Technicolor quality...
 
Notifications
Clear all

Technicolor quality?

256 Posts
45 Users
0 Reactions
83.4 K Views
jklingert
Posts: 1767
(@jklingert)
Noble Member
Joined: 16 years ago

To be honest, it sounds like it was not really just a bad batch of ICs. There were several different ICs listed in this thread that failed, and there were no waves of other WS2811 IC failures reported in other pixels, just the Technicolor pixels. A lot of the other failures that happened this year in other pixels were people who used silicone molded pixels, and they are known to be problematic with water intrusion.

It is more likely that there were a number of failure modes for the Technicolor pixels and the fact that they were a new design with a new molding might also point to mechanical failure and not just failure of the electronics. Folks should look at what is different about these pixels to find for reasons they failed. Zach posted on Facebook that there were issues with the molding ripping the pads and wires off the pcbs and that there were manufacturing changes midstream. The different heavy molding compound, the compressive stresses from the lens and locking ring are unique to these pixels and may help lead to some of the failures.

Until someone really does reverse engineering on a handful of the bad pixels, it will be hard to tell what really took them out. Lot's of speculation and comments from a vendor who is likely facing tens of thousands of dollars (or more!) in bad product shipped. The Technicolor group buy was almost 500 strings alone! I'm not looking to place blame, just help people learn what to lookout for in the future. Pixels are a massive expense and no one wants to spend thousands on pixels to have them fail.

Counterfeit, bad clone or recycled ICs are an issue that all electronic industries are facing with parts out of China now. http://spectrum.ieee.org/computing/hardware/counterfeit-chips-on-the-rise or http://spectrum.ieee.org/semiconductors/processors/the-hidden-dangers-of-chopshop-electronics Even the US Military has been burned. Take a look at how many clone pixel ICs have popped up on the market, it is possible that Ray's suppliers may also have received or used bad parts. No one has shared a circuit diagram of the pcb, it sounds like there were missing resistors that the datasheet for the WS2811 called for, hopefully there were no other electrical design issues. Still not sure why they are moving the wires around on the new design, was there a manufacturing issue with the wiring shorting out the 12v and the data line?

The quality of pixels have shot up over the last three years and most folks had very good success this year with other pixel designs. The first wave of silicone pixels 3 years ago were a complete disaster. They have gotten better, but they still tend to be more problematic.

Hopefully Timon and others can help sort out the underlying issues so that the quality of the pixels can all go up next year. This is a learning exercise for the community, just like most folks learned about the importance of waterproofing 3 years ago.

In the meanwhile, caveat emptor on any non epoxy pixel design...


Reply
dsmjuggalo
Posts: 663
(@dsmjuggalo)
Prominent Member
Joined: 15 years ago

Here's a couple pics in my phone.


Reply
Posts: 23
(@hauvega)
Eminent Member
Joined: 18 years ago

Here are some pictures of my tc failures.

1210 13
1210 10
1217 8

I have some more shots on my photobucket account. The link is http://s577.photobucket.com/user/Hauvega/slideshow/Pixel%20problems


Reply
MindTrik
Posts: 1542
Topic starter
(@mindtrik)
Noble Member
Joined: 16 years ago

did you figure out exactly what the failure causes were?...are yours the only ones to melt/catch fire?


Reply
Posts: 23
(@hauvega)
Eminent Member
Joined: 18 years ago

Mine did not melt. I figure mine had moisture inside. I had some with injection molding issues where the pcb was forced out of position. I did clean most of them up and coated them. I replaced about 10 ics on others that quit receiving and transmitting data. The 12v leaked over to the data line and killed some of the chips(low resistance on data to gnd measurements).


Reply
Page 29 / 52
Share: