Hey guys! I have an issue with my 3.5mm neopixel accent LED strips. The left strip is glitching up while the other side behaves as expected. Both strips are wired in parallel, share a 470ohm data line resistor soldered to data pad 3. Running the latest ProffieOS 7.14. See the Google Drive link for a video of the issue.
The chassis design makes the board quite hard to access. So my worst nightmare would be to completely disassemble the hilt, at which point I might as well do another install. And getting to the strip itself to resolder would be a huge pain. So I’m hoping someone would know what the issue is, and knows how to solve the problem?
https://drive.google.com/file/d/1XYkN-7jFTN1wJgtOgJNnsPKxVhSw3NwP/view?usp=drivesdk
In my experience those strips are really fussy. If you have even the slightest flux residue anywhere, they can glitch badly.
My advice would be to get a bunch of new wooden toothpicks and some isopropyl alcohol and just keep cleaning, cleaning, cleaning all round the solder joints. Do it for about four times longer than you could possibly think necessary before you throw in the towel and tear it apart.
I wanted to check the voltage with a multimeter, and I experienced my first magic smoke 
Both strips are now as good as dead, and even with the kill switch off they stay on.
Whenever you have two strips or neopixels hooked up in parallel and one behaves differently from the other, it’s always a wiring issue.
Too bad about the magic smoke though. That stuff is notoriously difficult to put back inside…
1 Like
Fortunately, the board itself still seems fine without any burnt MOSFETs. So I only have to replace the strips (I hope). However, I did have to cut all my wires and remove the the chassis part holding the board…
Merry Christmas!! 
I replaced the strips and their ON and OFF behaviors are as expected. However, a new problem has appeared: whenever I flip the kill switch off, the strips remain on. I have no idea how that is possible. The NPXL connector stays off however. The strips get their + power from the connector. What could the issue be? Could it be a faulty kill switch?
Video of the issue:
https://drive.google.com/file/d/1XuDlK6Qgk2uszMkjTdCwEja_fDGXa8Le/view?usp=drivesdk
1 Like
As mentioned in my reply on Facebook, you might have fried the MOSFETs just enough so that they are not actually off when powered off. This would leave them in a half-on state maybe, which could explain the dim pixels.
If you moved to an available set of LED pads and adjust your config appropriately, it would likely work correctly.
However I see you decided to replace the board entirely.
So, I would mark the bad pads on the first board and consider using it with other pads if you decide to use the board later.
1 Like
I already disassembled the board by the time I saw your reply on Facebook 
But I cleaned the other board and will keep it separately, I have a build in mind for which I can use that board.
After I’ve tested the new board and completed my install, I’ll mark this thread as solved. Thanks for the explanation btw, it perfectly explains why my LED strips were acting up.
In some cases this is caused by a manufacturing defect where the pull-down resistor for the FET isn’t connected properly. When that happens the FET will start in a random state, and may stay on when the board turns off.
Proper testing should assure that this doesn’t happen, but not all manufacturers do proper testing.
Hm, I see. I haven’t encountered such a defect with the boards I’ve used so far. But if I ever get a board where that happens right out of the gate, I’ll know what the problem is.
In this case, the issue only came up after I fried my LED strips. So a damaged FET is the most likely cause.