The "I don't know" keyboard w/ "beats me" switches

User avatar
snacksthecat
✶✶✶✶

16 May 2019, 03:46

Okay, stupid title. But what is this keyboard and what are the switches?

I only have some photos for now but I'll be taking a shot at converting it.

Gotta go grocery shopping now but I'll also be posting a disassembly video soon.

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

Image

User avatar
Sangdrax

16 May 2019, 04:17

Well, I'll be. Some sort of Fujitsu leaf spring membrane board. Wanna call it Gen 4? I can't imagine what else it could be.

User avatar
snacksthecat
✶✶✶✶

16 May 2019, 20:43

Double "duh" on my part...
  1. xxhellfirexx pointed out that I actually own(ed) another keyboard with the same switch
  2. The company that made this keyboard is right in the title of the ebay listing... MAI
I did a quick google and it sounds like MAI made lots of terminals for banks and small businesses. This unit belonged to a VDT called the Basic 4.

I peeked into the service manual for this machine but it doesn't go in depth about the keyboard. That was the only doc for this machine on bitsavers but I did see some other material out on the web. Just gotta find the time to dig in.

Oh I also had an opportunity to plug it into my logic analyzer last night. Didn't produce any output but I didn't really expect it to yet.

Thing that is currently on my mind; is NKRO possible?

Findecanor

16 May 2019, 21:16

snacksthecat wrote:
16 May 2019, 20:43
Thing that is currently on my mind; is NKRO possible?
Not with that membrane. But the membrane is in two parts (rows and columns) and a middle spacer-membrane.
You could perhaps construct a custom circuit board (or membrane) that replaces the bottom membrane, on which each pad leads to a diode for that switch.
If not a custom printed board then maybe traces made with some kind of metal tape.

User avatar
snacksthecat
✶✶✶✶

17 May 2019, 01:09

Findecanor wrote:
16 May 2019, 21:16
snacksthecat wrote:
16 May 2019, 20:43
Thing that is currently on my mind; is NKRO possible?
Not with that membrane. But the membrane is in two parts (rows and columns) and a middle spacer-membrane.
You could perhaps construct a custom circuit board (or membrane) that replaces the bottom membrane, on which each pad leads to a diode for that switch.
If not a custom printed board then maybe traces made with some kind of metal tape.
I suspected as much, based on a little bit of research I did on the train looking for similar mods to Model Ms. I figured the M is such a common board that if no one's done it on that then it's not been done at all.

Okay so here's the disassembly video. It's a little dumb to post now that a lot of the mystery has been lifted.

It's not a very interesting video and I also say the word "um" about 1,000 times but oh well. Plz be nice :ugeek:
I do have some theories about the protocol based on some other boards I've seen. I'll post the info once I figure out how to make my messy thoughts more coherent.

User avatar
snacksthecat
✶✶✶✶

17 May 2019, 02:14

Oh no, I'm getting that horrible feeling I get when I start to realize I did something dumb.

I'll give someone a little prize if they can guess what it was.

Hint: what I did is mentioned somewhere in this thread

User avatar
User101

17 May 2019, 02:49

snacksthecat wrote:
17 May 2019, 02:14
Oh no, I'm getting that horrible feeling I get when I start to realize I did something dumb.

I'll give someone a little prize if they can guess what it was.

Hint: what I did is mentioned somewhere in this thread
Did you blow it up by plugging it into your logic analyzer? ;)

User avatar
PlacaFromHell

17 May 2019, 03:10

I guess you put the middle membrane that separates the rows and the columns under the other two membranes, that causes all the keys to be electrically pressed. Membranes are not double-sided so you don't need to isolate them from the backplate.

User avatar
snacksthecat
✶✶✶✶

17 May 2019, 17:12

PlacaFromHell wrote:
17 May 2019, 03:10
I guess you put the middle membrane that separates the rows and the columns under the other two membranes, that causes all the keys to be electrically pressed. Membranes are not double-sided so you don't need to isolate them from the backplate.
Ding ding ding!

I don't know what you won yet but I'll see what I've got and reach out through PM.

User avatar
PlacaFromHell

17 May 2019, 19:42

YESSSS :lol:

User avatar
snacksthecat
✶✶✶✶

17 May 2019, 22:58

PlacaFromHell wrote:
17 May 2019, 19:42
YESSSS :lol:
You might even say, "Hell yes!"

So I've done a few things since the last post.

#1 - I made sure the keyboard functionally works.
Like when you press a key there's continuity along the path. Should really have done this earlier.

#2 - I made sure the crystal is working
Not that I thought it was broken, but I just wanted to get some output from the board. If I tested it and got nothing, I'd know something was wrong.

#3 - Started tracing the lines looking for clues
I looked up the datasheets for the ICs online to see if they could be of any help. That's what gave me a hint on a different board. Here's what I found:

Image

So the big boy chip does all the heavy lifting, figuring out which keys are pressed. Then it ouputs that data to an inverter chip. One of the lines on the keyboard cable connects to this inverter. The other data (?) line is connected to a transistor (Q1).

Brb looking up what transistors do.

sncbrax

18 Jul 2019, 07:03

Hey snacks, thanks for sharing! I recognized it right away as I had found a MAI board that looks just like it structurally some time ago. Yours looks pristine though, mine was pretty busted and dusty when it arrived, not to mention nearly brown it was so yellowed.. so I set it aside until stumbling on your post earlier. I cracked it open and by looking at the SN and full PCB it seems like it might be a bit of an older model? It has quiet click Type 3 Fujitsu Leaf Spring switches. The tactile event and click are almost non-existent :lol: but the travel is still pretty smooth despite the condition. The dyesubs look nice. Here's the album with the pictures of it that are nowhere near as pretty as yours :) https://imgur.com/a/5ooZ06n but I think I got all the necessary codes captured. I actually managed to forget to take pictures of the PCB while I had it out but I remember the code that corresponds to the one pictured below- it starts with the same first four characters: N86D-2867-R010/02 HANDA. Funny that the code on either of our PCBs reads N86D but then the sticker on either plate says N860. Is that like a Chinese Whispers thing? :P

Image

Post Reply

Return to “Workshop”