Model F XT Keyboard Stange Issue With Replacment PCB

[Flamin Joe]

Hi, got a bit of a strange one here.

I brought a replacement PCB some time ago for my Model F XT Keyboard to replace one which was giving the keyboard 301 error after my attempts to fix it came to nothing. After going through all sorts of fun getting it back together I was getting the 3B 301 keyboard error which is a stuck key but after many attempts I was unable to find which key was causing the issue. So I decided to test it with just the bare PCB carefully held in my hand to rule out any funny stuff with the PCB and that's where the weirdness starts.

If I hold it by the edges carefully not touching any of the pads the system will boot up fine and I can touch the various pads to confirm it's registering key presses. All good there.

If I put it down on the table, which by the way is a non-conductive wood surface, I start getting key presses??? Put it upside down (so it's the side your typing on facing the desk) and it's just the B key which is being pressed. Put it the otherside so the typing face is up and I get all sorts of keys getting pressed. Carefully pick it up from the edge and it stops.

I'm baffled as to why this is happening. I've cleaned both sides of the PCB with circuit a board cleaner so it's spotless but yet as soon as you put it on any surface it triggers the typing. Thinking maybe there's the possibility it's flexing causing these random activation of keys I held it carefully by the edges and gently flexed the board but nothing. I also tried a different cable from my other working Model F XT and it does the same thing.

At least now I know why I was unable to find the "stuck key" as it was obviously getting triggered by just being in the case but baffled as to why it's getting triggered in the first place. What am I missing?

 

[ibmapc]

Hmm.. Pretty strange. Have you checked the ground connection? I've heard of strange behavior when the ground is interupted on the model f.

 

[Flamin Joe]

I actually didn't think of the grounding as I was trying to be smart and quickly testing it outside the keyboard chassis before going through all the trouble completing it so it was never hooked up. But in saying that I did try that now making sure it was grounded but it's still the same problem.

Is there anything in particular I should be looking for to confirm the ground connection is ok? I did a basic continuity test with a multimeter and I'm not seeing an issue.

 

[evildragon]

Model F's absolutely need to be grounded since they are capacitive. My Model F started acting weird when this circuit on the house lost the ground until it was fixed. But, a Model F should actually be in the grounded barrel plates with the ground wire attached to the screw which clamps against it.

Also, if when together you have issues, make sure the foam you are using is something like neoprene foam. Some foam can build up static electricity and again, wrecks havoc.

 

[Flamin Joe]

I did actually replace the foam as the old foam had disintergrated quite badly and I did ensure the replacement was Neoprene after I read through a few foam replacement threads on other sites where it was recommended. This is the stuff I used: https://au.rs-online.com/web/p/rubber-sheets/7336731/ But then this still doesn't explain why it is occuring with just the bare PCB which I'm now making sure is grounded.

Also I didn't mention this before but out of curiosity when I opened up my working Model F to swap cables I was interested to see if it would have the same issue. So I took the keyboard assembly completely out of the chassis and without attaching the grounding wire to anything it didn't throw up any 301 errors and typing was normal. This is what is making me wonder if there is an issue with the PCB itself but then like I said if I carefully hold the bare PCB in the air by the edges (grounded or not) it appears to work perfectly normal.

I can't help but think there is something I am overlooking here as this should be pretty straightforward.

 

[modem7]

Model F's absolutely need to be grounded since they are capacitive.

Various parts of the circuitry will be connected to ground by the design engineer, for various reasons (e.g. signal return, power return, ...), but the design engineer will not have said to themself, "I will need to ground this keyboard because it uses a capacitive type key matrix." In electronics design, there is no mandate to ground one end of capacitors.

 

[evildragon]

Various parts of the circuitry will be connected to ground by the design engineer, for various reasons (e.g. signal return, power return, ...), but the design engineer will not have said to themself, "I will need to ground this keyboard because it uses a capacitive type key matrix." In electronics design, there is no mandate to ground one end of capacitors.

I can assure you, an ungrounded Model F can exhibit a mind of its own. The grounded steel places are without a doubt, very important to prevent RFI. My Model F with the plates removed does what OP describes, it has a mind of it's own. Only when the plates are clamped properly around the PCB and grounded does it actually stop ghost writing and stays silent until I press a key. You are saying ground one end of the capacitors but I never said this. Without a proper ground encasing the PCB where the matrix is, RFI will indeed ghost type, just putting on a table, near a wireless router, etc. That ground wire isn't attached to the screw and plate for just show. It's functionally required.

 

[modem7]

I can assure you, an ungrounded Model F can exhibit a mind of its own. The grounded steel places are without a doubt, very important to prevent RFI. My Model F with the plates removed does what OP describes, it has a mind of it's own. Only when the plates are clamped properly around the PCB and grounded does it actually stop ghost writing and stays silent until I press a key. You are saying ground one end of the capacitors but I never said this. Without a proper ground encasing the PCB where the matrix is, RFI will indeed ghost type, just putting on a table, near a wireless router, etc. That ground wire isn't attached to the screw and plate for just show. It's functionally required.

You wrote, "Model F's absolutely need to be grounded since they are capacitive." That is not the reason they are grounded. As I wrote, they are grounded for other reasons.