32MX PMP apparently failed due to incompetence
On three generations of a 32MX795F512 board I had been supplying the AVdd with its own dedicated regulator and, using the board to run a TFT display through its 16 bit PMP, I was getting an intermittent problem initializing the display. Since I actually don't need to use analog for the time being, I have followed the advice from people here and supplied the AVdd from the same regulator as the Vdd - the rationale being that the chip might otherwise be harmed if the analog and digital power were not applied simultaneously. I was grateful for the input. I have been doing this same thing with a lot of 24FJ applications (which needed analog) without failure but I have now learned my lesson.
After fixing, the intermittent problem with the display is still present with two of the boards so I've run a check, sequencing all the PMP ports (configured as GPIOs) on/off and checking each in turn. Several GPIOs seem to have failed, remaining either on or off so I have to assume that my earlier bad practice must have damaged the chips - although I'm surprised they work at all.
On all three of my boards I am finding that RE5,6,7 have apparently failed - RE5,6 remaining on and RE7 stays off.
Looking at the spec sheet I see that RE5,6,7 have only two functions - GPIO and PMP so, before I condemn the 32MX to death and hand-remove it (not a happy task), I would like to confirm that setting the three TRISes to zero and putting PMCON=0 ought to free these three ports up as GPIO outputs? Being fairly new to digital I am always surprised how tough these chips are and how often an assumed hardware failure is actually down to software.
post edited by Spottymaldoon - 2019/02/03 15:19:17