Have you ever noticed,

Archive FastLED Archive
1

Have you ever noticed, that in a WS2812B setup you have from time to time (every some hours) a white flash of one or more leds for approximately one frame?
It seems to be completely independent from the running code and used controller board.
I did not change the power supply and the wiring in the my setup yet. The data cable is shorter than 30 cm an unshielded.
A friend of mine has witnessed the effect too, so it is not about my perception of reality.
Any thoughts?

2

One thing we found today was. . . have you tried using a different clock/data pin?

3

I have seen this happen often when using 3.3v MCU’s and 5v LEDs. The only reliable way I have found to stop this from happening is to make sure that I have 74HCT245’s between the MCU and the LEDs. I’ve seen it happen far far less often with 5v arduinos. Power spike on the data line causing the next frame of data to be read shifted some bits off? (Or power drop causing the WS2812’s to interpret everything as 1s?)

4

I lied, the second reliable way to stop this from happening is to use APA102’s :slight_smile:

5

@Andrew_Tuline what we found today was a far more limited situation - namely, you had to be using the power management code and using pin 13 for your led data :slight_smile: (though, that is now fixed on master)

6

I’m pretty well done with the WS2812’s and just ordered some APA104’s. They’re not a whole lot more expensive, and will solve several problems.

7

The APA104 is pretty much the same as the WS2812 - same data protocol, everything - aka 3-wire led chipset with tight timing requirements and a slow data protocol

The APA102 is a 4-wire led chipset with no timing requirements and can be driven at speeds from 12-20Mhz :slight_smile:

8

Good to know that I am not alone.
Yes I tried different pins and with the teensys I use always a level shifter. Hopefully the APAs make it to Europe soon.

9

I get it too, roll on my apa102’s :slight_smile:

10

Oh weird, I haven’t seen that happen on APA102’s at all. Does it happen with any kind of frequency? Also - what are you driving the APA102’s off of? (Trying to decide if this is a glitch in the chip, a glitch in the line, a glitch in power, a glitch in the code, a glitch in the MCU’s memory, a glitch on the data line, etc… etc… etc…)

11

Lol Noooooo I meant can’t wait for my apa’s lol been soldering up 30m of 2812’s today in to 50cm strips! :confused: I now have hexagons in front of my eyes lol

12

@Daniel_Garcia Oh crap. Gotta cancel those 104’s. . .

13

I used to have this very same problem on a set of LPD6803 and WS2812B. Turned out to be a matter of (not enough) decoupling on the power lines and/or the led’s. These led’s cause a huge spike when switched on/off and that creates great current rushes. These rushes cause the data to the led’s to get corrupted, especially on longer cables. After adding more decoupling all these flashes were gone. Same hard-/software setup. Decoupling is your friend.