I have 120 24v 7 LED pixels controlled by a Pixelblaze. In testing everthing worked fine but now I am having signal propagation issues. The signal seems to go for ~25 pixels and then the test pattern built into the LEDs takes over.
Thought everything was set in testing but now on site at install, I am having issues. The only difference is a 1m long extension cable in-between every 12 pixels, but I wouldn’t think that would make it stop so soon if it was the issue.
The PB signal doesn’t have to reach the last pixel under its own power right? Each pixel should repeat the signal right? Or is that not always the case?
I am going through double checking connections but wondered if an SPI signal booster right after the PB would help or not?
Unfortunately this is common with longer runs before first pixel or between pixels. I assume you tried swapping segments if possible to see if there’s a bad pixel not passing D1 AND D2?
In theory each GS8206 chip regenerates the data signal, but in practice you can see all kind of weird artifacts with a scope like noise building on the rails, impedance effects (scope shows overshoot or smooth edges on square data the further you get from the controller), or reflections (look for superimposed / additive patterns on data).
If it’s power rail noise, you can try adding large capacitors across +24 and ground, spaced along your total length.
If its the other two, try putting your SPI “signal booster” (buffer op amp) inline closer to where the signal starts to degrade.
You can also try soldering the extensions if you’re using connectors - the connectors always add a little inline resistance.
And if none of these work, the two remaining solutions are to distribute more Pixelblaze (shorter channels) in Sync mode, or use an output expander to make a star topology, then use RS485 balanced transceivers on either end of the long runs to each starting pixel of the segments you’ve subdivided into.
Unfortunately I think you’ll find the same issues with any controller - it’s usually a feature of the physical led wiring.
Replacing some of the extensions with pixels worked to keep the data going. I also had a couple bad pixels that I replaced.
Now the problem seems to be too much power injection. If I disconnect the power injection and add more pixels, the signal keeps going, but when the power injection T-connector is attached, the pixels around that section don’t get the signal.
Seems that if the voltage drops to around 20v then the signal goes through but 21-24v the signal is overidden by the test pattern.
I am going to try to see how far I can go in between power injection but it seems that once a single Pixel has full voltage, then its internal test signal is too strong.
Interesting. Out of curiosity how many total 6-pucks are you running off one controller or channel, and how often were you power injecting with those (roughly… in terms of ft)?
4 sections of 103 to 132 pucks per section. I was power injecting ~ 20-24 pucks.
Since every time I removed a T-connector, the signal went farther, I removed all the T’s as a test. It made it to about pixel 125 before showing flicker. I added power injection at the end of the 132 puck section. This made the last two pixels go full white, so I added 2 pixels to the end and taped over them. It seems it takes 2 pucks before the test pattern stops conflicting after power injection.
Good news is that so far it is working with power only at the beginning and at the end. Even works at 100 brightness although I am not setting all the pixels to white at once.
Wifi Pixelblaze followers working too.
Wish I knew how to power inject and not have signal issues since I ran all the cables. Next time I will avoid pucks with a built-in test pattern.
With power injection you want to make sure to have a solid connection for both positive and negative (GND) sides of power. Negative (GND) will always be connected between the LED power zones (before and after the injection point) or you will definitely have increased data issues. If you are using the same power supply, you can keep positive connected as well.
It was strange behavior. All injection is from the same supply using factory made t-connectors with solid positive and ground. Tested with multimeter as well.
It works with the injection just at the end of the relatively short runs of 130 pucks since it is 24v.
The Pixelblaze Wifi follower and mapping setup works great. 4 PBs running, each with part of the full 3D map along with “dummy” pixels for scale. It is really responsive and works great.
