Light Spectrometer with Display

Greetings,

I'd like to build a Light Spectrometer with the Arduino and few sensors. The ideal design should have a sensor that is aimed at light, then a colorful spectrum is displayed (on a small display). I did search the forum and googled but nothing so far.

Any hints/leads would be appreciated.

This has started from the need to measuring the light that grow plants emit ... to more of a curiosity, and possibly application on a telescope, to 'identify' the color spectrum of few bright stars.

I believe Forrest Mims did the original research on this, using LEDs to measure different colors of light from satellites.

Thank you. I looked him up, and while it's very interesting, he focuses more on the 'fundamentals'. That is absolutely fine, though I'm more interested in a rather 'practical' device that (I envision) all I need it to point it at a light source, and a display would show a graph of the light spectrum.

I know SparkFun has a breakout with the AS7265x sensor. Not sure yet if that is the actual sensor I'd be needing, still got some reading to do.

https://www.digikey.com/product-detail/en/sparkfun-electronics/SEN-15050/1568-1955-ND/9739876?utm_adgroup=Evaluation%20Boards%20-%20Expansion%20Boards%2C%20Daughter%20Cards&utm_source=google&utm_medium=cpc&utm_campaign=Shopping_Development%20Boards%2C%20Kits%2C%20Programmers&utm_term=&utm_content=Evaluation%20Boards%20-%20Expansion%20Boards%2C%20Daughter%20Cards&gclid=EAIaIQobChMIv5PI7ZPe6AIVSP_jBx375AP7EAQYASABEgL_c_D_BwE

There is also a youtube where the author hacks a camera ...

Check out the Public Lab spectrometer.

Thank you. It seems there are many 'copies' of the original design from an MIT course in 2013.

And https://spectralworkbench.org/

jb63:
Thank you. I looked him up, and while it's very interesting, he focuses more on the 'fundamentals'. That is absolutely fine, though I'm more interested in a rather 'practical' device that (I envision) all I need it to point it at a light source, and a display would show a graph of the light spectrum.

I know SparkFun has a breakout with the AS7265x sensor. Not sure yet if that is the actual sensor I'd be needing, still got some reading to do.

https://www.digikey.com/product-detail/en/sparkfun-electronics/SEN-15050/1568-1955-ND/9739876?utm_adgroup=Evaluation%20Boards%20-%20Expansion%20Boards%2C%20Daughter%20Cards&utm_source=google&utm_medium=cpc&utm_campaign=Shopping_Development%20Boards%2C%20Kits%2C%20Programmers&utm_term=&utm_content=Evaluation%20Boards%20-%20Expansion%20Boards%2C%20Daughter%20Cards&gclid=EAIaIQobChMIv5PI7ZPe6AIVSP_jBx375AP7EAQYASABEgL_c_D_BwE

There is also a youtube where the author hacks a camera ...

DIY Light Spectrometer - webcam & diffraction grating - Applied Science at its Best! - YouTube

I don't think that sensor gives you much spectral resolution. You'll just know that you had light at some wavelength in a very wide range.

I built something once that used a diffraction grating and a photodiode array that was showing promise. But I abandoned it and then took it apart and then lost the design. Wouldn't be too hard to put back together, I still have the PDA and I remember it being just a couple of op-amps to get the voltages to levels I could sample with the Arduino.

This was the PDA I used: https://www.mouser.com/ProductDetail/ams/TSL1402R?qs=%2Fha2pyFaduhiG1ugejqqrIjsBg02c4P4pLMwFWbxQe0%3D

This is the very basic code I used to read it:

To give you an idea how old it is, the example file has a .pde extension. You'll have to change that to .ino to run it.

It basically gives you 256 bins of 10bit data. If you put the whole rainbow on the sensor from say 380-740 that gives you a spectral resolution of 1.4nm. So each bin of data is telling you about a spectral line about 1.4nm wide. If you select a smaller portion of the spectrum you can get more resolution.

This particular PDA wasn't very good at UV wavelengths and the readings had to be adjusted even at the very blue end of the spectrum. But it worked really well I thought. There are tons more available, some with far more pixels.

The AS7265x does the job quite well - to get a rough spectrum of a light source, the bands are 20 nm or so. That's great for measuring e.g. the PAR (photosynthesis active radiation) for plant growth, which is why I got this set of sensors. I added a UV sensor for two extra UV bands, the AS7265x has several bands in the near IR but nothing in UV.

It's not a spectrometer as what you would use in measuring the spectrum of a star for element analyses, for example. Then you'd have to be looking for 0.1 nm resolution or better, which no doubt exist and are no doubt a few orders of magnitude over my electronic toy budget.