Some LK201 clones with Cherry MX switches


18 Oct 2021, 07:54

The wiki has a page on Clones of the DEC LK201, and a number of them were made by Cherry with MX black switches. After working on some related keyboards, I dug the ones covered here out to focus on. They aren't Cherry boards, but do have Cherry switches.

First up is a Wyse WY85, 840105-01, with thick double-shot keycaps, from around 1987.
WY85-case.jpg (1.3 MiB) Viewed 316 times
WY85-back.jpg (915 KiB) Viewed 316 times
It is hard to tell from the photo, but the legend colors are navy and yellow.

The switches are in need of a good clean, which probably means desoldering. But it's still operational, if a bit scratchy.

This is the version without an actual controller. The strobe pulses just clock a 393 TTL counter.
WY85-top.jpg (1.44 MiB) Viewed 316 times
WY85-bottom.jpg (1.28 MiB) Viewed 316 times
This is the same model that Soarer used for their WYSEverter and it works fine with that, provided a pull-up resistor is added.
Wyse-Converter-bottom.jpg (275.72 KiB) Viewed 316 times
The protocol is so simple, though, that it might be worth reimplementing for TMK/QMK at some point; the same converter hardware should work.


18 Oct 2021, 07:59

Next is a Qume QVT-61 from around 1994. This is what got me started on this set, after doing a few older QVT keyboards that cloned a VT100, sort of.
QVT-61-case.jpg (1.47 MiB) Viewed 307 times
QVT-61-back.jpg (1.1 MiB) Viewed 307 times
According to a table on Bitsavers, this part number, 303686-11, was also used with a number of other models, even including the QX15 X terminal, back when those were a thing. If so, then it must not include the key layout: compare the keyboards for some of these models here.

This has thin double-shot caps. The switches are in pretty good shape and certainly much nicer to type on than the earlier membrane / rubber-dome QVTs.
QVT-61-top.jpg (1.28 MiB) Viewed 307 times
QVT-61-bottom.jpg (1.27 MiB) Viewed 307 times
It has a 4P4C RJ cross-over cable.

I read online somewhere that it was compatible with Wyse keyboards. That might make sense, since Qume was reselling Wyse by this time and was eventually acquired by them.

But, no, it is in fact more or less the same protocol as the QVT-101+, but with two start bits instead of one.
qvt-61-A.png (30.04 KiB) Viewed 307 times
qvt-61-nokey.png (30.71 KiB) Viewed 307 times
Compare the traces here. The scan codes aren't the same, since they evidently represent the physical matrix, but the shift encoding matches. To set the LEDs, the same serial encoding is used, with the terminal pulling the signal line low to start. The same converter hardware can be used, with slightly different firmware.


18 Oct 2021, 08:08

I do not have much information about the third one, which has Cangjie legends. The base keycaps are thin double shot and then these black sub- and blue front-legends are printed on.
KB-199-case.jpg (391.44 KiB) Viewed 299 times
KB-199-front-typewriter.jpg (312.92 KiB) Viewed 299 times
KB-199-front-numpad.jpg (254.29 KiB) Viewed 299 times
The case a small label on the edge that says KB-199. The PCB says BX-220.
BX-220-top-before-1.jpg (433.45 KiB) Viewed 299 times
BX-220-top-before-2.jpg (1.21 MiB) Viewed 299 times
BX-220-bottom-before.jpg (1.3 MiB) Viewed 299 times
There was a QC label.
BX-220-QC.jpg (302.64 KiB) Viewed 299 times
But it seems to just say 键盘生产工艺流程条 and so does not give any clues as to the manufacturer.

Some damage was done before it got here. The 8749 controller was pulled out of its socket. The cable was hacked off, so I do not know what connector it had, other than that it had five conductors. Maybe DIN?
BX-220-connector-remnant.jpg (160.43 KiB) Viewed 299 times
Some optional components, such as a buzzer, were never there. But the 72LS174 flip-flop used to buffer the LED states is still present, so we should use that. This chip also gives an approximate date of 1993.

It appears that there was some attempt to harvest the vintage MX blacks, because there is some sloppy half-done desoldering. This also seems to have broken some of the switches.

So, first step is properly removing all the switches.
BX-220-top-bare.jpg (1.32 MiB) Viewed 299 times
BX-220-bottom-bare.jpg (1.18 MiB) Viewed 299 times
As I feared, this reveals that some of the pads have been lifted.

I put in some clean MX blacks (plus the one gray for the spacebar), taking care to attach to what was left of the traces, and added the remaining two LEDs and their resistors.
BX-220-top-after.jpg (1.26 MiB) Viewed 299 times
I might one day try to program a NOS 8749 to do something simple like serial. In the meantime, I used one of the few Teensy++ 2.0 boards I have left here (they have been discontinued) to do the scanning and USB. It can fit into the same socket with just a little hacking to move ground and eliminate analog reference pins that are in the middle of port 1. Since this new controller just pulls right out, all the original circuits are preserved.
BX-220-controller.jpg (669 KiB) Viewed 299 times
There aren't any switch diodes. But there are a few diodes to prevent the key switches from interfering with LED state loading. This forces a particular matrix scan direction. Which in turn means that the anti-ghosting disables Ctrl, (Left-)Shift, and Meta combinations. So not really up to being a modern keyboard if one needs those, like in Emacs.

User avatar

19 Oct 2021, 08:23

Very nice project!

A closely related board with MX Blacks:
Pinout & Soarer's logic analyzer - Tatung Terminal


Post Reply

Return to “Workshop”