Updated Mar. 12, 2020
Charger interface board debugging:
...
Remove R6 to check if current leaking through the 5V, Opamps, and through D5.
Result: 8mA current. This is the LED current (1mA), the opamp supply current (<1mA) and the Iq of the 5V regulator (5-6 mA ish from datasheet). No board heating
This now makes sense why the opamp was heating up → 12V in to opamp, 2.5V, 50mA out, through the diode D5 to the 3V3 rail, and that’s not an insignificant amount of heat.
Also try removing D5 if this does not make any changes. (after replacing R6)
Result: 13mA current draw. This is due to the extra current from the resistors used as voltage dividers after the opamps. So everything still works properly.
If this fixes it, then the opamps should be fine (as the inputs are both at 0V thus there is no leaking through them), and it is something downstream causing the issues (MCP3515 or another board component).
If we are still having issues, then try removing the opamps and/or the MCP2515
Removing MCP2515 (after replacing other components)
Current consumption stays at 13mA.
Resistance from 3V3 - GND is now on the order of 4kOhm
Now replacing the TCAN332 (with the MCP2515 still removed)
Current consumption from 12V is 14mA. The 3V3 line is at 2.5V.
Pinout of the MCP2515 on the board was compared to the datasheet pinout, they are correct.
MCP2515 removed, the resistance from VDD to VSS of the chip was 65ohms, compared to a new chip that was 10M+ (or something crazy like that).
VERDICT: Dead MCP2515 - how did it happen? No clue - could have been a bad DC-DC on a controller board that exposed it to too high of a voltage? Could have been a probing error that connected it to 12V?
The MCP2515 was replaced, and the resistance 3V3 - GND = 4k (ish), current draw from 12V = 14mA.
Board is now good for firmware testing!