Red and Sticky

So recently I came upon a fantastic find, a Roland D-50 listed for sale at an absolute bargain price with those four little words I love so much: For Parts or Repair!

Wait, wait; actually my bad, this is the Roland D-70, the much loathed and resented synthesizer that was based on a completely different line of synths, and only gained the D- prefix when Roland realized that people really loved the D-50 and decided that phoning in a follow-up would be a really great plan.

It was actually the bridge between the Roland U-20/U-220 (and can even take the same expansion cards) and the later JD-800 in terms of lineage, and other than a few notable faults (like an underpowered CPU that can bog down under high load leading to slower than usual ADSR envelopes and LFO modulations, and the sound structure where tones are shared between patches) it wasn't a terrible synth. It just wasn't a D-50, a crime for which it will never be forgiven.

Anyway, that's really neither here nor there. This one is mine now, and it needs some fixing, so click through the break to read more.

As you might imagine from the bits of blue tape on the keys, not all of them are producing sound. This is because the keybed shares a fault common with the U-20, the JD-800, and a few other Roland synths of the time period.

You see, this keybed is what's known as "semi-weighted", which means that it's a pretty regular synthesizer keybed, but the keys have little metal weights glued to the underside to give them a more substantial feel, a bit closer to what you might get with a piano (though, in reality, it was really just a minor step up from a basic synth key feel, and nowhere near a hammer-action keybed).

The glue they used, however, would turn out to be a bit of an issue, as due to a fault in its chemistry, it would break down over time and ooze out over the underside of the keys, dripping down into the mechanism and making a huge mess of things.

And this is the source of the problems with this unit. So, as you might imagine, step one is to get the keybed out.

These little plastic strips on the back need to come off to free the keys from the frame.

They're held on with double-sided tape and just peel off pretty easily.

Once they're off, the keys pop out by tugging them forward gently and then lifting up in the back. This can be slightly more difficult on the black keys as the metal weights in them can drop down in front of the centering post at the front of the key.

And that makes sliding them forward a little tricky.

Once the keys are out, we can see how the red glue has dripped down onto the key matrix strips, mucking them up considerably.

This actually isn't as bad as it looks. The glue is relatively harmless just sitting on the surface of the coverlay like this, though it does tend to wick underneath the key contact strips and get pretty close to the contact terminals underneath. Getting those contaminated with melted epoxy would not be good, so cleaning this up is on the to-do list.

To do this, we need to pull the silicone contact strips off, which you should do carefully so as not to accidentally tear one of the domes clear off the strip.

No comment.

But once everything is disassembled we can get a better look at things, and start cleaning the two parts separately.

Or three parts, technically, but the contact strips are really easy to clean. They're made of silicone, and literally nothing sticks to silicone, so the epoxy just rubs right off without any effort while you're putting them through a regular cleaning with some dish soap.

Anyway, to get to the whole reason I actually bought this keyboard: I wanted to do science. It's been widely known for a while that soaking the keys in a bucket full of lye solution will strip off the red glue and leave the keys sparkly and clean, but that has the downside of having to deal with a bucket full of lye solution, which is not ideal in a number of ways.

Some people have experimented with washing soda, which is at least a little less offensive to your ability to continue to have skin on your body, but still a bit harsh to work with.

Instead, I wanted to try using borax, which is safe to handle with only a few of the usual cautions (like don't drink it, or pour it in your eyes), doesn't smell bad, and is cheap and readily available at most grocery stores.

So for the sake of science, I started with a white key where the weight had fallen out and become hopelessly lost.

And into a saturated solution of borax and water it goes.

And the next morning...

It's spotlessly clean.

It even works pretty well at cleaning the red glue from the chassis of the keybed, as can be seen in this quick test.

I also confirmed that ammonia works, with the downside of being offensive to the senses and less safe to use around nonferrous metals like copper (on the PCBs) and aluminum (such as the main chassis of the D-70). So it'll do in a pinch, but borax is still my recommendation.

Anyway, the rest of the keys need cleaning, so into the big orange bucket they go for a little soak.

Cleaning the key matrix strips would, however, be a more manual task.

I found that working on long sections at once was a good tactic: I'd wet down a whole long stretch with the borax solution with a cotton swab, then when I came back around to the beginning again the surface of the glue will have started to break down and I could agitate more of it away.

Also note in the above photo how the glue was wicking under the contact strips: it generally stayed away from the contacts themselves due to surface tension, but it's better safe than sorry.

Anyway, while I was waiting for the keys to soak, I did tend to replacing the tact switches. This task made a brief appearance on a previous blog entry.

