Flickering SK9822

Hey guys,

New to Pixelblaze, I guess technically LEDs in general. I’ve got 10 meters of SK9822 set in 1m channels that connect with 4 PIN 22 AWG wires. A new power drop is provided at 5 meters, and only ground and data is connected at the power drop point. I’m getting flickering that’s occurring in the set where the new power drop is provided (data is running from left on the bottom back to right on top). I’m not sure what’s causing the flickering, I’ve dropped the frequency which seems to help on some patterns and other patterns it doesn’t seem to be a problem. Not sure where to start with troubleshooting.

Example of the problem: https://youtu.be/kb9LurY54sc

I’m converting my entire setup to run on SK9822 LEDs, I was using a bunch of lights provided by Govee before set in the same channels and arranged in a matrix that looks like this (https://youtu.be/jkdN5YBQZJ4) when setup to give you an idea of what I’m doing. The channel strips/sticks are mobile and are set in vary levels of diffusion to give those a different effect and to be able to change how they are arranged.

Flickering randomly happens when a signal is close to the threshold - where its not clear between a 0 and a 1, and noise causes it to jump between them. Sometimes that happens when a signal is run too fast, or when a clocked protocol like the SK9822 gets out of phase slightly with the data, or it can happen because the data/clock signal voltage is too low or too high.

I wonder if there’s enough voltage drop at the tail of the strip that the outputs aren’t close to 5v, seems to flicker more when more LEDs are on. With a multimeter, you can set it to a static pattern and measure the voltage on both sides of the jump.

SK9822 need a 5v data signal, and normally thats what you’d get, but if the output chained to the other power zone is itself far away from its power source, the voltage will drop below 5V over the strip and can be low near the end. Sometimes that shows up visually as whites shift toward red, and blue/green drops out. If that is the case, to fix you can run additional power lines to the end where the LEDs cross into the other power zone.

It also can’t hurt to throw some capacitors at the end of long power runs to help smooth out power.

You mention connecting ground between the zones, and thats good, though if it were loose or a weak connection, it could explain flickering as well.

Some apa102 and clones experience clock drift over some number of pixels, and have to be clocked slower when there are a lot of pixels in series.

You mentioned changing data speeds helps, what values are you using? Does changing the brightness help? Even at higher rates?

1 Like

Was running at 1 MHz. Dropping the brightness seemed to make things improve. I’ve identified a few bad solder joints with the multimeter, although in that process I managed to short something and damaged a pixel (Doh!).

In terms of capacitors, should these only be connected to the 5v rail, or on all the data lines as well?

Capacitors help smooth power, but would disrupt the data line. Many capacitors large enough (100uF+) will be polar, meaning its important to get the positive and negative connected in the right way.

You’d connect the negative lead to the negative side (GND) of power and the positive lead to the +5v side.