General • Re: ADC creapage on the RP235X

What is your analog circuit for measuring ?

Open circuit, though the Pimoroni Pico Plus 2W I am using has an on-board in-line 1K between the pad and the GPIO40/ADC0 on the RP2350B.

I didn't see such behavior (and I should have seen) when tested ADC DNL on Pico2:

You wouldn't have unless only reading the ADC with greater than 10 minute intervals between reads. I suppose I ought to check to see at exactly what interval it does change behaviour.

If it's on the lower end of the scale it could be simply analog bias/offset, that could raise with temperature or other factors. And, yes, it's everywhere.

Some offset and bias, plus noise, I was expecting.

But the weird thing is that it's fine, near constant, as it also was on the RP2040, providing ADC is read every 10 minutes or with a shorter interval.
I would have put it down to temperature or other factor if it were creeping up consistently-ish but it isn't, it's behaviour does seem to drastically change after a gap of 10 minutes or so.

Also, leave it to rise, read it, and it resets to the same baseline; every subsequent sub-10 minute later reading still returns that baseline.

Not an errata, it's better fleshed out in the datasheet's ADC section. This thing is everywhere; so it's more of a physical limitation that may be more obvious to the end user for 40 nm chips. It's a good teachable item to put into the datasheet.

I would be interested in hearing what it is or seeing a pointer to an explanation.

I don't understand it and I have never observed this behaviour with any other ADC I have used.

Statistics: Posted by hippy — Wed Nov 20, 2024 9:51 am

---