Some Model M refurbishments

[cpl]

I got three sold-as-seen Model M's back in March, as I reported here: http://deskthority.net/keyboards-f2/mcd ... it=douglas.

First, I decided to clean up the working one, missing its external birthday label. It turned out to be a Jan 1988 model, from the internal birthday sticker. It has the yellow wired-up LED PCB, but not the brass-coloured plate. Following the wiki here, that would make it a "1.5" genration model since it fits neither the first or second gen. descriptions.

Anyway, I started with this:

Dirty

DSC01272s.JPG (318.72 KiB) Viewed 5648 times

And achieved this:

Clean

DSC01275s.JPG (315.48 KiB) Viewed 5648 times

Internals, for those interested:

Inside Scottish 1988 Model M

DSC01271s.JPG (368.28 KiB) Viewed 5648 times

Next I opened up the dead captive cable M, born in Scotland in 1994. The PCB header was detached from the PCB, so I soldered it back on. This brought the keyboard back to life. However, the following keys are not registered: 'Pause' 'Num Lock' '/' '*' 'Enter' (the big one) 'R Shift' ' R Ctrl' 'Z' 'X' 'C' 'V' 'M' ',' '.'

I wonder if this combination points to a particular failure mode, such as a broken track on the membrane? I suggest this due to this previous work: http://danson.grafidog.com/2009/02/model-m-repair.html

I also have a SDL cable 1993 blue logo model that I haven't opened yet. Apart from missing a spring at 'L ctrl', the following keys don't register: 'B' ' N' '?' 'Alt Gr' the spacebar and the left, right and down arrows.

Any hints or tips would be most welcome

[off]

the following keys don't register: 'B' ' N' '?' 'Alt Gr' the spacebar and the left, right and down arrows.

Sounds like a single trace to me.. No clue on IBMs whatsoever, but these keys are so in line with eachother that a single trace seems very likely- so it could be cut somewhere, or just not making the jump from the membrane to the pcb (dirty contact). Check some matrices and crack that baby open is what I'd say
GL, nice find (it seems)!

[Icarium]

How did you clean the casing? I put one in the dishwasher once and even at only about 55 degrees it got badly warped. :/

[cpl]

the following keys don't register: 'B' ' N' '?' 'Alt Gr' the spacebar and the left, right and down arrows.

Sounds like a single trace to me.. No clue on IBMs whatsoever, but these keys are so in line with eachother that a single trace seems very likely- so it could be cut somewhere, or just not making the jump from the membrane to the pcb (dirty contact). Check some matrices and crack that baby open is what I'd say
GL, nice find (it seems)!

Thanks for the tips. I'll start with the connection from the membrane to the PCB. If that doesn't fix it, it looks like I'm in for a 'bolt mod' type activity to get into the bowels

[cpl]

How did you clean the casing? I put one in the dishwasher once and even at only about 55 degrees it got badly warped. :/

I used some multi-purpose kitchen cleaner (Cilit Bang) diluted in a bucket of hot water, and scrubbed it with a sponge. Some low-abrasive Scotch Brite rubbing was need for the most stubborn dirt, rather than the sponge.

The keycaps were cleaned using denture tabs - no rubbing needed.

[off]

ps, should be so much easier if you have something like a multimeter (continuitymeter) available to you, even without dismantling the board any further, finding someone's traced matrix will be semi-necessary though

[cpl]

ps, should be so much easier if you have something like a multimeter (continuitymeter) available to you, even without dismantling the board any further, finding someone's traced matrix will be semi-necessary though

I do indeed have a digital multimeter. I'm hunting down the trace matrix information now, mostly by trawling threads here and at geekhack.

[off]

Not sure how far you've gotten; but just found this:
KEY_B KEY_N KEY_spc MOD_RALT KEY_larr KEY_KPminus KEY_darr KEY_rarr KEY_slash
source
Which seems a match!
So I'm assuming it should be the last pin; check it by unplugging the sheets and holding your multimeter in continuity mode on that pin on the controller and tapping the pins on the controller that belong to the other sheet; ofcourse while plugged in and having notepad/aquakeytest active.

