Model M with keys failing to register

[escargot]

Greetings all, I've been reading the forums for some time now but haven't felt the need to post so far, until I got my hands on an old Model M (1391506, Lexmark, 1993, Made in Mexico, spanish layout) that looked pretty much healthy - until I started using it periodically.

From one moment to another some keys start failing to register, specifically the numbers (non-kp) and some Fx's and I'm sure some of the Insert/Home/Pg Up row, but I haven't been able to double check.
Why? Not a clue! The issue pops up without any predictable pattern. At the moment it's working perfectly, no problem at all, but suddenly the keys will stop registering and so far the only thing that seems to address the problem is a subtle touch of the cable at the keyboard's end.
It obviously sounds like a cable-related problem, but moving the cable around in an attempt to reproduce the issue has been futile.

My first approach to fix it was a thorough cleaning of the keyboard, so I opened it up after finding a way to go past the 5.5mm screws without a nut driver (swept the whole town looking for one - I ended up using a normal screwdriver and a ~7mm steel tube that I managed to grip onto the screws) and cleaned it entirely. I also checked the rivets which were undamaged and the board is tightly attached to the backplate. The controller and everything on that department looked just fine, although there were some dark stains on the contact membranes, which I cleaned with Qtips moisten with alcohol, but they look pretty healthy.

So yeah, it looked pretty good inside, nothing to worry about, so I thought some of the stains in the membranes might have been the reason of the failing keys, but wrong I was. The problem is still happening with no apparent reason. I suppose it is the cable, but wouldn't a cable problem affect all the keys and not just a bunch of them?
The keys affected are all clustered together, 12345, 7890, some Fx's; so it's apparently physical, although the rivets are all healthy and everything looks just fine inside.

I wanted to do the bolt-mod anyways, but I'm living in Chile/Argentina and I couldn't find 2mm bolts anywhere, the smallest I found was 3.5mm and 3cm long with some huge nuts. I still want to visit some other shops in a near town hoping I find smaller bolts, but until then I'm stuck with a pretty random malfunction.

I've also been searching for more Model M keyboards, but they are pretty rare down here and I'm not really willing to pay huge shipping prices by getting one from the US.

Any ideas around?

[bettablue]

Greetings all, I've been reading the forums for some time now but haven't felt the need to post so far, until I got my hands on an old Model M (1391506, Lexmark, 1993, Made in Mexico, spanish layout) that looked pretty much healthy - until I started using it periodically.

From one moment to another some keys start failing to register, specifically the numbers (non-kp) and some Fx's and I'm sure some of the Insert/Home/Pg Up row, but I haven't been able to double check.
Why? Not a clue! The issue pops up without any predictable pattern. At the moment it's working perfectly, no problem at all, but suddenly the keys will stop registering and so far the only thing that seems to address the problem is a subtle touch of the cable at the keyboard's end.
It obviously sounds like a cable-related problem, but moving the cable around in an attempt to reproduce the issue has been futile.

My first approach to fix it was a thorough cleaning of the keyboard, so I opened it up after finding a way to go past the 5.5mm screws without a nut driver (swept the whole town looking for one - I ended up using a normal screwdriver and a ~7mm steel tube that I managed to grip onto the screws) and cleaned it entirely. I also checked the rivets which were undamaged and the board is tightly attached to the backplate. The controller and everything on that department looked just fine, although there were some dark stains on the contact membranes, which I cleaned with Qtips moisten with alcohol, but they look pretty healthy.

So yeah, it looked pretty good inside, nothing to worry about, so I thought some of the stains in the membranes might have been the reason of the failing keys, but wrong I was. The problem is still happening with no apparent reason. I suppose it is the cable, but wouldn't a cable problem affect all the keys and not just a bunch of them?
The keys affected are all clustered together, 12345, 7890, some Fx's; so it's apparently physical, although the rivets are all healthy and everything looks just fine inside.

I wanted to do the bolt-mod anyways, but I'm living in Chile/Argentina and I couldn't find 2mm bolts anywhere, the smallest I found was 3.5mm and 3cm long with some huge nuts. I still want to visit some other shops in a near town hoping I find smaller bolts, but until then I'm stuck with a pretty random malfunction.

I've also been searching for more Model M keyboards, but they are pretty rare down here and I'm not really willing to pay huge shipping prices by getting one from the US.

Any ideas around?

Escargot:  I have an online friend on FaceBook who lives in Argentina.  

