Restoring a rubber-dome IBM Model M2 keyboard

erin-pinheiroErin Pinheiro wrote 07/16/2019 at 01:17 • 4 min read • Like

Not a project, but wanted to share something I did the other day.

Emma, my trusty Model M keyboard has been ill with a case of rotten membrane contacts since the day I was given her for free. Some conductive ink helped mitigate it, but without a proper method to paint the traces decently and a good quality ink, this was only a temporary (albeit long-lasting) solution.

I tried a few last-resort methods of exposing the traces underneath the conformal coating before retiring her, which is why it looks even worse for the wear in this shot. While I haven't fully given up on her yet, I needed a temporary substitute.

Meet Marlena, Emma's younger sister. Despite being an M2 keyboard, sadly she's rubber dome and not buckling spring. Still, she feels like a fair substitute to her older sister. I rescued Marlena years ago from being tossed away from a hackerspace I worked at. The problem is she seemed just as reluctant to work So as Emma, so now I have two defective keyboards on my hands...

So I started to take Marlena apart, and in doing so I found some interesting cost-cutting methods I'd never seen in a keyboard before. Where metal braces would be on longer-width keys such as spacebar and backspace, Lexmark molded plastic springs instead. Except for the enter key, which had the standard metal brace for some reason, go figure.

Time for a well-earned bath... the keyboard had seen better days, I never properly cared for it since I rescued it years ago (which is why I never noticed it wasn't working). I thought it was only dirt keeping it from working right, but there was more to it than met the eye...

When I put everything back together once it was dry, half the keys still wouldn't work. In fact even with a clean membrane underneath, pushing on it directly without the rubber layer, some keys showed no sign of working at all... until I pushed down on the PCB. Ah. These membranes connect to the edge bus on a controller PCB through pressure alone, and on Marlene's case, this was provided by a piece of foam that was long since gone into mushy hell. First, I cleaned the contacts on the edge of the membrane, and then padded the space under them with a few bits of cardstock, which properly pushed them together.

...Yet, still not quite there. There was only one column completely not working, so that was an indication that something was wrong in the membrane traces, and it was time to break out the continuity tester. Sure enough, there was a break somewhere along the way, thankfully in a spot where there is no conformal coating. Nothing a bit of conductive ink couldn't fix.

And presto! That revived the last stubborn column and now Marlene is happily working again, sans a plastic clip I accidentally broke while trying to remove the decoder PCB. And now we're happily working together writing up articles while Emma awaits a new repair. :)