Yamaha electronic keyboard—some keys work, some don’t

[iXod]

Yamaha DGX-630 / YPG-635
- - -
The power amp IC was dead;  replacing that was straightforward. I presume this instrument had multiple failures that must each be repaired so as to discover the next one. 

Now only some of the 88 piano keys are functional. It’s octave related: eg, the E and F-sharp keys are the only keys that work (in several—but not all—of the octaves). All other of the 88 piano keys are non-functional.

I replaced the FFC cable between the key-contacts PCB and the “DM” processor board. I cleaned the contacts in the FFC connectors and added a little Deoxit cleaner. Touched up solder joints on the key-contacts PCB. Opened the rubberized key-contact strip (each contact strip covers approx. one octave of keys) and cleaned the rubberized contact and its PCB contact using dilute dish soap and water and dried prior to reassembly. Replaced all electro caps on the “AM” PCB (contains the power amp and several v. regs). No change. All DC voltages are correct.

The accompany (“the backup band” automated play) and other digital features all work. Still, most of the musical piano key presses are not working.

Anyone else diagnosed such a problem? Ideas?

Thanks.

[bob91343]

I am guessing that the keyboard is okay but there is an electronic problem.

[perieanuo]

other ideas beside veryfying the matrix connections continuity (rows, columns) and verify the resistive contacts or whatever switch you have there (along with the resistors and tracks on the pcb), like the pub, "what else"?

[Audiorepair]

It's almost always dirty rubber contacts and/or the contact PCB.

If you have any Isopropyl alcohol, clean the rubber contacts and PCB pads with that, I'm not sure soap and water is a good idea.

What also happens is that drink gets spilt onto the contact PCB, or some other corrosion happens, so that some of the many many traces become open circuit. 
Inspect it visually for areas of possible corrosion, you should be able to measure continuity across the whole PCB without difficulty.
This will generally cause blocks of notes to stop working, e.g. all the G's, or every 8th note or something, as the contact arrangement is a row and column matrix.
And resolder the ribbon connectors, these can get dry joints/corrosion too.

What you can also do is to swap the most worn contact strips with the least worn ones, this can sometimes get notes working again and may extend the life of the keyframe considerably.

[iXod]

Thanks for your reply. I’ll try those suggestions.

When I press any of the rubber contacts I read approx 400 . What should this resistance be?

Thanks again.

[Audiorepair]

Thanks for your reply. I’ll try those suggestions.

When I press any of the rubber contacts I read approx 400 . What should this resistance be?

Thanks again.

Don't know, I've never needed to measure it.
The note either works or it doesn't, the resistance will be highly dependent on pressure.


Incidentally, if you don't have any MIDI monitoring equipment, you can shove an LED into the MIDI out socket into the 2 pins either side of centre.
It is polarity dependent, so it will only light up  one way round. 
This will give you a visual indication of whether the keyframe is triggering notes or not, just in case the problem is actually not the keyframe.

[iXod]

If the keyframe OR the associated key-press sensing circuit is at fault (which I suspect), the LED won’t be any help.

I tried triggering notes by jumping keys using a 200 resistor across the 2 pins of IC17 (Yamaha 44QFP) for a key. Tried several, none that I tried resulted in any notes.

The inputs I can check are present (several V’s, grounds, osc.). The outputs are not easily confirmed if I can’t trigger key-presses. That’s why I’m circling IC17 (the keyframe scanning circuit) as increasingly suspicious.
 
The oscillator for IC17 is generating a clean 5 MHz clock.

Thanks.

[Audiorepair]

Not sure if your attempts to trigger notes by shorting pins on the IC would work.

Normally each note has two contacts, both with associated diodes as part of the switching matrix.
One contact is made first by depressing the key, then the second contact a short time later to give the time difference between the two contacts, which is used to work out how fast  the key has been pressed to generate the correct volume of that note.

If you just short a random contact, the note might never sound.

If there are no diodes in your shorting circuit then all bets are off.

[iXod]

Each contact has a diode in series (x 2 contacts = 2 diodes per key).

Also confirmed: closing a single rubber key contact (instead of both contacts for a key) will not sound a note. Both must be closed.

[perieanuo]

Thanks for your reply. I’ll try those suggestions.

When I press any of the rubber contacts I read approx 400 . What should this resistance be?

Thanks again.

that's up to what type of conductive rubber they used. some of them are sub-ohmic, some of them into kohm range.
the idea is they are paired with matrix controller ic, so all you need is to measure a good responsive key and take that as default.
or just test the pads with something conductive, anything, aluminium foil for example or just the tweezers without scratching the contact pads.