Corrupt Pixel Data?

[mbl111]

Hey guys, I seem to be having some trouble with a certain run of 12v bullet pixels in my display. When I put them into a test pattern from my PixLite 16 controller, whenever they go red the green will flash on and off strobing the lights between Red and Yellow until the controller sets them to green and they behave.

Tonight I'm probably going to do some troubleshooting but I was wondering if anyone had encountered this issue before and had some insight before I wasted a bunch of time pulling things apart.

Thanks in advance!

 

[Nuzzy]

Possibly a voltage problem, how far is the data run and do you have any power injection points?

 

[mbl111]

From my controller to the first pixel I have a 5m data run, but its boosted by a Hanson Electronics Null Pixel booster and has worked fine in all other testing.

From there, I have approx 480 pixels, 4 10w WS2811 flood lights with 4 power injection points evenly spaced across the run. I ran the numbers through that online voltage drop calculator and everything seemed fine, however I'm going to check all that with a multimeter tonight.

Could the signal be degrading with such a long run even though it's through pixels?
Is there any kind of signal reflection issues that I should be terminating the line to fix?

 

[Nuzzy]

I would say that’s not enough power injection unfortunately. Any distance between the pixels and the flood lights or between any group of pixels? If it’s a continuous run then the signal will stay good, once you go more than 3 or so metres from the last pixel to the next the signal will drop dramatically and corrupt.

The power injection for nodes should be every 50 or 75 but every 100 at minimum. You could try turning down the brightness to around 60% in the controller to see if that improves things without extra power injection.

The 4 flood lights should only need one power injection between the second and third light but it is dependent on how long the cable run is.

My spiral trees are strips, there is 50 pixels per tree and about 4 metres of cable between them, I have to power inject at each strip and also have a null pixel between each tree to have them anything over 50% brightness.

These lights can be very fussy sometimes.

 

[Martin Mueller]

I see many people trying to limit the number of power injection points. It does no harm to inject more often. You do run a risk of unstable operation if you do not inject often enough.

As for the null pixel driver, That driver helps for the single hop in which it is installed. Every pixel regenerates the signal and drives the next pixel in the run. Depending on the quality of your signal wire, you should get 3-4M between any two pixels. However, that assume the driving pixel is working at full designed capabilities. Pixels do not all work at the same level. Some can drive further than others and some have a problem driving as little as 5cm. You may just have a single pixel that is sensitive to the voltage drop and cannot drive to the next pixel.

 

[djgra79]

Yeah no harm in adding in more power injection. That will be the test really, add in another power injection point and see if there are different results.
Also, you can try reducing the model brightness in xlights or at the controller level output. Another quick test is to do a ramp up and down of white and if they don't flicker the entire time whilst ramping and only when it's at it's brightest, then that would indicate lack of power.

 

[mbl111]

Just wanted to update and say that I've solved the problem. Turns out I'd somehow fried the chip in my Pixlite 16 Reciever. Quick swap out and everything was up and running again