async communication dspic33ev256gm004

[Ttelmah]

Silly thing to try. Use the internal oscillator, instead of your crystal.
If it gives correct frequencies, then it might suggest something silly
in the oscillator (like the crystal itself), while if this gives the same error,
it suggests something in the PLL setup.

[chaphill]

Ttelmah, I think your suspicions about the necessity of the gymnastics to change the clock setup are probably well founded. I was reprogramming several boards with code based on this setup so I just reprogrammed the "suspect" board with the same program All four of the ones I programmed that were normal gave a solid 49.997 or 8 hz for the frequency of the timer output, while the "suspect" gave 45.6 hz and the osc2 pin showed a clock of 36.5MHz compared to a 40.09 MHz for the other four. I checked the input oscillator and it was 7.3728 MHz, all measured with the same oscilloscope. What I am going to conclude from all this is that the chip is definitely bad, and is getting worse.
It may very well be that the default setup that CCS performs is just fine, but I learned a lot about how the clocks are set up on power-up.
The really bright spot in all this is that I found that I can determine the baud rate, number of bits and a lot of other "settings" of an unknown telemetry with my logic analyzer. It had no problems whatsoever with my legacy 78125 baud 9 bit communications or determining that the "bad" board was trying to communicate at 110 kbaud. I had no idea it was so versatile. Thanks for all your input! -Ray

[newguy]

A friend recommended getting a Saleae recently. Whenever we'd be on a teams call, he'd share his screen and go over the logic analyzer traces he was sharing. I had never seen anything so versatile or easy to use. I got one myself, and I have to tell you it's the most useful development tool I now own. Simply awesome.

Regarding your problems with the dsPIC - are they all the same revision? You can check the processor ID via the CCSLoad interface. I'm curious if they're all the same or if the problem unit is a different rev. I also have to ask how strict your ESD measures in your lab are? The kind of stuff vexing you now, in my experience, is usually caused by ESD.