Diagnosing a dead switch

Plasmodium · 20 Dec 2015, 17:49

So, I've been working on a THKB prototype board and after soldering it all together, everything works...except for the "Z" and "LGUI" (Win?) key. Now, I want to diagnose if this is a problem with the switches involved or the diodes or solder joints or something else? How would I go about testing this with the minimum amount of desoldering?

Pictures can be provided if needed. I shan't be able to actually do any real desoldering till Tuesday, but I'm happy to test other things in the meantime!

RoastPotatoes · 20 Dec 2015, 18:08

Plasmodium wrote:So, I've been working on a THKB prototype board and after soldering it all together, everything works...except for the "Z" and "LGUI" (Win?) key. Now, I want to diagnose if this is a problem with the switches involved or the diodes or solder joints or something else? How would I go about testing this with the minimum amount of desoldering?

Pictures can be provided if needed. I shan't be able to actually do any real desoldering till Tuesday, but I'm happy to test other things in the meantime!

With it plugged in get a bit of wire or something to join the two pins. If it is the switch then you should be able to create an input that way.
Also of you have a multimeter that could be handy.

Plasmodium · 20 Dec 2015, 18:18

Thanks. No multimeter, unfortunately - but presumably you could make a ghetto one with a battery and a small lightbulb?

Also, re: bridging the pins, can I do that while the switch is still in? I thought Cherry MX switches sent the signal by breaking the connection rather than making it - or am I making this up?

RoastPotatoes · 20 Dec 2015, 18:26

Plasmodium wrote:Thanks. No multimeter, unfortunately - but presumably you could make a ghetto one with a battery and a small lightbulb?

Also, re: bridging the pins, can I do that while the switch is still in? I thought Cherry MX switches sent the signal by breaking the connection rather than making it - or am I making this up?

Hmm I have to say I don't know. But I think this is a thing I have done in the past. Have a go I think I have done it on my gon before.

Plasmodium · 20 Dec 2015, 18:36

Cool. Out tonight but I will have a go tomorrow morning and post the results!

jou · 20 Dec 2015, 21:14

Plasmodium wrote: Also, re: bridging the pins, can I do that while the switch is still in? I thought Cherry MX switches sent the signal by breaking the connection rather than making it - or am I making this up?

You can do this with the switch soldered in. Cherry MX are normally open, i.e. It makes contact when you press it

Plasmodium · 20 Dec 2015, 21:50

jou wrote:
Plasmodium wrote: Also, re: bridging the pins, can I do that while the switch is still in? I thought Cherry MX switches sent the signal by breaking the connection rather than making it - or am I making this up?
You can do this with the switch soldered in. Cherry MX are normally open, i.e. It makes contact when you press it

Thanks! That's useful. In that case, there's a decent chance it's just the contact inside that's a bit bent - because one of my spares seemed to be touching all the time when I opened it up to have a look.

jou · 20 Dec 2015, 22:27

The contact leaves do touch without the slider. The slider keeps them apart until you reach the actuation point:

Plasmodium · 21 Dec 2015, 11:45

Ok. Tried bridging the pins on the bottom and neither of the affected keys worked (but others did, so I was doing it right). What next?

Plasmodium · 22 Dec 2015, 13:28

How does one go about testing the diodes? That's the only other component involved apart from it being a fault in the PCB (caused by me or otherwise).

Matt_ · 22 Dec 2015, 19:58

Perhaps a problem with your firmware configuration?

To see if the diode is faulty :

Instead of bridging 1 and 2, bridge 1 and 3 to connect the row and the column while bypassing the diode.

Plasmodium · 22 Dec 2015, 21:15

Thanks for that. I'll try that tomorrow and get back to you. I was wondering about firmware myself, but it's just the firmware that ne0phyte pre-loaded!

Plasmodium · 23 Dec 2015, 19:10

It is the diode in both cases - ie, bridging the diode creates the expected output. Looking at the board more closely I see the LGUI diode is actually soldered in backwards (duh...my bad and I thought I double checked all of them!). The Z diode is in correctly, but is presumably faulty (or perhaps the soldering is not up to scratch. I will resolder both at some time after Christmas and report back then!

Plasmodium · 14 Jan 2016, 12:38

Just realised I never reported back with my eventual success! I posted on the THKB thread to show off the completed board, but I never came back here to thank you gents for walking me through this.

So, yeah, thanks for taking the time to help a newbie, chaps!

nicstreet · 14 Jan 2016, 13:36

Plasmodium wrote: It is the diode in both cases - ie, bridging the diode creates the expected output. Looking at the board more closely I see the LGUI diode is actually soldered in backwards (duh...my bad and I thought I double checked all of them!). The Z diode is in correctly, but is presumably faulty (or perhaps the soldering is not up to scratch. I will resolder both at some time after Christmas and report back then!

This is worthwhile test for anyone with issues, I had a series of four switches that appeared to be dead due to a dead diode, bridging the diode(s) allowed me to pinpoint offending item and replace.