*edit: if that pin checks out alright you might be lucky and have a bad contact between the membrane and pcb, like slightly misalligned. (as was just demonstrated to me after I'd cap-repaired and put back together my just-in-M2, a few slaps lodged most contactpoints into the right spots xD sadly I still might have to rip her apart again since I've got five bent springs *and* one membrane contactpoint doesn't have contact without external pressure- I'm betting that's the foam underneath that I've pushed down a bit when trying to remove the controller).
If unlucky it's the membrane itself that has broken traces, continuity-checking the sheet will take longer though
Also an M2 is a different beast, so don't go slappin' up your M.

[nfc]

Matrix is collected on connectors J3 and J4, see attached picture.

Just have a look at Model M matrix at (bottom of the page):
https://github.com/clee/rump/blob/master/doc.txt
Chris Lee did a fantastic work here... also with documentation.

It seems that your 1994 keyboard have a problem on J3.7 meanwhile the 1993 one have a problem on J3.8
Try to search in that direction


nfc

[cpl]

Thanks, off and nfc. This is very valuable advice! I haven't made any progress yet, but hope to find some time this weekend.

[cpl]

It seems that your 1994 keyboard have a problem on J3.7 meanwhile the 1993 one have a problem on J3.8
Try to search in that direction


nfc

I had a quick look this evening, and a visual inspection of the 1994 'board tells me that the exposed conductor on the track that connects with J3.7 is worn through. I've ordered some copper tape to sort this, and I'm hopefull that might fix it!

Alas, nothing so obvious with the 1993 'board. I will have to get the multimeter out when I have more time and do an electrical check.

[off]

I also have a SDL cable 1993 blue logo model:
the following keys don't register: 'B' ' N' '?' 'Alt Gr' the spacebar and the left, right and down arrows.

Well grats on fixing 1994!
And for '93, check if 'numpad -' works... if it does, there's a trace coming from the controller through there and on to the other ones that don't register; just worn through between NUM- and the arrows in that case

[cpl]

Well grats on fixing 1994!

Thanks! And thanks for the advice, you guys saved me a lot of time.

To finish that one off, here is the fully working and cleaned up 'board

Clean and fully functional 1994 Model M

DSC01301.JPG (299.11 KiB) Viewed 5176 times

Underside

DSC01302.JPG (293.52 KiB) Viewed 5176 times

[cpl]

And for '93, check if 'numpad -' works... if it does, there's a trace coming from the controller through there and on to the other ones that don't register; just worn through between NUM- and the arrows in that case

For the 1993 'board, everything is working except those keys I listed. I guess a track is broken in the membrane, as you suggest. Given the missing parts and the lack of its own SDL cable. I think I'll keep it for parts rather than bolt-mod it and seek the broken track.

1993 Model M - Faulty- Missing a spring, 2 stems, many keycaps

DSC01303.JPG (285.56 KiB) Viewed 5172 times

[nfc]

Congratulations also from my side!
Glad if somebody else can take advantage from mine and Off's tips or fault tracing method for the Model M.
Thank you for having reported reparation results. It was kind from your side.

Ops... we posted simultaneosly

[off]

And for '93, check if 'numpad -' works... if it does, there's a trace coming from the controller through there and on to the other ones that don't register; just worn through between NUM- and the arrows in that case

For the 1993 'board, everything is working except those keys I listed. I guess a track is broken in the membrane, as you suggest. Given the missing parts and the lack of its own SDL cable. I think I'll keep it for parts rather than bolt-mod it and seek the broken track.

DSC01303.JPG

Both looking good! Still re'93, that seems like a shame.. assuming the matrix matches yours (and it certainly appears so) it's just one track that's damaged in one spot; and that should be very easy to see once you pop her rivets.
Two (three) options:
A-repair trace on the ultra-cheap by pencil (just draw the line that was there in pencil; that's all).
B-repair trace with coppertape. challenge would be to find that, and semi-cheap.
(C-keep for parts.) I know this is deskthority not geekhack, yet such a simple repair seems mandatorily easy, like springswaps.

Hm, on second thought; you are missing lots of stuff... perhaps check if ascaii or someone has what you're missing floating about, if you have someone around who'd use her.

[nfc]

Just a stupid question... Did you try to swap the key-frames between keyboards? The fault should move together with the faulty component.

Never give up... I think Unicomp and/or ClickyKeyboard still have cables, keys, keycaps and membranes.

[ripster]

http://www.amazon.com/CircuitWriterTM-P ... B0002BBVQO

This stuff sucks
http://www.amazon.com/Electrically-Cond ... d_sbs_op_1