging/replugging with a hot tip should not trigger the "soft power limit backoff" either).

Use case / current situation with just a hard power limit:

This was tested with the 1.22 prerelease on s60 and a cranky baseus 120w pd/qc charger, supporting 12V/3A.

I have 2 tips working somewhat nicely (only one reset at low temp, recover+reheat and we reach solder temps).

3 more tips making the baseus reset without power limit, and even some downto a limit of 35W (ironos).

With a power limit of 25W in ironos 1.22, finally all tips heatup (slowly).

The behaviour doesn't seem to relate to clone tip vs JBC. And only my 2nd Baseus PD-charger-with-powerbank is more lenient and thus suffers somewhat less resets.

Still with the original firmware, I played with its "power percentage" setting, but I never noticed an effect, like e.g. a reduced number of resets during heatup.

PS: A very subjective impression, lacking a proper stop watch setup: original 1.11 firmware seems more aggressive/faster (both firmwares without power limits). But the original seems to round/dejitter over short periods to "better" display a temperature... . Not that I can revisit this lacking the hex of the original firmware :).

[jakobi]

Contrast to #1009: Above resets only happen after pressing a button, i.e. on actively heating the tip, i.e. the irons should be at a voltage other than 5v for the immediate resets.

If the power limit's nearly low enough to nearly keep the charger happy, the tip noticably starts heating, and I see 11v before the reset occurs.

On set-temp finally - IFF reachable - or with a suitably low power limit, the resets no longer occur.

I've that option to start heating immediately set to off (i.e. it is at the default).

[jakobi]

Note2: my s60 displays always QC. For the Baseus, which indeed can talk QC, but also for dc 12v power supply, which is a bit strange given a dc5521 cable :).

Note3: It would already help somewhat, if the power limit can be split into a specific limit for pd/qc and one for dc12v or global. Switching to dc would then run all tips at full power (or the overall power limit, with the pd/qc limit being ignored). PD/QC Charger friendly tips would however still be limited as the ones making the charger reset.