Attachment | Size |
---|---|
![]() | 24.01 KB |
![]() | 135.14 KB |
Hi everyone,
While discussing with someone I know about the Apple II+ and a pair of machines I revived earlier last fall, they mentioned that their own II+'s keyboard had "a few" defective keys and that they hadn't been using it much since. I offered to take a look and retrieved the ailing machine a few weeks later.
This substantially upgraded A2S1016 from the 36th week of 1980 sports a Datanetics 01-0551-01 keyboard, Apple part number 605-4117 dated from August 20, 1980. I took care of its Rifa filter cap, two electrolytics and a few cracked resistors in the power supply as well.
Cleaning flux was a bit of an afterthough in those days... (And for a few more years still, I've seen AEK IIs caked in the stuff.)
After running the keyboard test in Apple-Cillin, it turns out that the definition of "a few" here is twenty-one defective switches out of fifty-two, oof.
My multimeter agrees, no continuity even after exercising them a bit. A quick search tells me the contacts are supposed to be sealed so spraying contact cleaner directly in the housing won't do much. Oh, and the shell is "ultrasonically welded together" per https://telcontar.net/KBK/Datanetics/DC-50.
Cracking one of these open isn't without risk and one or two switches didn't survive:
After four decades the contact membranes are no longer sealed and there is some kind of contamination that prevents the contacts from closing:
It is possible to separate the membranes further to spray contact cleaner to "fix" the switch, gingerly reassemble all the parts then close the shell halves with a bit of polyimide tape - but I'd consider those on borrowed time.
To sum things up:
- Replacement Apple II keyboards of any kind are getting harder to come by and when they do show up they're not cheap.
- The Cherry MX based kit from ReactiveMicro is out of stock with no ETA.
- Datanetics switches are failing fast even though this machine is normally located in a climate-controlled office and there are no signs of spills.
- There are no replacement Datanetics switches available anywhere I checked.
- The Datanetics switches, while not outright terrible, are nothing to write home about in terms of key feel.
But I had an idea:
The Cherry MX switch footprint fits completely inside of the Datanetics' one! (That's a cheap Greetech clone I had lying around.)
Granted, Cherry's design is a decade newer (1983 vs 1973) but I'd still call its linear variant "period correct"; it has proven to be reliable, is easy to clean if needed and it's still in production.
So after getting my first 3D printer, many iterations on TinkerCAD...
...and print attempts in both PLA and PETG...
I'm happy with the results!
Whole keyboard with the "switch adapters" - I went with genuine Cherry MX Black:
With caps back on:
I have attached my model files here and also on Thingiverse (https://www.thingiverse.com/thing:5780203) - they're licensed under Creative C0mmons CC-BY-NC-SA: https://creativecommons.org/licenses/by-nc-sa/4.0/.
You'll need a reasonably well adjusted printer, I used PETG for my final prints and I suggest you do the same, PLA's weaker inter-layer bonds is not ideal for the DC stem in particular. The switch adapters need supports, the keycap ones don't.
Let me know if you end up using my models or if you have any questions.
Regards,
LR514
This is an absolutely wonderful contribution. I have this exact keyboard and i have resurrected it 2 years ago but random keys are beginning to get twitchy again and one has stopped working outright. When time permits, I am definitely going to give this a go.
Its a shame injection moulding costs so much to set up. These designs would be great to be able to buy.
But this is very cool. Time to dust off the 3d printer.
Great photos and back story.
Thanks heaps
Cheers
Dave
Hello,
nice job and thanks for sharing the project.
I was looking for a similar solution, I have an Apple J plus with a datanetics keyboard that has the same switches and I confirm that it is very difficult to revive a faulty switch.
How did you solve the misalignment of the cherry mx bottom pins with respect to the dcs51 pins and the pcb holes?
Is the new switch stem smooth as the original? Is it rough or does it slide nicely? What about keycaps wobble?
Sorry for all this questions
Best regards
Luca
This is fantastic... I wonder if it would work with the Datanetics clone PCBs that there are gerbers available for?
No need to apologize, these are good questions; I forgot to add a few photos to my original post and the forum isn't letting me edit at the moment, so keep reading...
I solder two (approx 2 cm) wires to the Cherry switch before inserting them in guide-holes on the switch adapter that match the Datanetics footprint. I used enameled "magnet" wire but any small gauge solid wire should work:
Enameled wires 1.jpg
Enameled wires 2.jpg
I'm keeping the stock Cherry stem to avoid that issue altogether, simply printing an adapter that mounts like an MX cap and converts to a DC-51 stem. Besides, printing smooth, quality stems with a filament printer is not really possible, perhaps with the exception of acetone vapor polished ABS but that comes with several other challenges. If you haven't tried them yet, I must say the new retooled "Hyperglide" MX Blacks I got for this project are pretty decent.
I checked and it is a little higher than the original DC-51 switches by virtue of the longer arm the cap is now on, but nothing noteworthy while typing normally. Comparable to, if not lower than that of Alps SKCC keyboards that shipped with similar adapters.
A few more notes:
The MX stem hole in the keycap adapter is slightly oversized: 1.25 mm thick cross whereas official Cherry spec says 1.17±0.02 mm. (Source: https://www.cherrymx.de/en/dev.html).
Thanks and Attribution: I used this Cherry MX switch model by user "gcb" : https://www.thingiverse.com/thing:421524 for part of the switch adapter design.
Thanks for all the answers!
Luca
nicely done. I may borrow some of this to do a similar thing to an SKCC keyboard