I have verified all four drive strengths on real ESP32 hardware. The IDF docs suggest that the same drive strength values apply to S2, S3 and C3 variants, but I have no real hardware to test this on.

The maximum source/sink currents come from note 8 in table 24 of the ESP32 Datasheet; the approximate internal resistances come from my own tests and this useful forum post.

The drivers are non-linear so pushing them to the limit results in much higher apparent resistance. I've gone with roughly where they stand when running at the maximum recommended current. I've included these rough internal resistances in the docs because they can make a big difference to circuit design – e.g., in terms of resistor choice when driving an LED.

Thanks for this, the feature looks good.

The only problem is the name of the drive constants:

• in the docs it specifies LOW_POWER, MED_POWER and HIGH_POWER
• cc3200 and samd ports use the above constants
• mimxrt uses DRIVER_OFF and POWER_0 - POWER_6

There was a discussion of the drive/power constants for mimxrt, but I'm not sure why it ended up being different to the docs. Probably because there are many values and it's hard to name them with low/med/high etc and understand the ordering.

I do actually prefer the names DRIVE_LOW, DRIVE_MED, DRIVE_HIGH, because it's clear what argument they should be used for, namely drive. This matches the pull constants like PULL_UP.

It would be good to make all of this consistent. Here is my suggestion: use constants DRIVE_0, DRIVE_1, DRIVE_2 etc (without a limit on the number of them). Then a port just uses 0 up to the maximum number of options that it has. It's very clear what the ordering are. And the word DRIVE_ matches the argument name drive.

@robert-hh what do you think?

@dpgeorge It is always favorable to have a common API, and I can go with any name, shorter being better. I'll discuss it with @alphaFred, because that part was made by him.

Hi @dpgeorge
@robert-hh relayed your message since that module was implemented by me. I would be fine with the name change. I totally agree - having a commons naming is much better.

Sounds good to me! I've updated my code and docs, and added a bonus commit updating the machine.Pin docs as well.

OK, so we go with DRIVE_n constants for all ports then.

Yes. I will provide a PR for mimxrt soon, too.

Merged in 33083bf through 3ebc370 (with minor changes to the last commit to further update the constants in the docs).