Originally Posted by electronicsfreak

Because I need it to be about equally sensitive to the visible spectrum, I'm pretty sure that almost any photo-transistor/photo-diode can be ruled out.

does anyone know of a light dependent component of which doesn't have a steep sensitivity curve? cost here is not a major issue (yet), but mass is (why I can't make a matricy of multiple photo-transistors/photo-diodes).

If not a specific component, Does anyone know what type of light dependent component would work best in this case?

For flat response curves, I have used chemical actinometry
( Actinometer - Wikipedia, the free encyclopedia ).

As that depends on a photochemical reaction, it quite easy to use at shorter wavelengths, but it is increasingly hard to find good reactions for longer wavelengths. For the infrared, I believe heat can be used (see: bolometers).

As a practical solution and considering that you need visible and infrared, I would suggest the appropriate photomultiplier, photodiode, etc. be used that has been calibrated for its sensitivity at various wavelengths. You then develop a corrected response curve and use that to calculate actual intensity at various wavelengths.

If you will be measuring polychromatic light, you will need to know its spectral distribution. But, if that distribution will be changing, I believe you are stuck with bolometry and actinometry.

You need to define more precisely the wavelength bands you need to measure and the accuracy you need. For very rough estimates, the published response curves of the detectors can be used.

John