WS2811 First LED failure

Status
Not open for further replies.
E

emkay

Member
Hello everyone. Thanks to a lot of information here, I built a fairly large custom-shape LED matrix. Using Teensy and the Octo. Everything works great. Mostly. Unfortunately, I’ve had a few ws2812b’s die on me. The initial dead units I attributed to manufacturing defects. However, there were a few instances of the first LED in a section burning out. I’m not sure why.

So far I’ve tested/done:
- Voltage is no more than 5V
- the power supply is one that is regulated, smoothed, with over voltage protection
- I’ve scoped the power up/down behaviour for both the Vin and Data lines and they seem to look great. No surges like I initially assumed were the cause.

Things I’ve thought might be the problem (just to put all my suspicions out there :) )
- since it’s the first LED, it feels like it’s caused by initial signal or power delivery. All following thoughts pertain to either Vin or DIN
- initial voltage spike: though likely ruled out through scope testing as mentioned above
- initial current spike or inrush current: still have to to do some scope tests to investigate/learn about this (though don't currently have any current probes). It seems some recommend capacitors to prevent onrush current, but it seems unnecessary given the capacitance and filtering in the power supply.
- Is an inrush current limiter necessary along the way?
- Is the Octo Resistor adequate for the data line?
- Is there are current limiting issue at the LED?
- grounding issues? If ground is poor, will it try to find ground through DIN? Or something along those lines
- could there be an issue with data line getting power before Vin if the controller is powered up slightly ahead of time?

Anyone that has any thoughts would be much appreciated. In the end, these failure could also be coincidental or even an accidental short.
Thanks,

Marc
 
Thanks for the suggestions so far. I did see the heat failure thread during the many hours of web searching, unfortunately it doesn't seem to be heat related. Though I'll take a closer look at that now that you brought it up. Tricky thing is that this is so intermittent, I'm finding it hard to test for it. Trying to repeat things over and over again to cause it to reappear though. Did at least 100 startup/shutdowns to verify power supply output issues :p I will load up 1.2 as well Paul. Thanks for the input. I'll have a look at that other specific bug issue as while I'm at it. Unfortunately the display was made for someone else and am trying my best to make it as reliable as I can before handing it over.
 
I've seen references to the need for a current-limiting resister between the drive and the first LED. 300-500 ohms are the values I've seen. Look at Adafruit's documentation for their neopixel...
 
PaulStoffregen said:
Looks like I might have added a bug in the library recently. Maybe related to this?

https://forum.pjrc.com/threads/43574-OctoWS2811-Unexpected-Colors-(7-bits-per-channel-)

Can you give the old 1.2 version a try and let me know if that makes the problem disappear?

I'm planning to work on this issue later this week. It doesn't happen on my old LED test board. Going to wire up some newer strips and set up to do more rigorous tests with a scope.
Click to expand...

For what it's worth, I've been experiencing similar issues with new 12V strips and what seems to be dead or stuck 1'st pixels. I've been doing some testing on a strip that I happen to have with me, and what's interesting is when I remove power from the Teensy, the first pixel lights to what it was last set to. I'm hoping this detail might be of use to others...

(I'm not at the same location as my workbench right now, otherwise I'd do some testing with wire lengths and gauges...)

Hope this helps. I'll know in a few days if this is the case with the other strips that I've been having issues with.

--AD
 
Status
Not open for further replies.
Back
Top