CTM-70 keyboard for late 70's 16-bit German business computer


10 Sep 2018, 12:15

Hi, I'm long time lurker and keyboard-geek, finally signed up-

Today I was handed the keyboard to the German CTM-70 business computer, by a friend who had seen my esoteric keyboard shelf and figured "10 is rookie numbers". This keyboard is now easily the oldest one in my possession, and also the first one of mine that doesn't 'Plug and relatively Play'.
IMG_20180910_1111422.jpg (2.91 MiB) Viewed 1546 times
keys for scale, pun intended

Note it isn't in QWERTZ layout, though the shape of both the Y and the Z keycaps seems to mismatch with the rest of their rows.

Obviously, this thing doesn't talk modern computer, and I'm actually a little bitter it didn't come with the entire computer. It connects using a male DB-15 (gameport-style) and it's brimming with 74xx TTL circuits, so I expect getting keystroke info out of this board shouldn't be too hard.

I intend to use this thread for documenting the progress, but I am very receptive to recommendations, help and pointers to out-of-the-box-solutions so I don't actually have to use my brain.

I found some info on http://www.horniger.de/computer/ctm/ctm70_4.html - Looking at the input card of the computer, the traces around the tastatur-connector suggest to me there's some 'parallel stuff' going on. (Note that is a DB-25 connector instead)
ctm70_4_17.jpg (481.17 KiB) Viewed 1546 times
The innards are labeled Cherry G80-0229 (pictures later), keystroke is a little heavy, linear and barely clicky. Keys are a little wobbly and catch occasionally. Alphanumeric part of the layout is offset to the left by about half a keywidth compared to IBM AT layout. Keycaps are concave, and are laid out about 6 cm above desk level at the highest point. Pictures suggest this keyboard was meant to be used in a recessed portion of the computer desk. The near-instant RSI my wrists develop when looking at this beast, support this.

I'll only be able to really investigate after office hours but I yearn to know: What sort of mechanism can I expect in here?

User avatar

10 Sep 2018, 15:56

A very vintage board in nice shape.

With ancient boards, all bets are off on converting them to usb. If it’s Cherry, it’s very likely to be regular “ohmic” switches at least, instead of something esoteric like capsense or Hall effect. You’ll want to take pictures of the PCB and switches, then likely get to work on mapping the matrix. Oh, and swap those Y and Z caps back to where they should be!


10 Sep 2018, 18:53

Sorry for the crud and the large pictures; last time I spent any sort of time on a forum I'd have had to put 56k warning in the thread title but I think we're past that as a society.

I had a quick oogle at the logic section and it's relationship with the connector
IMG_20180910_175157.jpg (249.77 KiB) Viewed 1449 times
Note the two 74LS157N multiplexer IC's. Their outputs go directly to the DB15 connector, and their select lines are tied together to one of the DB15 pins too. I'm taking this home and continuing the research there, but the setup suggests this is a single-keystroke deal where the offboard controller takes the lead in deciding which of the regions gets polled. I hope I'm wrong on this one, but I have to be realistic in my expectations.

It seems as though the shift-key is wired separately. The blind key above it is wired in parallel so as to provide crude but reliable case lock functionality. Nice. Cruise control for Cool has never been this reliable.
Muirium wrote: With ancient boards, all bets are off on converting them to usb.
Challenge accepted!!
Muirium wrote: If it’s Cherry, it’s very likely to be regular “ohmic” switches at least, instead of something esoteric like capsense or Hall effect. You’ll want to take pictures of the PCB and switches,
I can't find a lot on google about "Ohmic", how can I tell them apart, how do they work, what sort of maintenance do they require after 40 years?

Very depressed about these pull-down resistors
IMG_20180910_175208.jpg (268.93 KiB) Viewed 1449 times
Switches, exterior - I'm getting M7-vibes? wiki/Cherry_M7
IMG_20180910_175043.jpg (211.12 KiB) Viewed 1449 times
Partial frontal nudity
IMG_20180910_175020.jpg (125.5 KiB) Viewed 1449 times
Label says
Artikelnr G80-0229
Serienr 1072

PCB is marked "CTM 70 Tastatur T/a Bauteilseite 6.3.80" which dates this unit as a very late production; possibly a spare or replacement?

IMG_20180910_175322.jpg (173.21 KiB) Viewed 1449 times
Muirium wrote: then likely get to work on mapping the matrix. Oh, and swap those Y and Z caps back to where they should be!
Each pair is connected by wire or trace on the other side. I'm curious why they're doing that.
IMG_20180910_175327.jpg (182.98 KiB) Viewed 1449 times
Empty bottom shell
IMG_20180910_175004.jpg (181.1 KiB) Viewed 1449 times


10 Sep 2018, 19:19

Those look like wiki/Cherry_M7.

I think ohmic refers to switches that create a conductive connection upon activation.


10 Sep 2018, 20:21

I shot a quick video of the typing sound without casing for the curious of heart

CW: Vomit-focus, shaking hands, cruddy keyboard

User avatar

10 Sep 2018, 23:59

Yeah, sorry about “ohmic” — a name only IBM could love — it’s just lodged in my recall thanks to Xwhatsit digging it up from the old docs. It simply means “conductive switches” as we are used to them. Versus things like capacitive sensing, which aren’t as simple to understand or track down with a multimeter.

Now, your board looks to use good old Cherry M7 switches. They are the vintage ancestor to Cherry MX. A good bit bigger but same idea. You should be able to work out which pins in each of those clusters respond to the switches by listening to the beeper on your multimeter while pressing and releasing the matching key. Beware: the pattern might not be the same orientation for every key, as sometimes they’d rotate the switches. The logos on top should always give away their alignment.

So, what are all those extra pins for? Diodes, most likely. Have you read up on the concept of the keyboard matrix? That’s what you’re seeing there. A 2D grid of rows and columns — wavy and often downright inscrutable as they may be — which is how your keyboard actually sees its switches. You want to work out that logical grid. Diodes are a good sign because they are for [wiki]NKRO[/wiki] which means your keyboard is ultimately able to output every possible combination of keys simultaneously, even if its vintage controller cannot.

Oh, and you probably want to familiarise with the art of Teensy keyboard controllers. Even if you’d prefer to retain your original keyboard’s innards, it’s a good idea to know how they work. This thread has a bunch of pictures of what a matrix looks like during manual reconstruction, which can help. Your keyboard’s PCB is doing this same task, with all its tracks on both sides.

You’ll find the rows and columns are routed to a line or two of pins, for the controller to strobe and read. That’s your axis.

As for working out the current controller’s protocol, well, that’s beyond my knowledge. You know your way around a logic analyser? Apparently that’s how. Once you’ve worked out where to supply the power, and how much!

User avatar

11 Sep 2018, 07:19

These are plain old contact switches so like Murium said, easiest thing is probably to desolder the PCB and hand wire a matrix. Bonus, you get the switches loose and can clean them before putting everything back together. Should be about $10 for the wire, $15 for the Teensy 2.0 ++ controller, $2 for the diodes, but you can make it with every modern feature, layers and everything.

Worrying about terminal protocols is for stuff you can't just bruteforce.


03 Jul 2020, 16:40


this is an older thread, but is your CTM 70 keyboard now working?
I'm the owner of a complete CTM 70 machine and have also some manuals.
Probably this could help you - it's the schematics of the CTM keyboard:
https://www.dropbox.com/s/kui29ny5vpv4b ... r.pdf?dl=0

Btw: the second image is from the IO-board of my CTM ;)

Best regards,

Post Reply

Return to “Keyboards”