Just ran into an issue with a SP511e build (v0.13.1-bl4) with dual outputs to SK6812RGBW strips.
Strip 1 seems to consume 4 bytes per LED, but strip2 uses 20 bytes per LED?
My ESP32 instances show a small amount of overhead added for each new strip, but then the increment is the same for each new LED added.
I’ve got two strips of 225 LED’s each and a single strip of 450 is fine, but splitting the strip and adding a 2nd output gives me major warnings and mayhem.
I understand the 8285 chip has less RAM but the usage across two pins seems a bit high?
A little further research…
This is even stranger, the issue seems to be linked to the GPIO pin I use.
If I pick 3 (the default), I get large memory usage, anything else (I’m using 15 typically) it seems to be fine.
I hadn’t noticed as I’ve moved off of pin 3 for most of my SP511’s.
I tried on at least 2 different boards.
Aha, that leads to an interesting question - what NeoPixelBus Method is getting used when I set my GPIO to 15?
From what I read in Makuna’s Wiki it should be defaulting to BitBang no?
Which of course has dire warnings an flashing lights about instability and adverse effects when you try more than 3 pixels…
Good thing I didn’t read any of that before I tested it on a 215 LED SK6812RGBW strip I have setup - works like a charm so far at least with internal effects.
GPIO15 is bit-bangged.
My testing showed that bit-bangging is ok if ESP is not doing any other thing.
Unfortunately this is not the case with WLED so you may get dropped frames or occasional reboot.