My Scenario:
- 500 fairy lights controlled by an ESP-32 WROOM running WLED 0.15.1.
- Short distance to first LED (less than 2 ft), no data signal boosting, no resistors on data line.
- Power BTF-150 30A power supply 5V.
- Voltage injections at LED 300 and 400.
When effect is set at solid orange, or any other bright color everything was fine.
Animated effects were also good as long as none of the colors were black.
If an effect had large areas of black (or dark) color, those regions would sparkle red,
green, and blue. If I set the color to solid black the sparkling was bad enough
to trigger epilepsy. When I turned the power off using WLED, many of the
LED lights stayed on.
My solution was to cut off the first LED and make the connections to the second
LED (which was now the first). This made it work perfectly.
I had tried data resistors, I2C buck convertor, more voltage injection, different
power supplies all to no avail. I also tried a Dig Uno controller. I measured good
voltages throughout the string and had to conclude my problem was
data transmission somewhere in the LED string. Fortunately, the very first LED
seems to have been the problem. I did try to use WLED to skip the first LED but
that had no effect. I suspect that I abused the first LED with high voltage or
reverse polarity at some time, but not enough to make it fail completely.
Anyway, this worked for me and I hope someone else is able to benefit from
my experience if all else fails.