tris bits and ECCP...

[valemike]

For those of you using the ECCP in full bridge mode, you are instructed to make the RD7:RD4 pins as all inputs while you set up the ECCP, then turn them into outputs after your first PWM cycle.

I'm wondering what would happen if:
- I configure the ECCP in full bridge mode, whether it be forward or reverse mode, and only configure the two active bridge pins as outputs, while keeping the other two inactive pins as inputs.

Doesn't the ECCP run internally regardless of the TRIS settings; if so, then i can change my TRIS pins on the fly. Hopefully this gets rid of all ECCP glitch woes. (Yeah i've read a bunch of errata notes on the PIC18F458 and other PIC families).
I think i am experiencing a problem where updating the ECCPR1L to a non-zero number is initially accompanied by a glitch in the duty cycle. It happens so fast and i'm not that good at triggering a scope to only capture one transition, so i'm just playing it by ear.

[PCM programmer]

Buy or rent a logic analyzer. Then you can see the pulses, and
maybe solve the problem.

[valemike]

Yeah, it's probably cheaper in the long run than to spend R&D hours scratching my head.