Don't take it completely apart unless you have to, in other words, don't fix something that isn't broke.  The only reason to do a bolt mod is to either repair something underneath the keys or between layers that you can't get to otherwise.  Besides, the bolt mod is a lot of work to do only to find out that it didn't resolve the issue.  Although in the process, you probably would, simply because all of the connections to theh board would be reseated.  (More on that later)  

I would try to get a new cable and replace that first.  Contact www.clickykeyboards.com for pricing and shipping information.  The site lists cables starting at $9.00 US.  It can't cost that much to ship and in the long run, it could save you a lot of money buying another keyboard.  The problem could also be in the connector.  The cable connector is going to have more stress on it simply for being what and where it is.  

There is one other thing to try.  Open the keyboard and disconnect the controller board connections and re-seat them.  Look at Model M Rivet Replacement for Klutzes written by jpc in the modifications section of the forums.  You will find directions for disconnecting the controller in section 1.2.2.  

If all else fails, I do have a contact in Argentina who has a model M that is in pretty rough shape.   The keyboard is missing some keys, and the case looks like it came out of a war zone.  His keyboard isn't broken up or anything that serious, but it is very dirty.  He would probably be willing to give it to you if you can salvage some parts from it to fix yours.  Or you could take parts from yours and fix his; either way.

Hope this helps.  Of course, I will defer judgment to some of the rebuilders and modders here, like ripster.  In my opinion, ripster is probably the most knowledgeable when it comes to modding and repairing these boards.

[escargot]

I thought the same about the bolt mod, and even using the wrong sized bolts was nuts (no pun intended). Although the sturdiness of bolts and nuts instead of plastic rivets is really attractive.

In a very old store I saw some cashiers using antique IBM keyboards. I think I saw a classic Model M, but the rest were terminal designed versions. Some other cashiers were using some standard membrane Dell's, so I thought having a chat with the manager to try to get my hands on them by replacing them for cheap Dell/Genius/whatever-cheap membrane keyboards. But I had to leave town sooner than planned and missed the chance.
I'm eagerly looking for my next trip there to give it a try to my plan.

I also thought about contacting clickykeyboards.com and trying to get a cable replacement, but since I'm not 100% sure about it being the cable I want to avoid it.

When I opened it I re-connected the smaller membrane and I remember checking the wider one, and it sat well inside the plug.
Is there a detailed contact sheet from the full board contacts and it's connection to the membrane? That would help to know if the keys failing are actually in the same stream.

Thanks for the offering about that contact you have in Argentina, could be helpful if everything else fails.

[escargot]

So, a week has passed and the keyboard had not malfunctioned until today.
I cleaned up my desk and moved stuff around, keyboard included, after I was done I noticed the failure resurfaced.
I opened up Aquakeytest and checked what were definitely the keys not working:



The PrintScreen was also malfunctioning, but as with all the other keys failing they are not completely dead, after some strokes they will react giving one or even more inputs at once.

Anyways, I pushed the cable downwards towards the desk and now it works.

I'm not familiar with the electronics inside the keyboard, but I thought all the key inputs were sent through one of the wires of the cable - namely pin 1 in PS/2.
I have no clue what's wrong with the keyboard, probably something loose in the controller that by moving the cable -> socket -> circuit board gets fixed temporarily.

I hope the Aquakeytest screenshot helps to figure out what's wrong with my keyboard.

[theferenc]

Actually, it looks to me like printscreen is permanently depressed, resulting in blocking on other keys.

You did press each key individually and distinctly, correct? Because I get a similar picture by holding down printscreen and another key, then trying to press your missing keys.

Either that, or you used printscreen to take that picture, which is also possible.

[escargot]

Yeah, I took the picture with PrintScreen, I pressed it several times until it registered.

And yes, every key was pressed individually and there's no key giving continuous input.

Is there a detailed contact membrane sheet where I can see the relation between the membrane and which key goes through which conduit. That might help to investigate further.

[keyb_gr]

Is there a detailed contact membrane sheet where I can see the relation between the membrane and which key goes through which conduit. That might help to investigate further.

Like this, you mean?

[escargot]

Like this, you mean?

Thanks! That's exactly what I needed. Damn, I thought I searched the forum thoroughly looking for that, I just lacked the terminology.
Let's see if the mess is on the end going to the controller or if I'll have to prepare myself for a bolt mod to reach the whole membrane circuit.

