P10 Issue - Pixels lighting up where they shouldn't

Freman

Uh Oh Elf
Generous elf
Joined
Aug 9, 2021
Messages
118
So, got my P10's up and running, boy that was a mission.

I have this configured as a 3x3 with the panels being in the following layout (not sure that matters)

O1P3, O1P2, O1P1
O2P3, O2P2, O2P1
O3P3, O3P2, O3P1

Every panel but my first one seems to have a stray row of pixels or two that light up when they shouldn't, anyone know what causes this and what might be a solution?

1661248316357.png1661248295802.png1661248374937.png
 

dkulp

Full time elf
Joined
Jan 2, 2013
Messages
146
Location
Framingham, MA
Couple of suggestions: 1) Check the voltage and try dropping it to 4.8V or so. 2) If on a beagle, turn out the Output by Row option. 3) If output by row doesn't fully clear it, turn on the blank between rows option.
 
  • Like
Reactions: AAH

Freman

Uh Oh Elf
Generous elf
Joined
Aug 9, 2021
Messages
118
Couple of suggestions: 1) Check the voltage and try dropping it to 4.8V or so. 2) If on a beagle, turn out the Output by Row option. 3) If output by row doesn't fully clear it, turn on the blank between rows option.

It's a smart powersupply so thankfully when I set 5.2v it maintains 5.2v - I've checked each panel and they get 5.17v
Turning on Blank between rows helped with most of it but I have the same weird artifact on three panels now (1,1, 1,3 and 2,1) are missing some cyan and instead have faint strips of green

1661251639930.jpeg
 

AAH

I love blinky lights :)
Community project designer
Joined
Dec 27, 2010
Messages
4,190
Location
Eaglehawk
Can you swap the outputs around and see if the problem follows the output or the panels.
There is/was an option on the Pi version of FPP to have a processor slowdown. That's what I would have suggested first for the problem. What the problem appears to be is that there is mirroring of the pixels from 7 lines below where the glitch is occurring. This may be an indication of a cable problem.
 

Freman

Uh Oh Elf
Generous elf
Joined
Aug 9, 2021
Messages
118
Can you swap the outputs around and see if the problem follows the output or the panels.
There is/was an option on the Pi version of FPP to have a processor slowdown. That's what I would have suggested first for the problem. What the problem appears to be is that there is mirroring of the pixels from 7 lines below where the glitch is occurring. This may be an indication of a cable problem.
I've done a bunch of software settings and that solved most of the issues but the broken cyan line on 3 panels is definitely panel specific (I've swapped cables, ports and panels and it consistantly happens on those panels) and the dead pixel/missing led are definitely a panel issue.
 

Freman

Uh Oh Elf
Generous elf
Joined
Aug 9, 2021
Messages
118
So I took the cable from O3P1 and plugged it into O2P1 and the dead cyan pixel from O2P1 doesn't persist to O3P1

But the issue on O3P1 still exists (cos clearly the pixel is missing)

For reference this is the corner of row3 P1 (currently plugged into O2)/ and row3 P2.

The green blob is 3 X 11 pixels and should be contiguous and there's no spike (like the "screenshot")

PXL_20220823_112634832.jpg

But it renders like this, I've tried 3 different cables in two different outputs.

PXL_20220823_111801863.jpg

Obviously the dead pixel is a missing pixel

PXL_20220823_111829469.jpg

When I do similar tests on the other 3 panels that are misbehaving it's much the same only they don't have a missing pixel
 
Last edited:

Freman

Uh Oh Elf
Generous elf
Joined
Aug 9, 2021
Messages
118
Ok, I haven't finished testing every pixel (I'mma write me a little tool to help with that)

What I have done is load up the test in FPP again and poke around the known bad ones.
  • Pixel circled with yellow is very prominently mounted (it is way higher than every other pixel around it) - doesn't respond to a gentle poke, might need removing and re-soldering
  • Pixel circled in red is a poor connection, given it's every color I think gnd/0v isn't connected, if you poke it it makes connection and works (and the green lines above and below it go away)
  • Pixel circled in orange is much the same as the one in red
  • Pixel circled in blue just isn't working, doesn't seem to be anything wrong, feels normal (relative to pixels around it), isn't touch sensitive.
  • Pixel circled in green is just missing, no sign of it on the floor or in either package I received.

1661251639930.jpg
 

djgra79

My name is Graham & I love flashing lights!
Global moderator
Generous elf
Joined
Dec 27, 2011
Messages
2,163
Location
Cranbourne West
Yellow Red Orange all the same pixel on the panels, might be a manufacturing issue from the same batch
 

Freman

Uh Oh Elf
Generous elf
Joined
Aug 9, 2021
Messages
118
Yellow Red Orange all the same pixel on the panels, might be a manufacturing issue from the same batch
I suspect as much, yellow has a missing pad and appears to have been hand soldered... and I have repaired it, dismatle, remove led, remove the piles of soldar, a little bit of conductive silver paint to replace the missing pad, a few minutes in the oven at 150 degrees, way to much time trying to get the led back into place, but now you'll never know it was borked

before
IMG_1422.JPG
after
IMG_1442.JPG

a bit more flux cleanup and the plastic cover back on, invisible.

I will move on to red tonight, it seems to be a dry joint cos it comes and goes even without pressure.
 

Freman

Uh Oh Elf
Generous elf
Joined
Aug 9, 2021
Messages
118
And look at that, all working perfectly

Thank you all for your tips and thank you Alan for the spare bits.

PXL_20220827_071601803.jpg
 
Top