What is your experience with the reliability of APA102 strips? The WS2813 pixels now have a spare data line so when a pixel dies, your strip still will keep working.
So I was wondering if this really solved a real world problem? I’d like to keep using APA102 strips because of Global Brightness Control and their speed. However, how often do those pixels break, making the whole strip useless? I never had that experience, but I’m curious who does (and who doesn’t as well).
I never really had many dead pixels happen with wa2812 - but with apa102, one project I did last year I had ~15 5m rolls of apa102 that I had to cut out and solder over dead pixels out of 20-24 rolls that I had.
What happens when an APA pixel dies? Does it stay off, or does it influence other pixels as well? Or are there different variations?
I’ve had lots of problems with ws2812s, better luck with SK2812s, I now use APA102s and SK9822s. These have both been very reliable 
@Kasper_Kamperman when I was having problems with the apa102’s, it stopped passing data through to the following LED’s, so I had to cut it out and resolder the strip back together.
Pixel reliability is MOSTLY a factor of manufacturing process. 99.9% of the time its the STRIP, or Solder joints connecting the pixel to the strip that fail. I work with WD2812, SK6812, APA102 and SK9822 pixels heavily, I cut them up a lot for my projects, and the one thing i see that is a major factor is how they are cooked onto the strips. You need to visually inspect your strips after you get them from your seller. Ray Wu is good, but i have received garbage from him in the past. his quality control has improved drastically over the year. I also use other manufacturers and its a game of chance. 9/10 orders are perfect, and you get hat 1 where all the pixels are barely hanging on. Burn your pixels in before placing them in a production environment
Good to hear, that the cause could be mostly the solder joints and the strip. I’m hand soldering APA102 chips myself. I’ve found that when the ground of one pixel is not connected good, the data doesn’t pass to the next pixel (while the pixel in itself will light up).
@Leon_Yuhanov What is a good ‘rule of the thumb’ time to burn the pixels in?
Yes, I had to hand re-solder a pixel in the middle of a strip today, happily it worked though, and all working well. I didn’t change it, just soldered the pads again.
@Kasper_Kamperman RE Burn in…its hard to say… I know from pas posts that @Daniel_Garcia does burn in on his strips also. Its probably user preference. I make wearables for sale so I am really rough with my pixels tt make sure they last:
1.I run a rainbow sequence on each strip(I usually get them in 50cm or 1m strips)
2.I have a demo program that runs through all the colors of the rainbow and all intensities for about 2 minutes
http://3.At the same time i physically inspect the strip and flex/shake/gently abuse it to make sure all the pixels are well connected and that the strip is not faulty somewhere(this is where I find most faults)
I find that this roots out 99% of faults. here is my instagram https://www.instagram.com/wow_elec_tron/ if you wana check out my work
@Leon_Yuhanov Nice work. What is the reason to use a rainbow sequence? I was considering running them white on full brightness (probably with a blink to make sure data is updated) for a long while. I think in that way I can also test the maximum temperature, heat.
@Kasper_Kamperman Sometimes you get pixels with a dead LED inside(either the red, green or blue) but the pixel itself functions.
@Leon_Yuhanov I understand, but a pixel will never light up white if one or more of those pixels are dead, so I think turning them white might be a faster check.
@Kasper_Kamperman faster yes, but also much more painful. Have you ever looked at 144 pixels all at maximum intensity!?
I used to use patterns to test LEDs, then I missed the fact that one pixel was missing the blue element, it was real pain to get the thing apart again and change the led.
Now I test with red then green then blue then white…