EDIT: Well, I found this PDF with the whole circuit chart that shows exactly where the problem is. I'll check it out soon enough. Thanks again. More info about the internals of different Model M's here: https://github.com/rhomann/kbupgrade/tree/master/circuits

[escargot]

Well, I opened the keyboard and checked the specific terminal to the 4th column or the matrix - everything looked fine, the contacts, the traces, everything. I deattached it from the controller and did the good ol' NES cartridge trick to remove any dust inside of the connector.
And then I noticed something. The place for the controller board has some stoppers to grip it tight on the case, but apparently the mexican design for the case didn't match the taiwan made board, so the right side of it didn't reach the stoppers, making that every movement in the cable made the board move back and forth.
I took some high condensed cotton that won't release any fibers and tightened the board to the only stopper that gripped it from the opposite side of the SDL plug. Previously I did something similar to the upper holes of the backplate that wouldn't fit really "perfectly" to the nubs in the plastic case, making the board "float" inside of the case - it moved 1mm. back and forth, nothing dramatic, but still.

I also cleaned the whole case and the keycaps thoroughly, so until they dry I won't be able to test the oh-so-professional fix. If it works I'll take some pictures of they keyboard.

[escargot]

Well, of course it didn't work.

It's gotten even worse. I noticed some scratches on the contacts of the membrane, I'll be trying to get my hands on some circuit writer pen, but I think it won't be that easy to find where I'm at.

I took some pictures of the whole thing.



From the forum post given by keyb_gr I could tell that the 4th column was the one affected, and it should correspond to C3 according to the PDF with the circuit matrix, that being the 4th trace from right to left from the smaller membrane.
It's not the only one scratched though.


As you can see, the only broken rivet is at the bottom right side.


It appears to be fine from this side. (I made I mistake in the last post, it was not taiwanese, but Hong Kong made)


Same here.


What I mentioned about the PCB board not fitting correctly, only one stopper grabs it.


Unnecessary detail, but what the heck.

If I don't manage to find a trace writing pen, is it possible to trim the membrane?
In this post someone tried it, but it sounds as if he had to cut most of it, but in my case a couple of millimeters would suffice.

I hope someone can be helpful about this, this old rubber dome keyboard is killing me already.

[ch_123]

Silver gel pens are cheap on eBay, and the Chinese/Taiwanese often do free shipping to anywhere in the world.

[escargot]

Most likely somebody spilled some liquid that has seeped into the membranes.  

That's what kills most Model Ms.  I would think  SOME electronics distributor can sell 2x8mm bolt/nuts to Argentina.

I would really like to accept the diagnosis, but I don't understand why moving the cable would temporarily fix the issue. I still need to go to a bigger city nearby and check there for bolts and nuts. But I can't be 100% sure that that's the reason. Well, for that matter without proper testing I can't be really sure about anything, I know.

I suppose I'll try to get the circuit pen and try to fix the membrane and hope for the better.
Thanks for the responses.

[What is X?]

Is it the keys themselves or the controller malfunctioning? I have three keys on my 1397000's numpad that don't work, and they definitely feel weird. What's weirder is that the + key actually did work for a bit, and now it doesn't again. Furthermore, those keys actually don't come off! The keyboard just flexes when i try :/

[What is X?]

The Plus key won't come off?

Show Image

Yeah, the plus key, space/ key and ,* key. Also a couple of others on the numpad in the top right, but those three keys feel like they're not pressing the spring at all.

[theferenc]

How are you trying to pull them off? They should pop right off, just like any other key. And yes, I have one of those, and pulled off those keys to check (again...I already had for the bolt mod).

[escargot]

So I asked around, talked to people who knew about electronics and all that stuff and getting that circuit pen would be a bit difficult for now, so I "boldy" decided to trim the membrane. I plugged it in and the problem was still there, it even started to register double keystrokes, i.e. sending 3rd column-4th row (F1) with 4th column-4th row (2) after pressing F1.
I contacted a bolt factory in Buenos Aires and apparently they do make M2-8mm bolts, so I'll wait for their answer and see if I can get them in a reasonable amount. So a bolt/nuts mod might be underway.

But after one day of typing in my horrible old logitech rubber dome I just couldn't bare it. So I opened the Model M and used some rubber pieces to get the controller, the plug and the membrane really steady. It works again, don't know for how long, but I hope it does until I get my hands on a replacement or a hard fix for this one. We'll see.

Thanks for the help anyways.

[escargot]

Yup, I'm also leaning towards that theory, but until I get those bolts we'll never know.

[What is X?]

How are you trying to pull them off? They should pop right off, just like any other key. And yes, I have one of those, and pulled off those keys to check (again...I already had for the bolt mod).

I don't know, the keycaps simply do not budge. I'm sure they do actually have springs in them though, because the + key did work for a time.

FYI the keycaps are two-piece, and the top piece obviously does come off the same as the others.