Possibly fried xarmor - HELP PLEASE :(

[Bry]

Hello Geekhack,
So a couple of nights ago I was pretty drunk and sleeping when I awoke to a loud noise, which I ignored. Then the next morning I woke up and realized that a glass of water had fallen over and onto the floor. I didn't think much of it until now, but I think you can guess where this is going.

My U9BLS suddenly stopped working properly. The space bar, escape, g, h, ', and up arrow keys no longer work. I unplugged the board and took it apart. This is where it gets weird, I tried shorting the solder points for some keys, and the keys that I mentioned still didn't respond. The other keys I tried did respond to this.

So the problem isn't with the switches themselves, but with something else. (unless my logic is flawed, shorting the solder points should bypass a faulty switch)

I'm at university so I can't really desolder the switches (which I don't think would matter based on the line above)


Relevant info:

1. it was water that spilled, and I don't know how much actually got on the board
2. there was about 48 hours between the time it spilled and the time it stopped working
3. the affected keys are not located near eachother on the board except for the g and h keys
4. I tried the board on a different computer and the same symptoms exist
5. same thing happens when plugged in with USB and PS2
6. all the LED's still work fine


I searched through the Wiki on here to no avail. The section about fixing keys seems to focus on the switches being faulty or corroded, which they don't seem to be (because of the issue with shorting the solder points)


Well, that is about as much as I can say for now. If anyone has any ideas I'd love to hear them


Thanks <3

[mich]

If it's an NKRO board then there should be a diode (not to be confused with backlight led) connected in series with every switch. Did you try bypassing these diodes as well?

[Bry]

If it's an NKRO board then there should be a diode (not to be confused with backlight led) connected in series with every switch. Did you try bypassing these diodes as well?

I was wondering about this, so thanks for bringing it up. The problem though, is not knowing which solder points are for the diodes, as it seems there aren't one for each switch. I could be wrong though. I'll keep messing around with it later and maybe upload some pictures. Some people might like to see the internals of these xarmor boards anyways

Either way, if it is a diode that is shot, I'm going to have to desolder all the switches and LED's, so it might be a good opportunity to begin the ergo-clear project I've been wanting to have.
In the meantime, I'll be ordering a TKL Filco Ninja browns once I get the money for it in like 3 days.


By all means, if anyone has any more knowledge to share about the internals of these boards, please do. I'm gonna need to know the part number on the diodes if I eventually order them. If it isn't repaiarable I guess I have a ton of LED's, key caps, and brown switches :/

It sucks too because the board is mint condition besides being unable to type anything that requires the use of the spacebar lmao


Edit:
To answer your question, it is NKRO capable with the use of the PS2 cable (same as all NKRO boards).

[Bry]

http://solderslingers.com/cart/index.php?main_page=product_info&cPath=2_15&products_id=8

I think I've found the diode that this board uses between each intersection. Unfortunately, the way they are soldered, I cannot bypass them without access to the top of the PCB, which is covered by the plate. They pins don't extend through the bottom of the PCB


Anyone know whether or not these are the right ones? Im sure most NKR boards use the same things.

[simon_C]

go to the grocery store. buy a 5lb bag of rice. strip keyboard down to bare pcb and switches. put in plastic bag with all the rice. seal it up good. leave to sit for a few days. it will draw ANY moisture out of the whole system. try again then

[Pretendo]

Yes, best advice is to leave it alone for a few days.  Powering it on may cause a short circuit if it's still wet anywhere inside.  The last thing you want to do is fry the controller board.

[noodles256]

wlcm2gh

[Bry]

Well, it was plugged in when I realized they keys stopped working. It's now been unplugged, apart, and borrowing my turtle's heat lamp for 3 days to dry. They still don't work


What could be causing this? I would assume that if it were the controller board, it wouldn't just be a few random keys that would be affected. Anyone?

[mich]

What could be causing this? I would assume that if it were the controller board, it wouldn't just be a few random keys that would be affected. Anyone?

Yes, it seems like the problem is somehow local to the affected keys. Do you have a multimeter, or a friend who has a multimeter, or a friend who has a friend ...?

[Bry]

Yes, it seems like the problem is somehow local to the affected keys. Do you have a multimeter, or a friend who has a multimeter, or a friend who has a friend ...?

I have one, but it is three hours away at my parents' house (I'm at Uni alone in my little apartment). Even if I did have one, I'd need to desolder all the switches to actually be able to see the diodes, because they are hidden under the plate. The are not accessible from the bottom of the PCB, unlike the pins of the switches and LED's, which extend through to the bottom where they are soldered on. I am beginning to think this may become a winter break project lol.

[mich]

Here's a diagram of a simple NKRO keyboard circuit, in case you have never seen something like that before.
http://www.headfuzz.co.uk/midihack2

In a real computer keyboard the "vertical" and "horizontal" lines don't need to be exactly straight, but still there are two groups of lines and each key is connected to exactly one line from both groups.

Even without a multimeter, it should be possible to locate the exact source of problem by connecting various points of the circuit, e.g.:
- bridging input of a dead key with output of it's diode will result in a keypress if and only if all other connections and the microcontroller are working properly
- assuming the above, bridging output of a nonworking switch with output of some other switch connected to the same "input" line will result in the other key appearing to be pressed whenever you press the tested switch if and only if the tested switch works correctly
and so on.

[Bry]

Here's a diagram of a simple NKRO keyboard circuit, in case you have never seen something like that before.
http://www.headfuzz.co.uk/midihack2

In a real computer keyboard the "vertical" and "horizontal" lines don't need to be exactly straight, but still there are two groups of lines and each key is connected to exactly one line from both groups.

Even without a multimeter, it should be possible to locate the exact source of problem by connecting various points of the circuit, e.g.:
- bridging input of a dead key with output of it's diode will result in a keypress if and only if all other connections and the microcontroller are working properly
- assuming the above, bridging output of a nonworking switch with output of some other switch connected to the same "input" line will result in the other key appearing to be pressed whenever you press the tested switch if and only if the tested switch works correctly
and so on.

This reflects my basic understanding of the function of diodes, so I guess I need to go to walmart and get a soldering iron so I can actually get at the diodes to test this.