RGB LED pixels dying?

Interfacing hardware, sensors, output drivers, chips.
Post Reply
nadya
Posts: 1
Joined: Sat Jul 25, 2015 12:01 am

RGB LED pixels dying?

Post by nadya » Wed Jul 20, 2016 2:54 pm

The machine I'm building has a USC1903B (I think? That's the driver I'm using. USC1903 and WS2811 work just as well) RGB strip around the outside edge and I've have to replace a subsection of it a couple times.

Have you had a lot of quality control problems with these strips?

Am I doing something dumb just connecting it directly to a 12V power supply+arduino-- does it need an inline cap to protect it from voltage spikes or something?

paul
Site Admin
Posts: 706
Joined: Mon May 12, 2008 4:19 pm

Re: RGB LED pixels dying?

Post by paul » Wed Jul 20, 2016 7:36 pm

I haven't seen that before. You might monitor your power supply under some changing conditions to see if there appear to be spikes on the line. You would probably need an oscilloscope to see them however. You could try padding your 12 volt line with some largish capacitor like 1000u - make sure the current rating is 25V or more, to try to kill any spikes. The LEDs themselves are not inductive, but the wires definitely are somewhat inductive, so a long run of strips could definitely generate some spikes at switching transitions.

Also one break in a solder joint - will kill the data line after the break.

Backwards polarity will definitely kill LED strips but I doubt that is an issue. Are they outdoors? If so the issue is probably moisture / corrosion, which is is the only way I have seen LED strips die.

As to the data line, a 100 ohm (or so) resistor before the first LED is standard, to kill reflections on the line - after that the strips all have their own resistor on the data line.

Good luck with it and write back if you have new information or want to float an hypothesis on what's happening.

Post Reply