Printable View
I have a neglected A4000 keyboard that been boxed in the loft for about 10 years. It worked when boxed, but now I find only a handful of the letter keys work.
Any tips as to where I should start troubleshooting - or is this likely to be beyond salvation?
So far all I've tried is cleaning the contacts on the internal ribbon cable - with no effect.
When I get a chance I'll make a list of the keys that are working.
just strip it and clean it , i did it recently on here and made a thread - i'll find it :-)
---------- Post added at 19:52 ---------- Previous post was at 19:50 ----------
http://www.amibay.com/showthread.php?t=18014
@mjnurney
Nice!
@robinsonb5
If after cleaning as per mjnurney's suggestions you still have trouble, try exchanging the pressure pads on the problematic keys with ones from well behaving keys to see if it's the pads. I had this on an A2000 keyboard, ended up using the pads and keys from an A500+ keyboard and all was fine again.
Well now I'm thinking it's either the MCU or the connector between the membrane and MCU board, because of the way the working keys are clustered.
Here's a list of the working keys:
F3 F4 F6 F7 F8 F10
Backspace
Tab
Caps Lock
T Y I P [
G H K ; '
Right Shift
Right Alt
Right Amiga
Del Help
Cursor Right Cursor Up
Numpad: ( ) /
The cluster of working letter keys having the same "shape" in two rows seems too big a coincidence to ignore.
EDIT: And now I've found the problem.
I tried testing the signal path from the various pads on the membrane to the connector, and found that in many cases there's no signal path.
So then I tried testing from each pad to the others that are connected to it, by tracing the tracks on the membrane, and found several such links that are broken. I can't see any reason why these tracks are no longer connecting the pads - there's no obvious damage or corrosion.
So now what? Is this an ex-keyboard? I'm assuming soldering new links to the plastic membrane would be out of the question! Conductive pen, maybe?
Each row has many pads connected in series. So does each column. Resistance from the connector end will vary accordingly and might mislead a continuity test, so check continuity beginning from the connector end and proceeding between each pair of consecutive pads until you reach the open end. A single broken link would disable and entire row segment and this doesn't seem to be the case according to the working keys table you posted above, of course that's just a visual impression and your data could actually correspond to a specific set of dead links.
So the list you provided is 100% accurate? That is these keys are verified to work and the remaining keys are verified to NOT work?
Also, did you try swapping the pressure pads from working keys into non-working ones and see if it fixes it?
Easiest fix,
Buy any Amiga Keyboard (A2000, A30000) and swap out the circuit board, the cable just unplugs and all components are swappable, to go through and try to fix each and every key is just a bit too much.
I can even send you a known working board, but shipping might kill you.
Yup, I noticed that - I was using an el-cheapo analogue multimeter, and found that the buzzer mode wasn't going to work, so was watching for needle swings with the meter set to 10KOhm.Quote:
Originally Posted by BLTCON0
Or at least the portion of that row segment that's beyond the break.Quote:
so check continuity beginning from the connector end and proceeding between each pair of consecutive pads until you reach the open end. A single broken link would disable and entire row segment
Yeah, I think what I've got is random breaks some way down each row and column, and the keys in my table happen to be the intersections of the pads that are still working!Quote:
and this doesn't seem to be the case according to the working keys table you posted above, of course that's just a visual impression and your data could actually correspond to a specific set of dead links.
I can't swear to it, but I'm 90% certain the list's accurate.Quote:
So the list you provided is 100% accurate? That is these keys are verified to work and the remaining keys are verified to NOT work?
Nope, but I did use a pair of needle-nose pliers to bridge the pads by hand, and verified that the working pads on the membrane registered a keypress that way, and that the non-working ones didn't.Quote:
Also, did you try swapping the pressure pads from working keys into non-working ones and see if it fixes it?
@quarkx: Thanks for the suggestion - I'd noticed that the keyboard itself had "A500 stamped into it and crossed out, while the MCU board has "A3000" written on it! Anyhow, yeah, getting a new membrane's certainly the path of least resistance - but I'm still intrigued to know what's actually happened to this one and why. :)
lol, that's why sometimes I say that language should be written like math, with parentheses and well-defined operator priorities and relations ;)Quote:
Originally Posted by robinsonb5
Perhaps I should have written "entire segment of a row" to make it clearer, yes of course the harm would be beyond the break.
That'll do it as well... I'm with quarkx on the replacement solution.Quote:
Originally Posted by robinsonb5
If you test continuity between different points of the same pad, what are the results?
The stamped and crossed out A500 marking is also found on the back of the plastic backplate of the A1200 keyboard (a smaller A1200 has been stamped nearby)
LOL - yup, there are definitely times when that would make life easier!Quote:
Originally Posted by BLTCON0
Yeah, the pads themselves are fine - I did try that to make sure there wasn't some kind of corrosion I couldn't see. There are definitely tracks that no longer carry the signal - though I can't see any sign of damage to them. Maybe it's the joints where the tracks meet the pads - I don't know. I think I'm beyond the point where I can usefully spend more time on it though. I'll keep my eyes open for a "new" membrane at some point :)Quote:
If you test continuity between different points of the same pad, what are the results?
Yeah just get a working membrane.
Now let's talk pr0n :p (= photos)
Since this one is considered a 'dead' membrane, you could perhaps dissect it a bit :D
Remove the pads from the non-working keys with a sharp razor or something to the same effect, the conductive track underneath would then be exposed and you could directly check conductivity (ruling out - or ruling in - the tracks themselves as the actual problem cause).
If they're ok I suppose your theory is confirmed - conductivity has been lost between track and pad.