So i have a better video of my problem with the stripes.

Archive FastLED Archive
,
1

So i have a better video of my problem with the stripes. Anyone a idea what causes this?

4beae9e33c07481b3581214c657c3ccb.jpeg
4beae9e33c07481b3581214c657c3ccb.jpeg1920×1080 373 KB

2

Faulty LEDs, or faulty solder joints. Cut the affected LED out of the strip and solder the strip back together. Normally caused by excessive heat or poor manufacturing.

3

any signs of damage on the leds you press ?

4

No, no damage they are glued. Yes in the video you can see that i did allready several cut outs but they keep apearing :frowning:

5

Hi @Severin_N ,

I reported a very similar problem a while back. See this old post…

https://plus.google.com/106626345342202981932/posts/TqWBtiHJHjX?sfc=false

At the time, I concluded the problem was soldering solid wires on the back of them little PCBs and possibly heating the WS2812b devices excessively. But as you are seeing the same behavior on strips, I would change my conclusion to weak WS2812b manufacturing.

Last year I bought my first strips of WS2812b (had always used the little PCB style or 8mm RGB LEDs with WS2811) and drove it to white full brightness for a while (no more than about 15 minutes or so) and started losing one of the 3 colors on some pixels. Letting them cool down recovered all pixels and I now never get any WS2812b to run ‘full brightness white’ for any extended period of time.

6

@JP_Roy Hi welcome back ;)… i tried to add you to a other post but somehow google+ did’nt let me. (You where the guy who told me to try a teensy with octows shield and ws2812 at the first place :wink: close to a year ago.)
Thanks for that by the way, i think it was a great advice (except for the problem with the stripes for now.)

I have now a setup with the sloeber ide (based on eclipse) but still some troubles with the c++ for now. But i think i will overcome them. The good thing about a large project is that i’m beyond the point where stopping is a option :wink:

So mine where running at full brightness but not on white. mostly they did rainbow or the basic collor change from the octows.

But i have seen the recovery effect also on the tabletop. As you see in the video i was searching for the broken led wich was working again.

Did i understand right. your conclusion is now bad manufactoring? Or is it never run leds at full brightness?

7

Hi again @Severin_N ,

Both actually…

  1. The actual problem (I think) is WS2812b manufacturing/design. I find they simply do not like being driven at full brightness for more then a few minutes. Of course full brightness white is the worst.

  2. The solution: Don’t drive WS2812b devices at full brightness for more then a few minutes. This is an OK solution as I find full brightness to be excessive in most cases anyways.

8

@JP_Roy ​So how much did you reduce the brightness to get them stable? 80%?

9

@Severin_N I set MAX_BRIGHNESS to 128 which is 50%. I find that it is bright enough in all my builds and have not got a failure since. It is actually still too bright as I also read a potentiometer that dims the LEDs even more dependant on the animation. You need to experiment with it yourself but I would suggest to start with that level and see if you need more brightness for your setup.

10

It’s a combination of bad manufacturing, which can sometimes be exacerbated by running at full brightness. Also - while the strips are flexible, you have to be careful how much and how repeatedly you flex them, as that can put stress on the solder joints and make them loose : / It’s a really really really annoying set of things to deal with/have to work around.

11

@JP_Roy so i run the leds for several hours with brightness set to 150 and there didn’t appear new broken leds. So the problem seems to be located as you said.