This is interesting. Guy uses a CD/DVD disk as a prism to show the spectrum of daylight vs. CFL vs. LED light sources.
I made a spectroscope using a cheap diffraction grating, a cardbox box and a ~50mm projector lens as a collimator:
The grating - dirt cheap on eBay:
You shoot the grating with the lamp shining through the slit (two utility knife blades) at the other end (guess the type of lamp).
The nearest WB to "equal energy" seems appropriate. On my Sigma DSLR these days I remove the hot mirror (easily done with the thumb nail!) to eliminate it's UV/IR blocking effect.
Then convert to TIFF and open in ImageJ to get a response graph.
Here is a comparison between an LED Flood Lamp lab test and the method describe above:
The blue was not as evident in shots as the peak would suggest but, based on the graph, I now use their 3500K model with great success.
Here's a CFL compared to a GE Halogen flood:
Hope this is of interest.