deskthority

This terminal keyboard is a bit of a mystery to me. Only thing I was able to track down on the net was a terminal station connected to a Pr1me Mini Computer but not the actual terminal name.



It looked very nice from the top even in the auction, but one of the rear corners was completely sheared off and a big chunk of the rear case was broken off as well.






This was a major case screw location and it also needed a large gap filler for the big open space left. I wound up resculpting a corner from high strength epoxy putty and high strength epoxy on the busted seam. Works perfect now. The case is now sealed on the outer edge and still opens normally. Obvious repair but not ugly IMO.

I didn't do as much rigorous cleaning on the bottom of the the case for fear of damaging the original labels which were still in very good condition.








The internals were in worse shape than my 1982 UTS 40. The plate was nice but somebody had spilled something several times into the right side of the keyboard. Can't stop magnets that way though. So I did the usual teardown. Removed the covers, removed the sliders, scrubbed with alcohol by hand and then lubed and reassembled. The switch plastics were more brittle than my other ITW board though, and in the areas with the spills and jammed plastic, a few brittle sliders snapped one of their legs in removal. Fun part is, they only need one to work and the upper cover keeps them from popping out when put back together for extra security. Can't even tell when it's all back together.





It used a RJ12 type jack but with wider plastic, so an RJ45 is needed for a convertor. Testing a home flashed Soarer's with a spare teensy let me see the R05 errors when it connected and test all the switches for functionality since they always sent a start and stop code for each press. And they all worked great. The caps lock is a latching switch compared to the LED switch on my UTS 40 and uses a strange y shaped groove in the side of the slider that a hoop of wire sits in. Works nicely though.

And all back together, smooth, lubed and functional. Just wish there was an external convertor or even a different controller that would do the job of making it useable on a modern PC.

Oof. Very nice, well done!

Looks great. Is there a similar problem with connecting these to a modern PC like with hall effects?

Unlike some Hall Effects, these have integrated controllers. So they can be converted externally like a pingmaster or something. Main problem is you need somebody to actually reverse engineer these protocols because there's no standard option to be able to brute force it with a new controller like you can do with conductive or capacitive switches. I can do hardware, but I can't do code worth jack, so I'm praying Haata or DMA or somebody comes through at some point in the future.

That's why Thomas was super lucky to get a version of ITW magvalves with AT protocol.

Magnetic valve switches should be "easy" to poll in hardware. I'd fire up a function generator and see what the switches do in response to a 5V sine wave running at something slow, 100 Hz or so. My bet is that the square wave is echoed out the other side of the switch (perhaps with some attenuation, delay, and blocking of high frequency components). Surely you could take advantage of this just as the original switch designers did.

A square wave is indeed echoed, but at a significant attenuation, so an amplifier or comparator with an appropriate reference would work to detect key presses. IIRC, some switches conduct the pulse in the up position and block it when pressed. You may be able to use an Xwhatsit.

I'll get to mapping out the matrix and give it a shot. I have a Cypress kit laying around for a DMA capsense project. Probably gonna use the Sperry as my guinea pig though as I'll be less angry if I accidentally break something.

This could get exciting.

After looking at the Sperry PCB, I don't know if a regular capsense style controller replacement will work with the existing board. Like you have rows but all rows are interconnected by jumpers to basically be one row. They're all connected at two pins to outputs on a driver chip with four inputs from other parts of the board. Just a real odd duck. You can tell the original setup was current sense. Might have to do a handwire but I am super leery of busting the ferrite rings on a switch or ten during the desoldering process. Maybe just desolder one and do a proof of concept to test things out.

Anything obvious I'm missing? I never cross out the possibility I'm being really stupid about something.