Oberheim Matrix 6 CMOS Latch-Up
Have any of you have ever turned on a M6 and got the following symptoms: strange sounds out of the voices, droning, no key response besides popping occasionally, calibrate shows that EVERYTHING fails (BAD OSC 7, BAD WAVE 63, BAD RES 63, BAD VCF 63) even if you test the chips and get good signals?
Hint: You can do Master Edit -> 'D' -> "Ready?" -> Hit "yes" and hold down the 'C' button under the display. This will print out any errors found during the calibration routine.
Good news is that it is probably just one pin of one chip that caused the whole thing.
So... the CMOS ICs (4051, 4053, etc) get their power from a 10k resistor and diode divider setup between +12v and +5v, however the schematic erroneously shows -5v instead of +5v.
Because of the resistor this supply will have a small delay before it reaches full 5.6v. Now, on one particular IC (u711 4053 triple multiplexer) there is an unused channel that has its input tied to +5.
The +12/-12/+5/-5 will all switch to full power much faster than the divided 5.6v. What happens is for a short time the 4053 sees a higher voltage at the input than the power rail supplying the chip. The IC latches up and sinks current, causing the power rail to be pulled down. The chip will get very warm or even hot. On the M6 I worked on the 5.6v output read around 3.1v and, because of this, all other CMOS would not function correctly.
I replaced the 4053 as the current sink would have damaged or killed the chip anyway. Then I cut the trace to pin 5 which supplied the +5v and connected pin 5 to 6 (GND). The pins were shorted on the underside of the board.
After doing this modification, everything calibrated right away and plays perfectly.
It may have been that the chips originally used were more resistant to latch up than newer ones. However, frequent power cycling would kill the older ones anyway from the unbalanced voltages.