IR LED illumination

Here's to the experts...

I need to build a IR illumination for spectroscopy applications employing 2 different wavelengths LEDs (2x20pcs x 100mA LEDS, Uf1=1,3V, Uf2=1,5V)
I've made a first shot at that and verified that my greatest problems are temperature and supply current stability as these factors greatly influence the spectra characteristics, which should be absoulutely stable.
I know it's not 100% correct, but for the sake of saving some work I wired the stuff like that: 4 Leds in a row + R x 10 parallel makes 40 LEDs and it looked ok for me.

Anyway, what I need now is temperature control for my LED socket (aluminum preferably, but still flexible there) and a current control for 10 x 100mA = 1 A. Has anyone a fancy idea?

Thanks a lot
Melanie

What resistance do you have in series with the 4 LEDs? Suggest 39.2 ohms, 1/2 or 1 watt. How stable does the current have to be?

It is probably best to specify an operating temperature with this type of arrangement. So you power it on and wait until the system gets to ann equalibrium point. Then you take your measurements. You can heavily filter the power supply and place filters at your led board to remove any noise coupled through the power or ground planes.