Ws2805 24v: one pixel flickering, before and after are fine, replaced, still flickering

Issues
1

Im completely confused here, mostly because the problematic section was replaced, yet the problem persists.

136 pixels (about 14 meters), it runs in a loop and I have power going into both the beginning and the end. It’s a Meanwell PSU, at max its pulling about 75w from the wall, it’s a 240w psu, so I really don’t suspect power at this point. It also does this regardless of color/brightness. Changing it from 42 to 15fps seems to lower the intensity, but it’s still relentless.

The offending pixel is the second from last one on the first strip (pixel 48). The almost 2 strips after it are perfect, and the rest before it are also fine. No artifacts, no flashes, no flickers, effects look great. They’re all perfect. Except that one pixel.

I soldered a new section in to replace it, no difference.

When I first turn the power on, the first strip lights to (and including) the bad pixel, and about a quarter second later, the rest after that strip turn on, and the flickering begins.

Tried setting it as the starting pixel, still flickers. Set it as the last pixel, no flicker. I can turn on about 3-4 more after that and it won’t flicker… I can max out the brightness on a single color, then turn off everything after pixel 48, and it is fine. No flickering at all. But that obviously means no effects…

Should i be replacing the section before or after it or such? Some other issue? My arms hurt, soldering on the ceiling is painful, so I’m trying to avoid a “parts canon”. This one has me very confused.

I’ll include that I’m using a Gledopto 015-D running WLED 0.15. 16 gauge wiring for data+power. Length was about a meter from controller to light, but i read that could factor in, so I cut the data line to about half a meter. No change. Multimeter is showing 26.5v into the bad pixel, 26.5v out. DIN is showing 0.3v in, 0v out (again… This was what it showed before I replaced it as well), BIN 0.3v in and out…

Is my luck just that awful that i replaced a bad section with another bad section? Possibly (my luck is not great), but before I torture my arms more and solder on the ceiling again (on a step ladder on a bed… It’s not fun), is it possible I’m just completely missing something else here?

Thanks for your time, sorry if there was any rambling. It’s been a long night and I’m very sore -_-

2

Gledopto is known for such issues. you can try adding an external levelshifter, reduce the length of data line or lower the value of the resistor on the “levelshifter” they use

1 Like
3

I am not trying to come off as challenging, i’m sincerely asking to learn and understand. I am aware my blunt direct inquiries can come off this way, but it’s not intentional… And trust me, i really really wouldnt post if it was something i could find via search, as im not very social and very much prefer to find solutions instead of asking for them. But it’s this single middle pixel aspect that has me unable to find a similar scenario to compare mine to.

My initial clarification question is, would this apply to a single pixel sorta near the middle? My searching has lead me to a few paths, and they seem at odds with each other. One being the data line can be an issue with a long data line, but those seem to be more than one pixel, and the first IC effectively “repeats” the signal once it’s received. Which is where I’m confused, 47 pixels up until that one work fine, and every single one after that… If it’s poor data signal from the controller, why isnt it showing up in any other pixel? Why just that single one?

Snipping the data wire to half the length made no difference. A different Gledopto controller also made no difference either, and while it’s still Gledopto, i would suspect it’d at least shift the problem (slight manufacturing variations and tolerances being what they are), but it is persistently that one pixel.

I did mess with the wiring a little more, as it seems theres an opinion split in where to connect BIN. I had it bonded with DIN initially and changed it to ground, which did stop it from doing that quarter second delay thing on power-up, the entire strip illuminates immediately instead of half and half… Which is the very first time I’ve seen any change… But that one pixel is still an issue.

I am absolutely not saying “it’s not the controller”, just trying to understand the underlying cause responsible for this persistently peculiar oddity. Obviously i want to solve the issue, but i also want to learn how/why it’s this particular behavior so in the future i can figure it out logically instead of needing my hand held every time.

4

it is odd indeed. but there are many reasons (as you probably found out already) for stuff not going right.
if its always exactly the same pixel even with a different controller, its more likely an issue with power or wiring in general or it could just be that somthing kills that exact same pixel when you re-wired it. these are just reasons from the top of my head. if you want to get to the bottom of it, you need to measure voltages with a multimeter and even the data signal with a scope. you cannot debug electronics without proper tools or its like trying to count bacteria without a microscope.

5

Multimeter makes me think it could be that section. I certainly wish i had a scope at this point.

Thank you for the response. Wouldnt surprise me if my luck is just that bad. I’ll swap it out again and report back.

6

I just replaced the last 3 pixels on that strip to rule out any possible problems immediately around the bad pixel, and i don’t know if my luck was that bad of if a bad line immediately before/after was the problem, but all good.

Thanks again for your help!

7

Yeah my guess would be the pixel right before the first one that is flickering. Worth checking at least.

8

^^^^ Yep. Always best to replace the last working and first nonworking. The last working could have an issue with it’s output to the one that shows the issue.

9

Even though it was a relatively simple solution, i still learned from it. It could’ve been the one after effectively pulling DIN to short and making the flickering strip flip out, which would explain why din showed 0 at input, if the IC detected the short it might’ve just been trying over and over to push it and failing.

Another interesting symptom that i forgot to mention was each power cycle, sometimes it was 135 pixels, sometimes 136. So one of them seemed to be disappearing (oddly, they were all still working and animating regardless, but at 135, the last one wouldnt light, at at 136, sometimes it was one beyond)

Definitely a learning experience, but i enjoy learning :slight_smile:

Thanks everyone for the helpful info. Great community here!