HELP - major issue with failing pixels?

[Ktag]

HELP!

Having so many problems with not only my matrix but other props as well, including my PPD Wreath. Have swapped out more than 30 pixels, with no let up in site. Is it a BAD batch of pixels or is there something sinister underneath!!! I did mention some of this back in an older thread...https://auschristmaslighting.com/threads/intermittent-display-issue-in-matrix.14897/post-128108

At this rate I won't have a Halloween show which would be a nightmare (no pun intended) based on the amount of work I have put in.

- swapped out cables; reviewed all power injection wiring; disconnected power injection; swapped ports; checked voltages (from controller & at prop)(have used 14 guage wiring); added in Null buffers near controller, just in case (note PPD wreath is less than 2 metres from differential receiver); tried different diff receiver, different network cable, different power supply

I think it is the pixels, but when do I stop having to splice in replacement pixels (from the same batch). Is this possible?

All these pixels were bought earlier this year - 5000 x 5v. I haven't had any issues with last year's pixels... both 12v and 5v.

Thoughts on what I could try next... or do I just keep replacing pixels! Note: I've sent an email to Tom Zheng

 

[Notenoughlights]

Have you tested voltages to check that your pixels are getting the voltage they require, and nothing more?

 

[AAH]

String lengths?
Position in the string?
Position in relation to power injection?
Voltage from the power supply with no load?
Does squeezing the pixels do anything?

Can you post pics of the faulty pixels showing both sides of pcb.

 

[Ktag]

Answering your queries Alan...

String lengths? 610 ish depending on prop
Position in the string? varies, replaced about 25 pixels so far across multiple strings, identified at least 10 more spots where it intermittently doesn't push data through, or it is corrupted
Position in relation to power injection? on my matrix power injection at position 1, 61,183, 305, 427, 549 running at 30%
Voltage from the power supply with no load? 5.26v main cable; PI 5.24v (slightly longer run)
Does squeezing the pixels do anything? Tried but no, however, where it is "failing" it will correct itself for a time but then corrupt or no sending of data. It appears to be happening at a lot of places. Fix one, identify another etc etc. not occuring because I fixed one... just randomly.

 

[Ktag]

Just checked my other differential build which I tried as well with no change in result, and it has a voltage of 5.17v outputting through the cable. My main control box outputs 5.07v.

 

[AAH]

With squeezing fixing/breaking the pixel it looks like a manufacturer fault and replacement by them is the only option.
When the pixel goes faulty is everything downstream from it failing to work or are they just dropping a colour/s?
If you poke at the epoxy is it hard like plastic or somewhat spongey? Epoxy and the way it goes off can be affected by both temperature and humidity which sadly means that some pixel batches can be affected solely because it's been a hot muggy day versus a cool dry one.
When you next have a bad one can you specifically put pressure onto the side/area of the pixel over the black IC to see if that does anything.

 

[Ktag]

When the pixel goes faulty, everything down stream, stays on what the data previously had sent mostly. But on odd occasions it will flicker. Then of course it might right itself and work fine for a while.

this morning I have adjusted the voltage down to 5.04v, same problems. But when about to replace a pixel, squeezing it did make a difference.

Should I continue to replace each offending pixel, in the hope that by adjusting voltage down i've removed the source of the problem?

 

[AAH]

If squeezing makes a difference then the pixels are faulty with a solder fault and sadly replacing them is going to be the only solution

 

[uncledan]

Where did you store them? I lost a ton of pixels one year after storing in cargo container with no ventilation. The heat did a number on them.. massive failures. Other pixels stored in my garage from same order are still running fine. I tossed out 5k+ pixels from Mega Tree that were stored in cargo container. Not helping the current situation but a note for all to keep in mind where storing blinky

 

[Ktag]

Thanks. Stored in garage.
Just want to know if I’ve been pushing 5.26v through them. Would this have damaged them, if so … all of them or some of them?

 

[Notenoughlights]

If you are able to get some numbers from the chips on the dead LED's that should let us find a datasheet. I just had a look at ws2811 for the sake of it, and their max supply voltage to the chip is 6.0V-7.0V.

 

[Ktag]

Inscription on the black chip inside one of the pixels is
"WORLDSEM
WS2811
20210715"

Also attached is a clip where I managed to film a poor pixel coming back to life by squeezing. This has happened a few times now but in the end the pixel plays up down the track.

Contacting pixel supplier as I think I have exhausted all other possible reasons?? Admittedly when I bought these, they gave me an option to receive a batch from OS or one stored in the AU warehouse, of course I chose the AU warehouse! maybe I shouldn't have.


View: https://youtu.be/iyPd3jifUbo

 

[uncledan]

That sounds like a bad solder joint. If you've had this many issues sounds like some dodgy nodes. Definitely contact the supplier

 

[Ktag]

Spoken with supplier who whilst willing to help, is confused why the failure is happening as I described... ie. worked correctly when tested but once on display outside for a period of time has started to fail many, many pixels. Has offered 500 replacement pixels but the batch in question was a 5000 pixel order. I am worried that this will continually keep happening especially as these pixels were mainly loaded into my new high-density props and are extremely difficult to conduct microsurgery every time another pixel fails whilst up on the house.

A further example of the issue... On my 600-pixel PPD wreath (not the worst affected prop, that being my 1860-pixel matrix - see example, Note: Matrix wired in thirds, starting at bottom left) I have replaced so far about 8 single pixels and two 50 strings in the middle (due to giving up replacing individual pixels in that section). I then ran the prop on a full Red-green-Blue-White test cycle for a few hours - working as it should to start with but when I returned it looked like the attached. Obviously at least 1 pixel has faultered but based on the colouring downstream others may have as well. The wreath is wired clockwise from the outside to the middle.

Thoughts


View: https://youtu.be/paUl2oAwQ1o


View: https://youtu.be/rDuvzXFmQa0

 

[Notenoughlights]

I would put extreme pressure on them to replace the lot and have them take the broken lot you have back to test them. Sounds like they may not have had their solder hot enough to properly adhere to the components/PCB in the pixels.