Public Lab Research note

Calibrating Spectrometers from fluorescent light bulbs

by jholmes5 | March 18, 2014 01:55 18 Mar 01:55 | #10197 | #10197

I was at the Spectrometry Workshop in Brooklyn on Sunday and we were trying to follow the directions for calibrating the Spectrometers from fluorescent light bulbs.

We were having difficulty seeing with two things: 1 - seeing the three blue lines in the spectrum and thus not sure if we were selecting the "middle blue" in the calibration process. 2 - There was a lot of color between the bands, so we were not sure if there was a problem with the light source or our construction of our spectrometers.

So I had my Spectrometer in my office today and pointed it at the lighting in our office. Two things happened. 1 - I could see more bright bands in the spectra of these bulbs at work. In one image I could actually see three bands in the blue!


Photo 1: The above image is from the light bulbs at the PLOTS office. A very nice, even, "white" spectrum.


Photo 2: The above image is directly off the lights in my office at work, note the additional bright bands especially the three in the blue range. This matches more closely the images in the Spectral Workbench calibration guide. But still lots of color between the bright bars.

2 - When I pointed the spectrometer at a white piece of paper or the case of my aluminum Mac laptop the intensity of the light was reduced so I saw just the bright bars in the spectrum with reduced or no color between the bars.


Photo 3: This third image was produced when pointing the Spectrometer at the frosted silver (aluminum) surface of my computer. This reduced light intensity resulted in the drop of the color between the bright bands, but I lost that first Blue band

My next steps are going to be:

1 - Find the brand and "model" of the bulbs in our office (open the fixtures while no one is looking)

2 - Try capturing spectra of other fluorescent bulbs to see how much they vary, and if I can find a stronger "first" blue band and identify the best bulb for this calibration process.

3 - Get a mounting case so when I remove and reattach my Foldable Spectrometer I might not have to recalibrate (I wonder if that is true... I imagine that a small difference in angles might make big differences in the image location which, it seems, is what the calibration is about.)

4 - I also wonder if a narrower slit would be more accurate...

I imagine that these questions and ideas may have already come up in earlier notes, perhaps I need to review some of the earlier postings... sorry for any repetition, but we were all working around these challenges on Sunday and building our experience. :)

I'm interested in building my understanding of this tool and process to eventually use the tool to identify materials, solutions, water contaminants, ... I will have to figure that out once I better understand the tool.



Hi Jay,

This is a great example of how important it is to get a good exposure on spectral photos. You have to hit it just right to get the most information in each spectrum. If you miss by a few f-stops, you don't get much at all. There is probably a whole lot less variation among fluorescent bulbs than there is among your exposures. I haven't used a smart phone spectrometer in a while, but I remember looking for a good camera app that allowed easy manual exposure control. Taking bracketed photos is also a good idea, especially if the app will combine them for you. I never tried that. I did try a narrower slit, but that only made a little bit of a difference. I used 1000 lines/mm grating replica instead of DVD, and that might have helped some. Here is one of the first CFL spectra I made, and it's described here.

Reply to this comment...

Hey jay - this is a great post - it seems like your camera is just extremely sensitive and is being badly blown out (or clipped) by pointing your spec directly at a bulb. Doing a reflection instead to lower the amount of light is a great idea.

Reply to this comment...

Login to comment.