DIY Raspberry Pi Spectrometer Guide
github.com
github.com
Backstory: I've repeatedly encountered deep confusion about color, even among first-tier physical-sciences graduate students. Yet color is widely taught K-2. Apparently without great success. So what might a rewrite, a modern learning progression for color, look like? Perhaps one based on spectra, a modern colorspace, and building on current understanding of color perception? Tablets are used in K - "find and take a picture of a circle". So how about using them for color? There's middle-school work with color "arithmetic" (an <R, G, B> binary triple with addition(light) and subtraction(filter)). And phone spectrographs are a thing. Thermal IR inspection cameras suggest having a context image aids understandability, and phones now have multiple cameras, so might one do a more accessible sample-with-context spectroscope app? With the light path folded flat, not sticking out? And a high dynamic range to permit sampling objects under ambient illumination? Might one craft a spectra-based introduction to color? For K?
:) Partly-in-jest paraphrase: Most all models used in science education are needlessly ghastly flawed, leaving students and their teachers steeped in misconceptions. Some are useful - for important exams, or for collaborating with teachers in pretending topics are understood, though vanishingly few provide transferable or operational understanding, let alone integrated or interdisciplinary or rough-quantitative understanding. Students develop less dysfunctional understanding depending on their needs, which can be surprisingly limited. For examples, students empirically don't need to know the color of the Sun to be first-tier astronomy graduate students, nor the order-of-magnitude size of cells to be first-tier medical graduate students, so teaching the wrong color for the Sun, starting in K and continuing into undergrad intro astronomy, and teaching size/scale unsuccessfully from middle-school through undergrad, are in some sense not failing to meet student needs.
Shrug, ok. Also, it's unclear society needs, wants, or would appreciate, or even tolerate, students making sense of the physical world.
But... it can be fun, to at least discuss and explore, how we might go about it, were that an objective to be intensively pursued. No?
Public Labs even developed a modified design that works with most smart phone cameras, among their follow-up work (such as testing high-end cameras: https://publiclab.org/notes/stoft/10-23-2016/high-rez-webcam... )
The spectral response of the sensor is not linear, as it is designed to imitate human vision - and as anyone who read early 2000's digital camera reviews can tell you, even fancy cameras from well known manufacturers can have noticeably different color response.
One benefit of Rasp Pi cameras is that genuine cameras could be evaluated and characterized, but counterfeits and such will be a problem. Same is true of USB web cams, I suppose.
I tried something similar using just blu tack to hold the spectrometer to camera, from looking at the graph from it I think I possibly used the pi noir camera, as it can seem to see up to 900+nm or so.
I'd love a DIY mass spectrometer or liquid chromatograph for biohacking!
You can try building that one. It's a DIY raman spectrometer.
I'm still holding out hope someone will make an open source FT-IR design. I actually need one for a project I'm working on, and I'd prefer not to shell out thousands of dollars for a used machine.
Related: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4371691/
Beyond 1000 nm, silicon becomes transparent and ceases to work as a detector, so those longer wavelengths need a detector made from another material, notably indium gallium arsenide (InGaAs) which in one form can get all the way out to around 2700 nm. Anything that gets you away from silicon chip fab also gets you away from the fab-ulous economics of silicon. InGaAs sensors are super damn expensive.
Beyond 2700, thermal imaging cameras and the like use even more exotic sensor materials.
An alternative for those longer wavelengths is a monochromator (e.g., rotating diffraction grating detecting one wavelength at a time) and a single element detector which is cheaper than an array. If course your subject has to be sitting still for the duration of your measurement.
One of the reviews on that page mention implementing a spectrometer:
> The 1,000 lines/mm Diffraction Grating Plastic Film was used to make a high resolution optical spectrometer. First, the film was removed from its paper 2" x 2" card stock. The crystal clear film was next glued to an optical ring that was mounted in front of a very inexpensive 1,920 x 1,080 pixel webcam. The modified webcam was placed in the back oRead more about review stating Using the 1,000 line/mm Diffraction Grating in a Spectrometerf a pinhole camera-box and mounted at an incident-angle to the incoming light beam. In this configuration different light sources could be introduced to the new spectrometer and their light spectrum captured using a USB-computer input. Thus, a wide variety of light sources could be fully analyzed at a resolution of about +/- 2nm. Using this very inexpensive (1,000 lines/mm) grating, coupled with an equally inexpensive webcam, a high resolution optical spectrometer could be built for under $25.00
This is a prominent US supplier, and they offer a lot of useful info at their site:
https://www.gratinglab.com/Home.aspx
With all that said, the plastic grating is a very cool thing, and with it you can demonstrate pretty much all the basics of gratings and spectrometers.