One chip causes entire digital piano to become garbage (venting....)

e house and is rarely played. A few days ago some keys stopped working, an d the ones that do work are playing the wrong note. In addition, it plays this one specific note immediately after powering up. I don't play it howe ver, it's my wife's.

because if I hold down a key, the others that were working, still work. S o I take it apart.

y through-hole resistors, a few caps, a crystal, and a 40 pin DIP Winbond

8052-type micro. I believe it's called the keyboard processor. It connect s to the keyboard in 3 places, and a ribbon cable with approx. 20 positions goes to the main PCB off to the side with large SMD chips on it. I decide d to try something.

t still plays the same note on power-up. So I disconnect the keyboard PCB from the main PCB. I power up, and it behaves normally. Hmmmmm. I think it's safe to assume the keyboard PCB is bad. And it's probably the micro, although I'll have to look closer. Problem is, I can't get this board or c hip anywhere. Sure, I can buy the micro, but I need the code. The fix wou ld be so simple but the few kilobytes of firmware in there is what's standi ng in my way.

ctually disowned it, and of the two authorized repair centers, one can't ge t the part, and the other won't sell parts, and probably can't get it eithe r. Another website that sells this stuff also can't get one.

repair, if it can even be repaired, seems silly. How annoying when the de fective part is cheap and easy to replace, but is a micro. I've considered that possibly the code protection isn't on and maybe I can pull the firmwa re and program a new chip, but the chances of that working are almost zero. I'd hate to spend 50 bucks on the Winbond programmer with such a low like lihood of success.

e should I have to scrap the piano because of this.

ince we had a polyester blanket covering it, pulling it off might have gene rated enough static. Still, though, this seems a bit unlikely considering all the plastic it would have had to travel through.

Does cleaning the contacts help any?

Did you hang an oscilloscope on the VCC of that uC? Sometimes electrolytics dry out and then there's tone of 120Hz or switcher ripple on there. Does its clock still run? Crystals can age, fall out of their tiny holders inside the can or just plain corrode out.

I've had a Brother printer that slowly faded into lalaland because a supply voltage sagged away. Eventually a uC in there quit and the display froze up. But this time I didn't want to repair it anymore because it started to have a bit many paper jams anyhow, so I replaced it.

--
Regards, Joerg 

http://www.analogconsultants.com/

Look for conformal coatings on the board.

I've seen a few times where these coatings deteriorate over time and become conductive, fixed many RF problems in old circuits by scraping this crap off.

Also, if the PC board is of the cheap fiber type or that brown bridle type, it too can become conductive with moisture over time.. I've put them in ovens for a while at around 140F and fixed those.

The other problem could also be a floating input, due to a lost cap or pull up r.

Jamie

Are you sure they all play the wrong note, and not just half of them?

Perhaps you can figure out the protocol it's using, and replace the entire keyboard processor. The protocol's not likely to be complicated - it only needs to indicate which key is pressed, or released, and in the former case, how hard.

Depends on how you value your time, of course.

Sylvia.

Yeah only some might play the wrong note. But it's spread across the entire set of keys.

I have considered trying to reverse engineer it. We're in no hurry to fix/replace so I might give that a shot during my free time, which isn't much but I have thought about that.

I still have to put my scope on it and see what I can find. My guess is the uC got fried but I have assumed wrong before.

It had occurred to me that some of the wires are carrying a key code in parallel (otherwise, it sounds like a lot of wires for this application). If one wire is stuck high (say), then all the keys that should have that wire low will play wrongly. All the others will play correctly. If so, it probably wouldn't be hard to find. What to do about it is, of course, another matter.

Sylvia.

I have never used a 8051/8052, so here's a question: Can all micros running off this core use the same programmer? Some web searching isn't turning up much. My guess is no.

What I'd really like to do is feed some pulses or whatever it takes into this chip to see if it'll let me read out the firmware before buying a programmer.

If the product was made in volume the chip firmware might not be flash but instead in masked rom - may be extremely hard to reverse engineer. Have you checked all the parts around it or try cleaning up and baking out the board in case of contamination?

Same if flash because they may have locked it. But before going through all this hassle I'd first check the other things (VCC quality, clock, etc.). Also heating and cold spray. Normally a field-programmable that has become marginal over time then changes its behavior. If it doesn't then the chance that the problem is not inside this uC is pretty high and investing time and money into a programmer could be a waste.

Baking is a good suggestion. Cleaning also. Sometimes there is flux residue underneath that's hygroscopic or aggressive. Seeing if all data bits wiggle within voltage and timing specs also won't hurt.

About 90% of my equipment repair in the lab and elsewhere boiled down to contact issues and corrosion. Only once have I seen a programmables failure and that was the ERPOM in a Wavetek 23.

--
Gruesse, Joerg 

http://www.analogconsultants.com/

I'm sort of preparing for the worst, but will get my scope out and poke around first. FWIW I usually assume the worst and you're right, that is rarely the case in the end.

Interesting about the heating and cooling. I never thought that could work on a programmable chip; I always assumed the heating/cooling methods were for troubleshooting parts that actually generated some appreciable heat.

Six years is a long time ago for a digital piano. You might check to see how cheaply you can find the identical model used.

--
John

There's one on Ebay now for $190 shipped, much worse condition than mine, and the auction isn't over yet. That's 60% what I paid new.

Forgot to mention one more pathology: Sometimes it helps to alter the supply voltage slightly (usually an increase by 5-10%). If that makes a uC or other programmable come back to life then you know for sure that the chip has worn out.

--
Regards, Joerg 

http://www.analogconsultants.com/