Microprocessor ST9036 sometimes doesn't start

[MidiMania]

I am having trouble with a Midi controller PCB in a Studiologic MP-117 pedal board. It bases on an ST9036 Microcontroller. The code is in the internal flash memory of the controller and I cannot see or access it. I also have no schematic of the PCB but its a rather small circuit.
Here is a foto:


The problem: Sometimes the controller simply won't boot up. Once it is in that condition, even a manual reset at the RESET pin doesn't work.
If it boots, reset works fine by manually shortening the Reset pin of the uC.

What I have done so far:

1. I have played with the timing of the reset input which is fed by a 10k resistor with a 1uF cap to ground for slowing down the ramp-up of the reset pulse. I have added 4.7uF, 10uF, 47uF making the start reset slower - no clear effect.
2. I have checked the timing of supply voltages and it looks proper. Have added some electrolytic capacitors before and after the internal 5V regulator for making the ramp-up phase more defined, there where some instabilities on my scope.
3. The oscillator (25MHz crystal) seems to work well, same behavior in both conditions
4. There is for sure no mechanic problem like solder joints or so. The behavior changes without any mechanic influence on the PCB.

I am wondering if someone has an idea. What kind of condition is it, when a uC doesn't start and even RESET doesn't work?

 

[JohnRoberts]

I have never worked with that processor family but I ASSume that they all use similar firmware. It appears that the MCU is getting locked up in some invalid state that does not respond to a hard reset.

I suspect you are on the right track looking at the POR(?) power on reset. IIRC the POR holds the MCU turned off briefly during power up to give the innards time to settle.

Good luck... a data sheet for that MCU should provide preferred POR timing and pro forma circuit.

JR

 

[Philbo King]

One thing to check for is a latchup problem: Verify that no signal pin is ever outside the power rails range during powerup.
If you find any, one quick & dirty fix is to attach a small Schottky diode in reverse bias from the signal pin to the power rail, or to ground, as the case may be.

The non-dirty fix would be to modify the power supply startup so the thing sending in the signal doesn't start until the microprocessor is already started. That is usually a bit more complicated.

 

[MidiMania]

One thing to check for is a latchup problem: Verify that no signal pin is ever outside the power rails range during powerup.
If you find any, one quick & dirty fix is to attach a small Schottky diode in reverse bias from the signal pin to the power rail, or to ground, as the case may be.

The non-dirty fix would be to modify the power supply startup so the thing sending in the signal doesn't start until the microprocessor is already started. That is usually a bit more complicated.

I was about to give up but these are useful ideas, thanks. There is actually a pin that is suspicious. It is an ADC input and it is used for an optional volume control feature. It is getting some voltage from somewhere (I have no schematics) and I should be able to figure that out.
I'll check this in the coming days and will report any outcome.

 

[JohnRoberts]

I was about to give up but these are useful ideas, thanks. There is actually a pin that is suspicious. It is an ADC input and it is used for an optional volume control feature. It is getting some voltage from somewhere (I have no schematics) and I should be able to figure that out.
I'll check this in the coming days and will report any outcome.

Sometime MCU i/o pins have active pull-ups enabled (small current source pull-up). If internal, pin voltage should not be higher than MCU rail.

JR

 

[MidiMania]

Just wanted to report that I have given up on this one. Feels like waisting time without any value. There is a newer version of the board available as replacement for 45 €. I guess the older version exposed these issues not only in my case.
I do this very rarely but this time I go that way.
Thanks for all the hints! I'll keep them in mind for next time.