Here’s a funny one. Any idea what might be causing this? Every few fades, the last 25 LEDs in the string glitch out.
Interrupts turned off, on an RFduino. Replaced the remainder of the string at the busted point (a couple of times!) and still get the same result. Code is on Pastebin (http://pastebin.com/6upfBLxr). FirstLight works fine.
code looks clean…run a serial.println() each time you run fill() and watch it. weird though
I can’t see it clearly on the video but could it be that the glitch happens during color white ? -> power?
Not white, in particular—it’s picking a random hue each time and fading up a different colour. Not sure if it’s consistent. Will try all-White and eliminate power as an issue.
You need to do:
#define FASTLED_ALLOW_INTERRUPTS 0
before
the #include <FastLED.h> - otherwise it has no effect at all.
Well then, derp of the day award goes to me :-).
But even so: turning interrupts off with a full-white fade gives me a funny flicker on increasingly more of the LEDs on the down-fade as the thing warms up. Setting setBrightness(128) with interrupts off gives me a better result in the colour mode with some flicker on the way up, while setBrightness(128) with interrupts on gives me basically the same result I had originally, with the last 25 LEDs in the string going a different (usually dimmer) colour about a third of the time.
Could be a combination?
I also often want to massively over spec power. In part because I’ve seen a number of power supplies get wonky when they get above say, 80% draw, give or take. For example, the closer you get to the amp limit, the more of a voltage drop you start to see, and well, we know what sorts of problems happen with these leds and their data handling when that happens 
So is this a power issue?
Any chance you’re using pointer arithmetic and writing bad data to the LED array?
Done that one a few times…
@Luminous_Elements check his code: only random8 and sin8 and the FirstLight example. In void glow() {
the i++; is defined as uint8_t, so all variables should be in the range of 0 … 255
I’m still betting on power or cable cross-talk.
Yeah @Daniel_Garcia as do I, this is basically a test rig.