The key matrix strips came out magnificently clean.

There were a few spots of glue in the felt that runs on top of the aftertouch sensor, but I decided I wasn't bothered enough with that to do anything about it. The sensor itself is pretty fragile and failure-prone, and nearly impossible to replace, so I decided it was best not to tempt fate.

And good news: I got a replacement for the torn key contact strip. Though, this wasn't without a little mix-up when the seller accidentally sent me the strip for the low notes instead of the one I needed for the high notes.

But we got it sorted out without any trouble.

So things are almost ready to assemble. The keys came out of the bucket nice and clean, and all the lead weights dropped out of the black keys. The steel weights in the white keys stayed put, since there wasn't enough surface area in the thin layer of glue around the edge for the borax to penetrate, but that's fine.

I made sure to line up all the black key weights to make sure that none had gone missing.

There was only one loose white key weight, which had dropped out sometime before it arrived in my possession but hadn't been carelessly discarded by the previous owner like the others had been. This turned out to be rather fortuitous, as it meant I could get accurate measurements on it to reproduce it.

I went with a volumetric measurement and determined that if I bought a length of 1-1/4" wide 1/8" thick bar stock, then cutting off just slightly more than a half inch would get me precisely the same amount of steel in each piece, and leave plenty of room around the edges to drop it into place.

Speaking of that bar, it cost me a whole $15 to purchase from the local metal yard.

And uh... it comes in 10 foot lengths. Worth noting.

But getting to the point, I need it in slightly-more-than half inch lengths, not 10 foot lengths, so let's put my portaband to work.

This SWAG stand is fantastic, by the way. Highly recommended. About the only shortcoming is that the hammertone finish on the top isn't terribly durable for something you'll be scraping rough metal across on a regular basis.

Anyway, marking out the correct length is easy with a 6 inch square.

The dimension of the cut is 0.521 inches, by the way.

And, uh... I may have had to move the saw to get it more than 10 feet away from my car hole door.

Just something to keep in mind.

The cuts came out pretty accurate though, if I do say so myself.

And with a few minutes of cutting I was able to chop up the 21 weights I was missing.

And I even have a little bit of steel left over to use in some future project.

Anyway, next up is deburring.

Because I like having fingers that aren't bleeding.

Comparing the original with the new, it looks like I hit the weight pretty much perfectly.

10 grams, within a reasonable tolerance at least.

Now I need to put all these weights back into place.

And hopefully this epoxy won't melt and drip out all over the keybed like the original stuff did.

I think this might actually be the first time I've used the mixing tips. Usually I avoid them because I only need a small amount of epoxy at a time, and the tip wastes a bunch inside each time you use them.

But I think I'll be going through enough to justify the extravagance this time.

No pictures during the install: this stuff only gives you 5 minutes to work with it, and I got a lot of weights to get through.

At the end, though, it seems my approximation was correct.

This was exactly the correct amount of epoxy. This is all that remained to be squeezed out:

And all the weights are glued into place.

But we still have some original weights with some red glue hiding behind them, and those weights might drop out at some inopportune time, which would be, you know, inopportune.

So with that in mind, I decided to break open the second package of epoxy I bought and squirt a bead around the edges of those weights to seal in the red glue and keep the weights from ever falling out.

And that's finally all the weights secured.

5 minutes later, once the epoxy was set, it was time to reassemble the keybed.

Or at least it would have been, had I not noticed that some of the keys had gotten a bit scratched, scuffed and hazed over the years.

A few of them, like this, were pretty heavily scratched and so I went at them with some micromesh first.

And then hit them with the buffer and some car polish (which, despite the fact that it's used on "metal" panels, is actually designed to polish the plastic-like paint that forms the actual outside surface of most cars).

They didn't all come out perfect due to the depth of some of the scratches, but they were definitely improved.

So now the keybed gets reassembled.

Then it goes back into the synth, where I discover to my brief dismay that 3 of the keys are dead.

Now that's certainly an improvement from where I started, when half the keys wouldn't even move, but it's still falling a bit short of ideal.

So out comes the keybed again, and off come the keys so I can get at the contact strip.

And the secret weapon in this case is MG Chemicals 838AR, which is, as the bottle says, a conductive carbon paint.

This is roughly the same stuff as they use at the factory to form the contacts in the first place, so painting on a fresh coat will renew the contact back to factory fresh.

You can also dab a thin coat on the underside of the carbon pill in the silicone dome contact strip.

With those contacts patched up and the synth back together again, I'm happy to report that all the keys work, and as a bonus, so does the after-touch strip!

So that's a job well done, and it's all thanks to the magic of borax!