sparkytwd
11-03-2014, 15:37
This year we're using 5 120pixel WS2812 strips on our robot.
https://lh4.googleusercontent.com/-2OPmUL9nrQk/UxqIzQVcLUI/AAAAAAABqZ8/kLFVfp5aho0/s800/DSC07412.JPG
Unfortunately by the end of Saturday we were down to one strip. After finally getting time to diagnose yesterday, the reason is likely a fried first LED. After doing a lot of data line shuffling and power system inspection, we pretty much ruled out anything else. A resistance check on the first LEDs gave a clue. A good strip showed around 4Mohm resitance between data in and ground, however on the bad strips it was between 1kohm and 400ohm. It's not 100% obvious yet what the initial cause is, however best speculation is induced voltage due to the relatively high EM environment near the relatively too-long data wires.
To that end we're now rewiring with 150ohm resistors wired directly to the input of the first LED to filter out high voltage - low current transients.
https://lh4.googleusercontent.com/-2OPmUL9nrQk/UxqIzQVcLUI/AAAAAAABqZ8/kLFVfp5aho0/s800/DSC07412.JPG
Unfortunately by the end of Saturday we were down to one strip. After finally getting time to diagnose yesterday, the reason is likely a fried first LED. After doing a lot of data line shuffling and power system inspection, we pretty much ruled out anything else. A resistance check on the first LEDs gave a clue. A good strip showed around 4Mohm resitance between data in and ground, however on the bad strips it was between 1kohm and 400ohm. It's not 100% obvious yet what the initial cause is, however best speculation is induced voltage due to the relatively high EM environment near the relatively too-long data wires.
To that end we're now rewiring with 150ohm resistors wired directly to the input of the first LED to filter out high voltage - low current transients.