Hey,
I posted about this problem some months ago but have since stagnated my work on it. I really really want this LED setup to work properly so lets give it another go!
As you can see in the video after a while the led-strip freezes it’s animation beginning with the leftmost LEDs. When i turn it off and on again it’s still frozen. If I wait for a while before turning it on again it’s working again for a while.
What could cause this? Is it maybe cause by some overheat protection in the LEDs?
Might add that it’s a WS2812B strip
What can we say without seeing the code? Do the examples from the library cause the same effect or is it just your own code? If you can reproduce the error with any code I´d check the voltage of the leds. There is no overheat protection but if the voltage is too high they heat up until they stop working (sometimes forever).
Without source code you’ll probably not find much help. 
You probably have a bad pixel in the strip. I think I have a similar problem with a strip too… one of the pixels in my strip will jitter colors when it acts up
What’s the current rating for your power source?
Commenting because I’m curious to learn what this turns out to be.
I had similar problems when I use the 5V 500mA out of an Arduino. When too much power is drawn the animation just stops. @Ira_T has a good point: Check the power source and monitor the voltage.
I find that about 2 in a thousand or so WS2812 pixels have issues where they stop working after a shortish period of being on, especially at full white.
Try swapping out the pixels at the point where the lights stop behaving properly. If not then as people above have suggested then check your power. A good test is a slow white fadeup where if it stops working as it gets brighter it’s a very strong indicator that there’s a power issue.
I test every single project I do now with a full white multi-hour “burn-in” where i swap between a full white and patterns to check that no LEDs break the signal chain when warmed up, for the largest projects then at least one usually does this.
i don’t know how you have those strips chained? which end is D_IN? if its starting left to right the whole strip should have frozen when the first pixel failed… it looks more of a programmatic issue the way it progresses down the chain…but I’m a newbie so my say is only fluff…
@Pete_Brittain I like the full white test period idea. I’m working on putting lights on the exterior of my house. A few hours of testing now could save me a lot of trips up a ladder later!
I have exactly the same problem. In my installation the problem appears quite often when I go for 100% white. So I thought it’s a thermic problem.
I limited the total brightness to about 40% which is enaugh for me.
But the problem still appears on about 2 meters of 70meter WS2812B 144LEd/m.
I changed the problematic stripes and it semeed to work, but now another stripe has the same problem.
I’ll also mesure my power source.
What could be the problem about it? Is there a minimum Voltage for the Led’s to forward the information? All my LED’s light up white correctly, so I thought the power is allright.
Hi again, and thanks for all the replies.
I tried running with FastLED example code and even with a different microcontroller (Arduino NG instead of a RFduino). This setup had the same issues. I think i can rule out a code related issue.
@Ira_T It’s a dual channel (25 + 25 Watt) 5V supply. I’m only using one channel though. Can I safely connect the two channels in the same strip (to use the full power rating)?
@Jon_Bruno The D_IN is coming from right to left
@Pete_Brittain Thanks for the info! So they can really lose contact when warmed up? Just how sensitive are the strips? Is there a big risk causing more of these problems if the strip is slightly bent when handling?
I’m gonna measure the voltage over the ground and power line at different places in the strip and see if there is any major voltage drop. I think you’re on to something Birk. I should try it with a lower brightness.
I’ll do some measures and test this weekend at my installation as well and will post the results.
Hope we find the problem together!
Cheers,
Birk.
@Fredrik_Andersson you can use more than one power supply.
@Juergen_Bruegl Wow, never realized they drew that much power. That could be a reason.
I’m trying now with brightness set to 25 and hasn’t seen the problem occur yet.
I find it strange though, since I could previously run this setup for a longer period with full brightness. Then I had two led’s that was clearly loose since they passed through the data once pushed in. I patched together the strip with new led’s and a while after that I instead had this problem…
I’m thinking it might be cause of very tiny dissconnects when the temperature changes that started occuring after I pulled out the strip and maybe handled it uncarefully.
Measuring the voltage at full brightness gives me 4.7V at the end of the strip. At 200 brightness i get 4.9V. That should be enough for the WS2812. Although it’s definitely alarming that the woltage level drops at all.
@Birk_Schmithusen Did you find time to do some measures?
I just remembered that I had the strip running with a 3,8 amp power source without any of these problems (only effect was that the leds faded towards the end of the strip).
Although it could be that the new power starts to oscillate when it reaches it’s max rated effect, causing the LED logics to fail but with the LED’s still lit up.
Once I get home I’ll try it with only the first meter lit with highest brightness to rule out if it’s heating of power related.
I just did a test where I reduced the amounts of LEDs instead of the brightness. Even down at 25 LEDs the problem persisted.
Does this mean it’s likely heating related?
I did some measurements and had the same values as you have: about 4.8 volt at the end of the strip.
I connected the end to my power source and it changed nothing.
I talked to the guy who sold me the strip and he made some heat measurements with a thermo camera.
The tech specs allow a temperature up to 70°C. After one minute of full white he got 67°C on the outside of the LED. So probably inside it was already over 70°C.
I have 144 led/m in a plastic tube, which makes the effect worse. I’m pretty sure now, that it’s her related and that the stripes are not produced for full power white without extra cooling.
Cheers, Birk.
