Troubleshooting Mekanisk Nerd60 PCB

[Mad_Maxx]

So I've had a white GON Nerd60 Mekanisk edition from 1upkeyboards in an Fjell for about 6 months, and it's been working flawlessly......

Until yesterday.

Here's the timeline:

1. Spacebar stops working
2. Reflash, no dice
3. Look at joints, solder looks good
4. Replace switch, still no dice
5. Start messing around with software, realize that no switches east of left winkey work
6. Diodes look fine
7. Switches look fine
8. Check continuity between switches
9. Get a high resistance from one switch terminal to the next switch. I don't have a pic, so I'll attempt to ascii it (looking at it from the bottom of the pcb, o's being switch terminals):

Code: [Select]

   space*          alt*            winkey            ctrl
-----o------0Ω------o------30MΩ------o------0Ω------o
--------o--------------o----------------o---------------o
*not working

I'm no electrician, but all the traces look good as far as I can tell.
At this point, I'm not really sure where to go next. or if the pcb is bad.

Should I try to jump the high resistance terminals?

Many thanks for any support.

[suicidal_orange]

Sorry for the slow reply, somehow missed this thread...

That's a big resistance - are you sure there's a trace between the two not a diode?  If there is a trace you can solder a jumper wire between the two and it will probably fix all the keys, if there isn't then one or more keys will appear to be stuck but no harm done so give it a go.

As long as one key in each row and column works the controller is good, the rest of the PCB is just wires and can be bypassed.

[Mad_Maxx]

Sorry for the slow reply, somehow missed this thread...

That's a big resistance - are you sure there's a trace between the two not a diode?  If there is a trace you can solder a jumper wire between the two and it will probably fix all the keys, if there isn't then one or more keys will appear to be stuck but no harm done so give it a go.

As long as one key in each row and column works the controller is good, the rest of the PCB is just wires and can be bypassed.

Hi! I tried jumping them the other day. The row worked, but they registered the wrong keys...

I'm not sure what that means exactly. I assume GON doesn't have schematics available anywhere?

[suicidal_orange]

If they registered the wrong key you connected something wrong, there are no schematics available only the pinout.  If you post a pic of the back with the traces showing in the shine I'll have a look - should be fixable.

[Mad_Maxx]

If they registered the wrong key you connected something wrong, there are no schematics available only the pinout.  If you post a pic of the back with the traces showing in the shine I'll have a look - should be fixable.

Hey thank you I forgot I posted this! Here's a shot of the traces, hopefully the pic is good enough.

[suicidal_orange]

Ah it's this one - the 'columns' are actually U shaped, so it's not immediately obvious what should be connected.  I would say that as you have two pairs with zero resistance those are the columns.

The other side switch pins will not connect to each other but each goes to a diode and the other side of those diodes will be connected, possibly alternating?  Though that wouldn't explain the whole corner dying...

I need a bigger screen to look into this properly.

[KHAANNN]

If it's just the spacebar, you could suck the solder perfectly out of the spacebar switch, keep it in place, gently run wires to an empty switch spot instead, map that switch spot to spacebar

I just had a similar issue myself, it was a broken diode, how did you verify that diodes are fine? (You said diodes "look" fine, you can verify that they are fine by measuring their resistance, one way it's 0, the other way it's around 600)

[Mad_Maxx]

If it's just the spacebar, you could suck the solder perfectly out of the spacebar switch, keep it in place, gently run wires to an empty switch spot instead, map that switch spot to spacebar

I just had a similar issue myself, it was a broken diode, how did you verify that diodes are fine? (You said diodes "look" fine, you can verify that they are fine by measuring their resistance, one way it's 0, the other way it's around 600)

Unfortunately its the whole row, minus the two left-most switches. That's an awesome hack tho!

Poor wording on my part. When I said "looked" I meant I had measured them, and they were all good.

[suicidal_orange]

I'm looking at this again trying to get inside Gon's clearly disturbed mind (who makes a matrix like this?!)  and realised your picture shows the left two switches which work because it's the back of the board!  Should have noticed that earlier )

Actually the right side (when viewed from the switch side, left side when soldering) has one 'column' running along all the dead switches which means your problem does make sense.  Actually I think you have two problems - the dead 'column' which is probably a break between the spacebar switch and the controller,  possibly done when you replaced the switch, and whatever caused spacebar to stop working in the first place, possibly the diode. 

The diode should have continuity one way but not the other so you can test that - if it has no continuity either way it's dead.

Soldering between the 'column' and the controller chip the switch could be messy - can I get a picture of the spacebar switch and the controller like the one above to scribble suggestions for continuity tests/jumpers on?