A goal of the MeeBlip project is not only making our hardware open, but being open about what we learn making that hardware. We tended to keep some of that to ourselves at first, but now that we have something more meaningful to share, we’ll be a bit less shy. Getting things rolling, MeeBlip’s chief engineer James Grahame – no stranger to hardware design and manufacture – explains why making hardware can be … well, hard. Happily, with each mistake, you get a chance to get a bit smarter. But it can come down to something as simple as a ribbon cable. I’ll let James tell the story. -Peter
Designing and building boutique music hardware is hard. Anyone who tries to tell you different is either trying to sell you something or hasn’t tried it themselves.
A case in point — we shipped about a hundred MeeBlip SEs in mid-March. They started to arrive about a week later and I received an email asking for help to get a Quick Build kit running. That in itself isn’t that unusual, and we can usually get stuff sorted out in a couple of back and forth messages.
But this message was followed by another. And then another. For some reason, it seemed like quite a few boards were arriving DOA. That’s a rare occurrence, because the final assembly step before they leave is to program the latest firmware and play them for a minute or two.
The next few hours passed in a blur — carefully inspecting one of the boards from the batch for defects, spot-checking the power supplies, confirming that the firmware programmer worked and that the files were up-to-date. Everything checked out, leaving me completely stumped. There was nothing wrong, yet people continued to report that their synths wouldn’t power up.
In a final act of bewildered desperation, I decided to assemble a Quick Build kit from our inventory. It took about 20 minutes, and everything seemed to go smoothly. When I plugged it in… nothing. The power LED didn’t light and the synth didn’t make a sound. It was stone-cold dead.
It didn’t take long to track down the culprit: the ribbon cable that connected the little daughterboard on the back panel to the main circuit board. Each end of the cable plugs into a plastic connector that only fits in one direction. It’s designed that way to make it foolproof to assemble. Or so I thought.
Unfortunately, the cable manufacturer had quietly changed the ribbon cable design. Same part number, but the plug housings were installed differently. In fact, there were several different kinds of housings, oriented in several different ways. In short, there was no way on earth that these cables were going to work.
Things got a bit easier from that point on – we were able to help people get their synths up and running by modifying one end of the cable. We updated the assembly instructions, emailed about a hundred MeeBlippers to alert them and offered to send out replacements to anyone who needed them.
Still, it just goes to show how one little change can have a big impact when you’re making physical things.
Most of the time, we catch mistakes before they get released. That’s why we make (and shamelessly abuse) production prototypes. But sometimes things go wrong when you least expect it. Ed.: This also means a significant difference in open source hardware than open source software. Because software can be easily patched, it makes sense to distribute source code before it’s complete. Not so with hardware: you tend not to want people making a permanent, physical object before you’re ready. That’s especially true if you want to support the result, but arguable even if not. -PK
Check out the cases in the photo above. The one at the bottom is famous – it was the prototype used in many of our early product shots. Its companions were manufactured in the first production batch, from the same design drawings. And the top panels are cut upside down. In fact, the entire first batch was done that way. While the manufacturer was quick to take responsibility, it took a few weeks to get replacements.
There have been other little goof-ups along the way, too. Check out these panel overlays. They’re supposed to be cut to the same width, but something went wrong mid-way through cutting the roll — the machine quickly slipped out of alignment and some ended up too wide to fit the top panel cutout of the case.
As Peter asked the other day, “Is this sort of thing going to happen every time we engage the Infinite Improbability Drive?” Yes, probably. That’s part of what makes it so much fun. Besides, for every slip-up, there’s a happy accident that adds unique character to an electronic instrument.