higher current LEDs

[Paul Laskowski]

Hi Andrew, I wanted to see if you, or anyone on this list, has any good ideas for controlling LEDs with more current than 20mA.  We're preparing to bring the Garden of Missed Connections 2.0 back to Burning Man, and as part of our redesign, we want to get more brightness out of our lanterns.  (Last year's 20mA LEDs looked great at home, but didn't carry very far over the open desert, which is a mile or two across)  We're thinking about switching to flexible RGB LED ribbons, like http://www.amazon.com/Flexible-Changing-LED-Strip-Ledwholesalers/dp/B00476B908.  These would be nice because we can wrap them around a spool so they shine out at our lantern sides, without using a diffuser.  We're thinking of something like 300mA per channel.

The LED ribbons have built-in resistors, so we just have to connect them to a 12V source.  A friend who knows more about electronics suggested the following: "you'd take the output of the lightuino, attach it to the gate of a high current transistor. Then link your battery on the source of the transistor, and the LED on the drain. The board now drives the transistors that in turn drive the high power LED."  The project has 20-30 foot runs, so I'd want to connect the outputs to rj-45 jacks as we did last year.  Any ideas about how to accomplish this?  My first thought is using one of your rj-45 breakout boards to connect the lightuino to a breadboard with transistors, and use a second rj-45 breakout plugged into the breadboard to connect the output to our cat5 cables leading to the lanterns.  The breadboard might be hopelessly crowded with everything in there.  But this seems like it should be a fairly common problem, so I wonder if a board already exists to do all this for us.  Let me know if you have any good ideas.

Thanks!

Paul

http://gardenofmissedconnections.blogspot.com/

[Andrew Stone]

Hi Paul,

Yes that's exactly what you'd do.  I actually have a board just now in house that another user asked me to build (Scott, sorry for the delay if you are reading this, I have NOT forgotten about you!) that does something like this.  One clarification, you'd want to use PNP transistors on the high side -- its what your friend suggested but I just want to be clear.  The board I built converts the IDE to screw terminals and passes the signals optionally through a mid-size PNP transistor and then to a serious heat-sinkable transistor.  It turned into a pretty large and expensive board to build due to the size of the components and the trace widths.  I think at 300mA you'd be ok with something smaller.  The "mid-size" PNP transistors I bought are capable of 3.5A (if turned on "hard" which is what you'd do).

It may not matter, but to be on the safer I think you'd actually want to bring the power to your lanterns independently.  Put this transistor (and a nice sized capacitor) right next to each strip.  That way when you PWM the LEDs to change color you won't be actually driving 300mA on and off through the RJ45 cable.  Given your long runs, this kind of power might generate some RF and possible some unwanted inductance.

But if its inconvenient to locate the transistor in the lantern, just mock up your longest run first to make sure it works :-).  POE (power over Ethernet) specs 1A per conductor, so you should be OK with 300mA -- at least you won't melt the wire :-).  If you burn out the PNP transistor with a long wire run but it works fine with a short you have unwanted inductance and you need to put a bypass diode in just as if your wire was a motor coil...

Cheers!
Andrew