Flickering WS2812b near end of strip

[aplant92]

Hi there,

I've just put my display up, and noticed weird behavior on one of my WS2812b strips. Basically I have 2 5m 240 pixel strips, both connected to each other. There is 5v coming in at the start, in the middle, and at the end.

When I have all 3 points of voltage connected, the first strip works fine, but the second one works intermittently, during both FPP testing and running a sequence. Sometimes it works, other times it cuts out completely, other times it’s flickering.

However, if I remove the middle power injection, FPP display testing shows it all working as I’d expect (both strips lit up and moving), yet when running a sequence they appear extremely dim (can barely see them).

I’ve replaced the t injector in the middle, moved the power injector cables to different pigtails, and also resoldered the connections on the strips. Also ensured all wires leaving both ends of both strips are connected to the correct 5v/data/ground connections.

I don’t understand how the one strip acts all weird when the middle power injection is connected, but also how when the middle injection isn’t present, the display testing works yet not the sequence.

I should mention I only have one power supply and one F16V2-R driving all this.

Any ideas? I’m scratching my head!!!

 

[AAH]

You don't have the brightness turned down with Xlights or Falcon and that is causing the discrepancy?

 

[aplant92]

If I’m using the one power supply, is there any sort of common grounding I need to do between multiple of the power injection boards from David Pitts? I would assume because they’re all on the same PSU they already have a common ground?

Only reason I say that is because the behaviour looks very similar to a grounding type problem.

 

[aplant92]

You don't have the brightness turned down with Xlights or Falcon and that is causing the discrepancy?

Nope! I’ve triple checked all the XLights settings and can’t see anything differently between the two strips!

 

[aplant92]

Also shouldn't the lights work fine with just the power and data at the start, and power injection at the end? I do still have the T injector in the middle (just with no power cable coming in), so surely both strips are still connected and should receive data as normal?

 

[aplant92]

Also if it helps - I have noticed a clicking noise connecting the middle power injection cable (like a short). Upon checking the connections, I noticed the voltage I was getting from my multi meter was around 4.6-4.7V. Not sure if this is a pointer to something?

 

[aplant92]

OK - progress. This morning I decided to just unplug the t injector in the middle altogether, and connect the two strips directly to each other.

Strips now working properly (I think)! I’m wondering now if one of the pigtail extension cables I got from Ray Wu is wired differently in the run to that middle power injection point. I already replaced the T injector, so would think that isn’t the problem.

 

[cdbrown]

Perhaps the data, V or ground is swapped around and causing the issue although it's strange that it's intermittent. If issues are present it's always best to test each item by itself to eliminate the element. Then as you begin connecting elements and power injection makes it easier to identify the source. Next is to run that whole strip on full white to ensure all the leds are nice and white - won't need to inject in the middle.

I do still have the T injector in the middle (just with no power cable coming in), so surely both strips are still connected and should receive data as normal?

as long as there isn't an issue with the T piece. Does sound as though the data wire in the T piece is perhaps broken somewhere in there. Is it one of these https://www.aliexpress.com/store/pr...215.html?spm=2114.12010612.0.0.32c49c0cxmXEya ?

 

[aplant92]

Yeah doesn’t make sense to me either - the fact it was working during FPP testing but not a sequence was just strange.

And nope, it’s the one with 3 core in the middle!

 

[marmalade]

I had an issue on monday with a node on one of my snowflakes. The others daisy chained to it would intermittently flicker at full brightness.leading to a wild goose chase for power problems. But each flake had power injection, which was puzzling!

Turned out that one node in the middle of snowflake one had a bit of resistance across the in/out grounds which would corrupt the data coming out of it depending on the current draw. Took 4 hours to find (including pulling them all off the wall).. and 5 mins to replace.

 

[aplant92]

Isn’t that always the case!

Thanks for the tip though, if this rears it’s ugly head again, I’ll keep it in mind