Yamaha P45 keybed matrix scanning issue

Synth Repair
1

I have a Yamaha P45. Certain combinations of notes result in a spurious additional notes sounding. This repeats systematically across half-octaves (every 6 notes). However the exact nature of the fault seems to vary from one day to the next. At the moment, if I play a major triad in with A or E-flat, and add the 4th, I also hear a flattened 5th. Sometimes this only sounds if the bass note is also held down an octave below. Last week it was doing more or less the same, only with F and Bb.

I have seen various videos and forums online of people reporting similar issues, with humidity being cited as the cause, and the solution to install a dehumidifier. I do live in a very humid climate by the coast, and the problem has been more noticeable during more humid periods. However, it didn’t do this from new, rather after 4 years in a humid climate. So it seems likely that this environment has led to corrosion damage somewhere, which is causing signal integrity issues with the matrix readout from the keybed, worsened during periods of humidity.

The question is what is likely to be corroded or damaged? I have taken it apart and cleaned the three boards with the sensors. There was a lot of dust that had accumulated, and at first I thought this had fixed the issue. But no. So I took it apart again, removing each of the three PCBs and cleaning them carefully with IPA. I also cleaned the contacts for the FFC cable between the central board and the logic board. However, this didn’t seem to make any difference.

There was some slight discolouration of the contacts on the FFC, potentially indicating corrosion, but nothing that was obviously the cause of the problem (although I guess any discolouration is bad). I cleaned these contacts with IPA and a q-tip, although this didn’t seem to make much difference to the appearance. The three PCBs didn’t have any obvious corrosion; the resistors on the underside look like new, and the areas under the rubber sensors look like new. However, away from the rubber sensor actuators, the traces (which are covered with a protective polymer layer/lacquer) do appear dull in places.

The solutions appear to be (1) to replace the FFC, (2) replace the central of the three PCBs, as the other two are connected to this and since the fault repeats across the octave range it suggests the central board may be the faulty one, (3) replace all the PCBs and the FFC, as since it’s a matrix readout issue, maybe it makes no difference which board has the fault, and I may as well replace the FFC while I’m at it.

Any thoughts/insight appreciated. It’s a pain removing and reseating the rubber sensor actuators, plus it will be north of $100 to source the three boards, so I’m hoping to get away with the simplest and cheapest option.

2

Have you tried to check the solder joints and test the Diods ?

3

I didn’t check continuity/electrical function (I don’t have a meter). But I visually inspected them, and the diodes themselves look like new. I couldn’t see any cracks or obvious signs of corrosion on the solder joints, although some of them had a dull appearance.

4

My suspicion is the problem is in the keybed itself. Much as inspecting the rubber key contacts is a pain, it would be the first place I would look. Never hurts to clean the contacts. There could be some bit of something floating around in there to cause that in differing places. Also look to see if there is some mechanical problem with the action that would cause an adjacent key to fire. And get yourself a meter…does not have to be expensive. That is the only true way to check diodes.

5

You are on the right track - keys are scanned in sections of 6 - hence if there is a problem at one note it can affect every 6th note from there onwards (could be up the scale or down the scale or across the entire keyboard). A further observation is the point where the scanner enters the keyboard - the key contact boards are divided in to 3 section - Low, Mid and High. The scanner cable enters in the Mid board at C#2 (where C0 is the lowest C). That can help to further eliminate if the issue is across the lot or just Low somewhere or just High some place.

Remove all the keys and inspect the key contact boards - look careful for any corrosion specifically around the point where diodes are soldered to the circuit board below.

The effect you are describing is most probably caused by corrosion of solder joints or tracks. If the corrosion grows large enough it will start touching pins adjacent and then cause all sorts of problems.

You may also look at discoloured traces - if traces are black, not the usual green, then there is track corrosion caused by black mould growing under the lacquer.

The key contact boiards are available from YAMAHA and are not expensive.

6

I replaced the flat/ribbon cable that connects the board with the key contact pads to the logic board. This seemed to do the trick, and the problem has not returned. I’m surprised that this worked TBH, because it didn’t look very corroded, with just some mild discolouration.

Hopefully a cheap and easy fix for someone else with this issue, as it seems to be fairly common.