timon wrote:Please, this issue is being discussed openly and honestly.There is a lot of reasons how and why chips can fail especially when the manufacture has a bad batch. Ray makes pixels PCBs in large batches it would be quite easy for a bad batch of chips to only have been used in the TC considering a spool of WS2811s contains about 2,000 chips and a batch of pixels could be 100,000 pixels.
Maybe I've misunderstood the history...
We have had TC failures with BOTH WS2811 (bad PCB) and the TM1804 that did not have the same PCB issues. Sometimes I hear bad PCB as the cause, sometimes higher draw on the LED element, and sometimes missing conformal coating.
I've heard the problem described as a change in the design from what Zach originally designed/tested/approved.
Ray Wu Technicolor pixels had problems -- but I heard that HolidayCoro Firepix pixels ALSO had the same problems. I understood each to come from different manufacturers (factories). There were differences between Technicolor and Firepix (clock speeds were different preventing Firepix to work with DLA controllers) but the failure rates were common despite different sources.
Maybe there is misinformation here as we try to figure out the behavior and cause (but not necessarily the blame)... workarounds with voltage and intensity were ineffective. overheating and moisture intrusion might be part of the problem, but even that doesn't fully explain the issues.
I hope Timon's work will resolve the problems... but I don't know that it fully explains the behavior we've seen.
Steve Gase wrote:Maybe I've misunderstood the history...We have had TC failures with BOTH WS2811 (bad PCB) and the TM1804 that did not have the same PCB issues. Sometimes I hear bad PCB as the cause, sometimes higher draw on the LED element, and sometimes missing conformal coating.
I've heard the problem described as a change in the design from what Zach originally designed/tested/approved.
Ray Wu Technicolor pixels had problems -- but I heard that HolidayCoro Firepix pixels ALSO had the same problems. I understood each to come from different manufacturers (factories). There were differences between Technicolor and Firepix (clock speeds were different preventing Firepix to work with DLA controllers) but the failure rates were common despite different sources.
Maybe there is misinformation here as we try to figure out the behavior and cause (but not necessarily the blame)... workarounds with voltage and intensity were ineffective. overheating and moisture intrusion might be part of the problem, but even that doesn't fully explain the issues.
I hope Timon's work will resolve the problems... but I don't know that it fully explains the behavior we've seen.
It is a confusing situation, with several problems surfacing at the same time. Best solution, I think, is to keep the calm and not run around in a million different directions. While I don't know Timon personally (especially since I don't know who is behind the pseudonym), I think that it would be best to let him (or her) take the lead at this time, and not try to stir any pots now. While it rermains to be proven, I think that it's primarly due to quality control problems, rather than some unknown fundamental design flaw. It is very hard to resolve problems like this when anxious people keep coming at you.
I'm a he:biggrin:, a computer engineer turned System Administrator. Semi retired and returning back to my roots as a hardware/firmware designer.
Timon is my alter ego.
John
Steve Gase wrote:Maybe I've misunderstood the history...We have had TC failures with BOTH WS2811 (bad PCB) and the TM1804 that did not have the same PCB issues. Sometimes I hear bad PCB as the cause, sometimes higher draw on the LED element, and sometimes missing conformal coating.
I've heard the problem described as a change in the design from what Zach originally designed/tested/approved.
Ray Wu Technicolor pixels had problems -- but I heard that HolidayCoro Firepix pixels ALSO had the same problems. I understood each to come from different manufacturers (factories). There were differences between Technicolor and Firepix (clock speeds were different preventing Firepix to work with DLA controllers) but the failure rates were common despite different sources.
Maybe there is misinformation here as we try to figure out the behavior and cause (but not necessarily the blame)... workarounds with voltage and intensity were ineffective. overheating and moisture intrusion might be part of the problem, but even that doesn't fully explain the issues.
I hope Timon's work will resolve the problems... but I don't know that it fully explains the behavior we've seen.
Steve
I got all my Technicolor (FirePix) strings from HolidayCoro, and I have experienced every single issue as described through out this thread.
There has not been one day that I was not fixing/replacing bad pixels.
My entire house outline, windows, eves, gutters, etc.... are these lights.....
I for one am very hesitant on using pixels next year as I don't want to go through this again.
Here is a video of our display.
You will notice at the 2:09 minute mark that the right side of my gutter freezes up full white for a second or two, through out the video.
Two more video with all pixels actually working
Which brings up the question, there are obviously multiple chinese vendors for pixels. Are there any taiwanese manufacturers? Strange that they are now the higher quality, or even japanese manufacturers?
Who else besides Ray, would be a source for higher end pixels like a minleon quality manufacturer, and are their other brands that are of that level of quality?
Alan
timon wrote:Ray didn't have the date code but the bad ones are marked WS2811B. He's purged them and is using the WS2811S.
I've opened up about 10 of my bad technicolor pixels and all are marked ws2811s.
Interesting, I'll check and make sure I didn't get that backwards. BTW, what did it look like? Could you see any visible reason why it failed?
I have one with water damage but on the others I can't see anything. I'll get some pictures of them.
Steve Gase wrote:
Ray Wu Technicolor pixels had problems -- but I heard that HolidayCoro Firepix pixels ALSO had the same problems. I understood each to come from different manufacturers (factories). There were differences between Technicolor and Firepix (clock speeds were different preventing Firepix to work with DLA controllers)
Is there a source for this? I thought both used the 2811 protocol at 800Kbps. (The WS2811 can be configured for 400Kbps as well, but I don't think either Technicolor or Firepix use that speed). But I could be wrong, and I'd like to know.
Timon..I sent a BUNCH to Steve..he would have gotten them Monday the 23rd..probably hasn't had time to mess with them yet

