PWM and PIC18F55Q43

Woody · Joined: 11 Sep 2003 Posts: 83 Location: Warmenhuizen - NL

Hi,

The continuous joys of using a relatively new piece of hardware Very Happy

Let me start with the good news: I have PWM running on this chip. Four outputs producing different PWM signals to control the current through 4 different IR leds. The bad news however is that I am still not completely confident in how this works. And I hate that.

I can't find much info on this particular chip over here, so hence this (I hope) start of a discussion to get some more data on these 'new' PWM peripherals.

AFAIK from the DS the 18F55Q43 has 3 individual PWM units. Every unit has 1 'slice' and every slice has 2 outputs, making for a total of 6 outputs. Things like frequency are set for a unit (pwm1..pwm3), justification is selected for a slice and duty cycle is set per output.

Relevant code:

Ttelmah · Joined: 11 Mar 2010 Posts: 19539

There are actually six PWM's available, not just three.

There are 3 16bit PWM's which are as you describe. But there are also
three CCP modules which can also be programmed to produce PWM
outputs. These though use the timers for their clocks (just like the
normal PWM on most other chips), and only give 10bit available resolution
versus the 16bit on the modules you are using.

Yes, the CCS functions do have to program both duties at once, but if you
are only using one slice just leave as zero. If you look at these registers
they are marked 'DMA access only'. It doesn't appear that they are available
in the standard memory map!...

Thing that is not made at all clear is the PWM numbering used. CCS, seem
to have 'forgotten' that they also need duty settings potentially for both the
CCP PWM's and the 16bit PWM's....
I'd actually contact CCS, pointing this out, and asking for an example
using both the CCP PWM's and the 16bit PWM's.

The issue here is that this is not just a new chip, but it carries a very
large number of previously unimplemented peripherals. So where CCS
can add a standard peripheral they have done so, but they have not yet
actually written the code to handle some of the new parts.
You really need to talk to them.

Woody · Joined: 11 Sep 2003 Posts: 83 Location: Warmenhuizen - NL

Wow, I managed to completely miss the CCP-with-PWM possibility. That can come in handy as I might need 8 PWMs somewhere along the line. Thanks for pointing this out.

In my application I have pwm outputs all the time, only every now and then I need to adjust one of the outputs. I cannot let the other output on a slice be 0 so that forces me to write both outputs if only one changes. Not a showstopper in any way but to me it would be more logical if I could only write one output at a time.

I contacted CCS a fortnight ago about a PWM example. They sent me one using the pwm peripheral on this chip but I have to admit that the code was a bit beyond my working C. Lots of uint32_t, cTicks and what not.
I am basically just an assembler guy using C for readability and by then had my own solution working and was very busy with the rest of the code so I have not yet tested and toyed with their example.

Anyway, I agree that CCS needs contacting but (as was the case with my earlier I2C hassle) I first have to formulate sensible questions. In the meantime I use what I have and hope the next compiler update will solve some of these issues.

Paul