Measuring equipment?

I was wondering if you guys could outline what is needed for quality measuring of LEDs. I'm kind of at the stage where i know enough to be dangerously stupid, and am just now entering the field of optical engineering. I have a reasonable budget for this through work, and want "industrial grade" stuff (it doesn't need to laboratory precise - just something that will get used often and needs reasonably good accuracy).

I'd like to characterize the total output of high power LEDs (luxeon K2s, Cree, etc.), so i can get a measured baseline, and then measure the performance of different optics (beam pattern and efficiency, mostly). There is even an off-chance that i may be able to design a custom optic if the off-the-shelf ones don't cut the mustard.

AFAIK, an integrating sphere would be the best for measuring total LED output, but i don't know much about them... I would also like to accommodate not only LED, but halogen and HID sources as well (probably not hitting much more than 700 lumens). Also, is there a particular light meter to recommend? In addition to intensity, I also want to characterize color temperature (i'm probably always going to focus on white light).

It seems to me that i'd want some sort quality light meter, an integrating sphere to characterize the source (and would accommodate the light meter), and a wall in a dark room to shine the source (with secondary optics) on to measure the final output. Is this all i'd need?

Thanks for any insights - I'm really impressed at how active this forum is, and I'm looking forward to learning a lot.

-Damon

You need a spectrorradiometer with an integrating sfere.

There are some good manufacturers of them, offering a complete line, from "cheap" basic devices for hobbyist from about 3000$ to industrial grade for 40000$.

I saw recently a comparision between the two spectrorradiometers most used and difference in resolution is little (less than 5%).

Two weeks ago there was a new LED measurement device anounce: http://ledsmagazine.com/press/13836

Thanks for your reply, Kinnza.

I think 40,000 is a bit rich for my budget... I was looking more at:

this or this. Is there any reason why these wouldn't work with some sort of integrating sphere (not 100% sure i need one of those to get "good enough" results, anyway). Would those not be sufficient to get all the data i'd need? I used to work with an optical engineer, and i'm pretty sure he had an Olympus or a Minolta light meter which measured color temp, etc, and was good enough for our purposes.

I'm imagining that i am going to plot out a large grid on a wall, shine the light on it, and take cartesian data points (every inch or so...). For the integrating sphere... pretty much any hollow sphere with reflective paint inside and a couple ports (shine light in, measure light out) should do the trick, right? I'd like to stick an LED, or a halogen bulb, or an HID filament in this to get engine output, without taking into account optics.

Thanks,
-Damon