Yes they do look straightforward to construct. Everything is more difficult in the doing least for me :-)

PE raises a good point regarding calibration.

I understand that a fluorescent light source would have many spectrum lines in it. See the picture of a fluroscent light spectrum at:

Baker has some paragraphs on calibration in Photographic Emulsion Technique, 1948 edition. He mentions using an arc lamp (not something widely used today). I was thinking that Light Emitting Diodes (LEDs) might be a more modern alternative.

If I ever get around to making one I would use it for relative differences. For example seeing if the emulsion was now sensitised to green light. Exposing some panchromatic emulsion such as Plus-x etc would give a "reference" spectrograph. This avoids the need to calibrate.