LED driving for Solar Simulation

Discussion in 'The Projects Forum' started by kgon13, Jul 4, 2014.

  1. kgon13

    Thread Starter New Member

    Jul 4, 2014
    I'm currently working on a solar simulator as a part of a design project for school. I'm actually studying physics but this project has me designing a circuit for driving 3w led's in parallel. I've done some research and testing and have found that buckpuck's are going to give me the highest efficiency constant current I need for driving the led's. Here's a brief overview of my driver design requirements/plan:

    1. 12 LED's in parallel (discrete wavelengths on each string covering ranging over 350-1010nm, ~3v drop/led, .7-1A per LED)

    2. Battery powered (thinking Li ion in parallel but haven't tested this yet), right now I'm testing with an adjustable power supply.

    3. Arduino PWM for dimming control

    4. So far have tested different driving methods, but buckpuck works most efficiently for thermal management, consistency and efficiency.

    I have been testing LED's using the LDU1416s1000/LDU1416s700 buckpuck which work fine for driving 1 LED.


    Circuit for 1 LED (with filter):

    My issue is that they require an LC input filter for stable current output and when I try to rig 2 or more chips in parallel I'm getting unexpected current outputs and I have burned out a few LED's in the process. I think my problem is with the LC filter. Do I need 1 filter for all pucks or 1 filter for each string of the circuit?

    Thanks so much for any help you can offer.

  2. Alec_t

    AAC Fanatic!

    Sep 17, 2013
    Wiring LEDs directly in parallel is not a good idea unless the LEDs are perfectly matched for forward voltage drop (Vf), which they won't be (probably not even if they came from the same manufacturing batch!). The 3V figure is only the typical Vf. If the LED Vfs are mis-matched you will get current-hogging by those with lower Vf values; hence they will burn out and lead to a domino effect. Each LED, or LED string, should have a respective current-limiting arrangement (e.g. resistor).
  3. kgon13

    Thread Starter New Member

    Jul 4, 2014
    Thanks for the reply Alec. The first thing I tried was a simple resistive circuit, but Vf of each LED is not temporally stable and I was getting inconsistent current output after running the circuit for 30 mins. So I thought having each LED coupled with it's own buckpuck would effectively be it's own current-limiting arangement since they are acting as a constant current source.