Have I ruined my switches?

[Bobblybook]

Hi everyone,

I assembled a kbd8xmk2 some months ago. My switches seem to have issues with not always registering - I am using holy pandas that I cleaned and lubed myself.
I'm not sure what the issue is - but for quite a handful of keys, they trigger inconsistently. They also have "bad" and "good" periods (dependant on temperature maybe?).

I have desoldered a few of the problematic ones and checked them and they seemed completely fine, but when reassembled they are back to acting odd.


Basically, some key presses don't register at all. Also, when testing some of the bad keys, some repeated keypresses will double type the key. Sometimes there is a big difference in consistency when hitting the key directly centered vs. off-center (have taken the keycaps off to test this).

Also I desoldered a board of mx-browns and cleaned and lubed them, and some of them seemed to show similar behaviour after reassembling, though not as bad.

It initially seems like some switch/contact issue. Though I can't see any obvious problems..
Could there be some problem with the way I cleaned them? I just ran them in an ultrasonic cleaner with some mild soap, and rinsed them well afterwards.

I checked the leaf alignment and contact, it seems perfect. I have accidentally bent a few leaves in the past so I know what that looks like, these all seem pristine and I checked the clicking action when disassembled to check that the contact is made properly inside the switch..

I was extremely careful when lubing to not get any lube on the leaf or contact points, so this shouldn't be an issue. I just lubed the stem and housing with a small brush with 3203, and avoided the legs and any part of the leaf.

Soldering is fine, I'm not a noob to that and the joints are all good.

Is this sort of intermittent switch contact a known phenomenon when X or Y is done badly? I'm completely in the dark about what to try and test next, I'm just not sure where to go from here
Thanks to anyone who takes time with this.

[retoid]

Maybe worth reflowing the diodes, could be some of the solder joints on them were fractured which can happen.
Sometimes the fractures can be so thin you can't see them even under a magnifying glass, usually only with a microscope. Which will make them react as you describe, sometimes working, sometimes not.

[tp4tissue]

are you using lead free solder ?  Switch to leaded.  and check the pcb for cracks.

[Sup]

What is your way of lubing?

[Bobblybook]

Maybe worth reflowing the diodes, could be some of the solder joints on them were fractured which can happen.
Sometimes the fractures can be so thin you can't see them even under a magnifying glass, usually only with a microscope. Which will make them react as you describe, sometimes working, sometimes not.

I'll check this out, ty. I didn't imagine there would be anything wrong with SMD diodes from factory, but I'll see what I can do about reflowing a few of them.

What is your way of lubing?

I followed Nathan Kim's youtube video - thin amount of 3203 on a brush, thin film brushed on to appropriate surfaces. The behaviour of the problem switches seems in-line with residue on the contacts causing problems, but they are completely clean of any lube or oils or any gunk.

are you using lead free solder ?  Switch to leaded.  and check the pcb for cracks.

Nope, kester 63/37. I have a $$$ soldering iron and have done a lot of soldering before, so this is likely not the problem. However as retoid mentioned, there may be micro fractures in some of the diode connections, especially given the flexing of the pcb. I'll post some photos later today.. just need to have some breakfast first!

[Sup]

Maybe worth reflowing the diodes, could be some of the solder joints on them were fractured which can happen.
Sometimes the fractures can be so thin you can't see them even under a magnifying glass, usually only with a microscope. Which will make them react as you describe, sometimes working, sometimes not.

I'll check this out, ty. I didn't imagine there would be anything wrong with SMD diodes from factory, but I'll see what I can do about reflowing a few of them.

What is your way of lubing?

I followed Nathan Kim's youtube video - thin amount of 3203 on a brush, thin film brushed on to appropriate surfaces. The behaviour of the problem switches seems in-line with residue on the contacts causing problems, but they are completely clean of any lube or oils or any gunk.

are you using lead free solder ?  Switch to leaded.  and check the pcb for cracks.

Nope, kester 63/37. I have a $$$ soldering iron and have done a lot of soldering before, so this is likely not the problem. However as retoid mentioned, there may be micro fractures in some of the diode connections, especially given the flexing of the pcb. I'll post some photos later today.. just need to have some breakfast first!

3203 Shouldn't do weird stuff with the contacts. Try to bridge the contact points on the PCB and see if that has the same effect if it does then it's the PCB.

[Bobblybook]

Ok some developments on this - It's clearly a switch issue.
I reflowed a few of the diodes, made no difference.

Testing the switch contacts on the pcb with tweezers is a little difficult because a little slip of my finger can result in a double key press. Since the issue I'm experiencing is sporadic anyway (I went a streak of 30+ successful key presses today with the "dodgy" switch while testing), I couldn't be sure whether this test case was actually working fine or the same as the bad switch.

Instead, I have a box of tangerines here and I held one on the back of the pcb and pressed it a bunch of times. It worked perfectly, but the panda on the other side didn't.

I soldered some test wires to the switch contacts to measure the resistance of the switch with a multimeter. This was interesting..

A lot of the time, the resistance of the bad switch showed infinite, even when pressed. I mashed it a bunch of times, as I often have to to get the key to register, and I started getting readings. But every key press was a different resistance, and way too high.

I'd press it once, get a reading of 30 ohms. Again, 15 ohms, again 45 ohms, 8 ohms.. 3 ohms.. 30 ohms. Just all over the place, and occasionally an infinite reading. So it seems like the readings I am getting would correspond to the "working" key presses, but this shows that even those are not functioning correctly. I measured the tangerine switch and it showed 2 ohms every time, which is what I would expect for a closed switch.


I also noticed that wiggling the stem of the switch caused the resistance to change a lot, ranging from high single digits (~8) to infinite.

Edit: Here is one of the bad switches opened up.




I pressed the slider a few times while connected to the multimeter, and it showed the same bad behaviour. Very rarely a reading of <3 ohms.
It seems like maybe I need to slightly bend one of those two copper sheets towards the other. Which one do I bend?
I'm guessing the one on the left (the short, stumpy one).. I've bent the right leaf slightly on a previous switch and it alters the tactile bump significantly, and seems impossible to restore 100% back.

Or maybe I should not touch the leaves themselves but instead try and get the little contact nub to protrude a little more?

I never really bumped these contacts when cleaning or assembling, they seem extremely sensitive to their position and angle. Not sure why I seem to be experiencing this with so many of my switches..

[Leslieann]

It's possible these are just sensitive to how they're mounted, sort of like Matias.
Does the leaf move if you touch the contacts on the bottom of the switch?

[Bobblybook]

It's possible these are just sensitive to how they're mounted, sort of like Matias.
Does the leaf move if you touch the contacts on the bottom of the switch?

It doesn't move on this one. But these legs are now coated in solder and it holds them in place a lot more than it would have when they were untinned.