Is there any way to troubleshoot a PIC in a consumer device?

[Andy1845c]

Hello,

My wife has a Kurig coffee maker that died. I opened it up (apparently they don't want you doing that as it takes forever) thinking it would likely be a power supply issue. Doesn't seem to be. I have spent several hours reverse engineering the board and it keeps coming back to the PIC 16F677 that's the brains of the board. I have 5v between Vss and Vdd and have determined what pin the power button is connected to and that it is indeed pulled high when the button is pressed. There are several small LEDs that appear to be driven directly off the controller pins that should light when the power comes on, and they never do, so I am really suspect of the PIC.

Is there any further testing anyone can think of I could do before I throw the whole unit away? Mostly out of stubbornness as I have hours invested in this now.

I used to screw around with these controllers when I was younger but its been forver. Was there a wat to put a scope to see if the ocillator is running? Would that even tell me anything?

 

[sagor1]

If the power supply is ok to the PIC, and the PIC is not behaving like it should, it may be "dead". However, check that pin 4 is not being held low. That is a reset/clear pin. If it is low due to other component failure, the PIC may not run.
Oscillator may be internal, and unless they send some clock signal out one of the pins, there will be no "clock" signal anywhere to measure. The only pin that can be easily set to output a clock signal is pin 3.
Other than that, probe around with a scope to see if any of the pins are changing, and if so, look it up on the datasheet to see if it makes any sense. Not likely, but worth a try before throwing it out.

 

[Pommie]

Can you post a clear picture of the board?

Mike.

 

[Andy1845c]

MCLR is pulled high.

I don't see any external oscillator.

Here are a couple pictures

 

[Pommie]

Looks like it uses the internal oscillator (no crystal on pins 2 & 3) and no ICSP connector so not much you can do.

Mike.

 

[Andy1845c]

Looks like it uses the internal oscillator (no crystal on pins 2 & 3) and no ICSP connector so not much you can do.

Mike.

Kinda what I figured but I thought it was worth picking the brains of some people who know these better than I do.
Its a nice PCB. I wish there was a way to get a schematic and the source code for the PIC. haha

 

[Andy1845c]

ICSP connector

Now that you mention that... the round pads to the right of the PIC.... do you suppose that is what those would have been for? One is tied into MCLR... didn't trace the others. I assumed it was for a connector that this model doesn't use but maybe it was to program the controller?

 

[Pommie]

An interesting project would be to work out the schematic and program the pic yourself but without ICSP it's pretty much impossible.

Mike.

 

[Andy1845c]

An interesting project would be to work out the schematic and program the pic yourself but without ICSP it's pretty much impossible.

Mike.

I honestly considered this... I was thinking I would have to program one and then solder it in. I only played with the DIP controllers, never got into SMDs

 

[rjenkinsgb]

Now that you mention that... the round pads to the right of the PIC.... do you suppose that is what those would have been for?

That was my thought, looking at the photos. Only five pins are used on ISCP / ICD connectors.

Check for AC across the DC power, and on the reset pin - if the PSU is failing (eg. capacitors dying) it may be getting ripple or power dropouts that cause continuous resets at the CPU?

 

[Pommie]

Now that you mention that... the round pads to the right of the PIC....

If it is (and it looks about right) then the connections should be Vpp (MCLR), 5V, GND, Data (pin 19), Clock (pin 18). Do you have a pickit at all?

Mike.

 

[Nigel Goodwin]

Now that you mention that... the round pads to the right of the PIC.... do you suppose that is what those would have been for? One is tied into MCLR... didn't trace the others. I assumed it was for a connector that this model doesn't use but maybe it was to program the controller?

I would say it absolutely certainly is, it's a fairly standard way of doing it.

However, processors in general (and PIC's in particular) are EXTREMELY reliable, it's very rare to find a duff one - and even then it's usually been damaged by external events. In my many years of servicing I've replaced numerous processors, and most of the time it's not helped - the rare exceptions are things like a particular model of Sony CRT TV, where the LOPTX arcs internally, and takes out the processor.

If you've got a PICKit you might try and see if you can read the existing PIC, although it's quite likely it's read protected, so you may not be able to read or copy it.

 

[Diver300]

Now that you mention that... the round pads to the right of the PIC.... do you suppose that is what those would have been for? One is tied into MCLR... didn't trace the others. I assumed it was for a connector that this model doesn't use but maybe it was to program the controller?

Those look like they are wired so that the processor can be programmed with a PICkit. I would certainly try connecting to those and reading the PIC. It will probably be read-protected but it is worth a try.

 

[Andy1845c]

I totally still have my PicKit2 someplace. Haven't used it in probably a decade. I will have to dig it out and relearn how to use it and see what I can come up with.

Thanks everyone for your responses!