I’ve wanted a digital piano for some time now, and I still occasionally glance over the ‘spares and repairs’ category on eBay in hopes of a cheap deal. Finally, I found one:
According to the previous owner, when she bought the piano, it was working fine. Over time it developed a crackle and eventually stopped producing any audio output whatsoever. It does however turn on and the digital stuff appears fine, the screen and buttons were all working.
Opening up the keyboard, I found a problem pretty quickly. Leaky capacitors:
Electrolytic capacitors are composed of a thin sheet of (usually) aluminium bathed in electrolyte. At the factory a dielectric oxide layer is formed on the aluminium by applying a voltage to it and thus creating a capacitor.
Over time the electrolyte in these capacitors dries out, usually causing the capacitance to decrease and the impedance to increase. In some cases, gas build up inside the capacitor and then bursts, leaking electrolyte onto the PCB, which is what has happened here.
The electrolyte is corrosive and as you can see from the photos below, can be quite damaging to the copper traces around it. Nearly all of the capacitors on the main board have leaked and corroded the copper around them, so they all had to be replaced.
The worst offenders were to be found in the audio output stage of the board, directly after the DACs and before the connector for the amplifier. There are two 5V regulators here with electrolytics around them, so I’d imagine the heat from the regulators accelerated the leaking process and caused the worst corrosion.
I started off removing the faulty capacitors with a soldering iron resulting in a few lifted pads, and soon switched to a hot air station. The electrolyte seems to ‘stick’ the caps to the board so this wasn’t as easy as I had hoped but in the end, I got them all off.
Reading around on electronics forums seems to show that deionised water is the best option for cleaning off the old electrolyte so I brushed the board down with water and alcohol. After removing the solder I ended up with a slightly cleaner board.
Unfortunately, the electrolyte had actually corroded away a number of traces and vias completely, which is almost certainly what caused the original problem. I fixed these traces with mod wires and solder and replaced the lifted pads with solder braid where necessary. I had to remove the -5V regulator to clean up the corrosion beneath it and in the process snapped off a leg. So that has now been replaced with a TO-220 package.
In some places the electrolyte had migrated along the tracks underneath the soldermask. In these cases I tried to scrape off as much soldermask as possible and clean the copper underneath to prevent further corrosion.
I ended up replacing the caps with a variety of caps, mainly because it’s much easier to solder them in (especially where broken traces and mods occurred) and there weren’t any particular space requirements.
Turning everything on, the keyboard works fairly well, there is still a slight tone over the notes being played, this is barely noticeable unless the treble is turned up but I will try to diagnose it eventually.
I have been thinking about designing a USB midi board to interface with the keyboard matrix, allowing me to replace all the aging electronics with a single board linux computer, but this is probably some time off.