# Low power RGB LED for wearable device

I am trying to prototype a wearable device which will have about 120 RGB LEDs. I was considering the WS2812B but they seem to draw about 50mA and so may not be a good option since I will be powering them from a battery.

I would be glad to know of other options or ways to implement the LEDs in order to consume less power.

• It would help to give an actual figure. "Less power" is pretty vague. How much, per LED, are you hoping to consume? (1 mA might be your hope, but I mean a practical hope). Obviously 120 LEDs at 50 mA each will consume 6 amps. Are you hoping for 500 mA? – Nick Gammon Sep 3 '17 at 7:08

You can reduce the current by simply reducing the duty cycle. For example, to display red, rather than sending 255 as the red component, send 42.

If you divide each of the R/G/B values by the same figure you get the same colour, just duller. In a quick test I found that a 24-pixel strip only used 67 mA if I sent red=0, green=0, blue=42 which made them look a dull blue. If you send a value of 8 (in a single channel) then it takes 25 mA for 24 pixels.

A bit more testing reveals a base consumption of 21 mA even if you are sending black. That is, the overhead is 21 mA, which works out at about 1.1 mA per neoPixel.

Based on that, you would consume 335 mA to make 120 NeoPixels show a moderately bright red.

I don't think there is going to be some "magic bullet" that lets you light up 120 LEDs brightly, but not use much power. However using the NeoPixels which lets you have fine control over the brightness of each of the R/G/B pixels is probably going to be about as close as you can get.

The figure of 50 mA which you quoted is to have "full brightness" on all three colours (ie. bright white).

• "If you divide each of the R/G/B values by the same figure you get the same colour, just duller." : oops, that's a common misconception. color scale values are not linear. you will find colors get very "bandy" at the low end like that. the rest is good stuff though. it would be interesting to PWM incoming power and still use 0-255 instructions to maintain relative contrast... – dandavis Sep 4 '17 at 20:25