ATSAMD11 PWM pins? about atsamd HOT 13 CLOSED

@jessebraham

from atsamd.

Just got back from out of town, I'll take a look at it in the next couple days here.

from atsamd.

I've added (and tested) Pwm2 to my fork (36afab5) which makes use of TC2.

As far as using multiple pins, the datasheet seems to imply it's possible:

Timer/Counter -> 24.8 Signal Description

Refer to “I/O Multiplexing and Considerations” on page 13 for details on the pin mapping for this peripheral. One signal can be mapped on several pins.

I believe it is possible to use TCC0 as well. However, unless I'm mistaken, in both cases our pwm! macro needs to be modified as it currently expects a single TCxPinout as well as a TCx parameter.

@sajattack , am I on track here? The pinout member in the $TYPE struct does seem kind of like magic and doesn't appear to be read anywhere yet is still working as expected, so I'm clearly missing something there.

from atsamd.

One signal can be mapped on several pins.

The thing I'm aiming for is to have 4 separate PWM counters. From a not-too-careful reading of the datasheet, it seems like this might be possible with TC{1,2}/WO{0,1}. I might be misunderstanding the way the WO channels work - the datasheet says Waveform output channels for TC : 2, whereas it says Waveform output channels per TCC : 8. I'm not sure if this means 2 WOs per TC or 2WOs total shared across both TCs. If it's the latter, I think I would have to use TCC0.

from atsamd.

The mode we use the WO in uses 2 channels for one pwm output.

from atsamd.

One thing that's confusing me about the rust API is that the definition of atsamd11c14a::tcc0 seems to be missing some stuff compared to what's mentioned in the datasheet/C snippets. E.g. it seems like cc should have some additional structure to account for the fact that there are 4x compare channels. All of this stuff is autogenerated right?

from atsamd.

Could've been missing from the SVD

from atsamd.

Got it, thanks. So what's the best way forward here? Manually edit the SVD file? Ask the manufacturer? Just give up and write everything in C?

from atsamd.

FWIW, I was able to get this working for my needs via the C FFI. So no longer any urgent desire on my end. Still probably worth fixing though.

from atsamd.

I don't think our HAL supports TCCs yet anyhow so we can worry about it later.

from atsamd.

Sorry, had a busy week. Glad you got a solution figured out.

I have started trying to adapt the pwm! macro to use TCC0 instead of TCx, just as a proof-of-concept. If/when I get it working maybe there can be some discussion on how to move forward.

from atsamd.

Can this be closed now?

from atsamd.

Sure, although Jesse, let me know if you want me to reopen so you can track adding TCC0 support here

from atsamd.