Crystal failures are always a problem. We used to chuck a bag of them in a tumbler for an hour then on a tray on a vibration table for an hour. About 20% would fail on test from being entirely open circuit to being way off frequency.
That's physically almost impossible. Even if the crystal sheet were to drop off the mounts it would have to do something electrically hard to explain to cause a short. The contact areas are usually the only conducting parts on the crystal itself. Failing open is easy for the exact same reason: if the crystal is mechanically damaged that will usually dislodge one or both of the contact areas. This is because those are the only places where something touches the crystal to give it as much freedom to vibrate as possible, any kind of further constraint would impede function.
I used to find that the crystal filters used in certain VHF radios would fail after as few as a couple of hundred hours if they were mounted a certain way in certain machines. I always assumed that it was down to some particular vibration frequency from the engine coupling through the cab.
Although that was nothing to the amount of 0.1μF SMD ceramic capacitors that would go leaky in the same product, and cause all sorts of bizarre faults like drastic loss of sensitivity (decoupling capacitor in IF amp bias circuit pulling down the bias rail) or sensitivity being okay but radio not going into service because the squelch would never open on the control channel (decoupling capacitor in the squelch threshold circuit pulling down a different bias rail).
Bloody things. Fine when they worked, not too horrible to work on when they didn't, at least.
He started selling Floppy Emus in 2014[1], bought a run of 10,000 DB-19 connectors in 2016[2], mentioned that he sold "several thousand" in 2021[1], and decided he no longer needed to horde as many DB-19s as he had held onto in 2025[3]. So based on all if that, probably somewhere between 4,000-8,000? (One of which is currently on my desk for connecting to an Apple IIe I recently acquired.)
> Every time I ask a contract manufacturer to assemble a batch of more Floppy Emus, there are always a few that don’t pass QA testing.
It seems like this would still be 60 out of a number less than the total production over all time. I was just curious what the failure rate was. We recently had a build with one failure out of 500. Which I'm happy with but isn't that large a sample size.
This is a huge problem in manufacturing. Unless you're on the latest "node" your options are bleak: Decades-old stock that's been rotting in a warehouse, parts pulled from boards and resold as new, cheap knockoffs. The crystal he's using are huge and outdated by modern electronics standards, and there really isn't anything he can do about that.
Crystal failures are always a problem. We used to chuck a bag of them in a tumbler for an hour then on a tray on a vibration table for an hour. About 20% would fail on test from being entirely open circuit to being way off frequency.
Did they ever fail short?
Nope.
That's physically almost impossible. Even if the crystal sheet were to drop off the mounts it would have to do something electrically hard to explain to cause a short. The contact areas are usually the only conducting parts on the crystal itself. Failing open is easy for the exact same reason: if the crystal is mechanically damaged that will usually dislodge one or both of the contact areas. This is because those are the only places where something touches the crystal to give it as much freedom to vibrate as possible, any kind of further constraint would impede function.
I used to find that the crystal filters used in certain VHF radios would fail after as few as a couple of hundred hours if they were mounted a certain way in certain machines. I always assumed that it was down to some particular vibration frequency from the engine coupling through the cab.
Although that was nothing to the amount of 0.1μF SMD ceramic capacitors that would go leaky in the same product, and cause all sorts of bizarre faults like drastic loss of sensitivity (decoupling capacitor in IF amp bias circuit pulling down the bias rail) or sensitivity being okay but radio not going into service because the squelch would never open on the control channel (decoupling capacitor in the squelch threshold circuit pulling down a different bias rail).
Bloody things. Fine when they worked, not too horrible to work on when they didn't, at least.
Did they say how many they built total? Was this 60ish boards out of 100, 600, 6000?
He started selling Floppy Emus in 2014[1], bought a run of 10,000 DB-19 connectors in 2016[2], mentioned that he sold "several thousand" in 2021[1], and decided he no longer needed to horde as many DB-19s as he had held onto in 2025[3]. So based on all if that, probably somewhere between 4,000-8,000? (One of which is currently on my desk for connecting to an Apple IIe I recently acquired.)
[1]: https://www.bigmessowires.com/2022/02/16/bmow-business-updat...
[2]: https://www.bigmessowires.com/2016/06/04/db-19-resurrecting-...
[3]: https://www.bigmessowires.com/2025/06/30/bulk-lots-of-db-19s...
> Every time I ask a contract manufacturer to assemble a batch of more Floppy Emus, there are always a few that don’t pass QA testing.
It seems like this would still be 60 out of a number less than the total production over all time. I was just curious what the failure rate was. We recently had a build with one failure out of 500. Which I'm happy with but isn't that large a sample size.
This is a huge problem in manufacturing. Unless you're on the latest "node" your options are bleak: Decades-old stock that's been rotting in a warehouse, parts pulled from boards and resold as new, cheap knockoffs. The crystal he's using are huge and outdated by modern electronics standards, and there really isn't anything he can do about that.