Sometimes Experiments Fail
Changing the Design
The original design was very cheap by spectroscopy standards, however after beginning the build a few things became apparent. One was the Raspberry Pi controlling and interfacing the system being very unnecessary and not cost effective. At first it seemed like a perfect way to control the system. The board and camera module allow for powering the LEDs and capturing the spectra. The issue is that, if you are trying to reduce costs of the lab, it isn’t efficient to have an entire computer system that is only doing spectroscopy acquisition. From this, I decided to make the spectrometer much simpler, a USB device. This drastically reduces the system cost. A CMOS webcam can be purchased for $3 and has adequate resolution for collecting spectra, LEDs can be ran using a battery and switch (I used an arduino board to power mine as I had the board lying around), and the rest of the work can be done with software to produce spectral data.
The First Design
Since building a webcam spectrometer is not something unique, I based the design loosely off those of others, but had no success. The design used a DVD as the diffraction grating, a $3 webcam from a dollar store, cardboard, and some bright white LEDs. Unfortunately I was not able to collect a spectral pattern. A picture of a captured spectra on the CMOS from the second rendition of the device is shown to the right and was the goal for the first model.
The system is not complicated but relies on a few needs. If we were to break this down into two criteria that must be fulfilled, I would say it is detecting enough light and having good alignment of the components. The first try likely failed because of poor alignment which in turn meant to light was detected. A few things I did note is that the camera was quite sensitive. This is great news because it means a narrower entrance slit can be used on the spectrometer giving better resolution of peaks. The sensitivity also means that the goal of capturing fluorescence spectras is not out of the question.
While the first build was far from a success, much was learned and an emphasis on carefully checking alignments and how the instrument will function can be taken into account on a second build, hopefully ending with a working system. Some simple issues that should be worked on that I hadn’t thought of prior to the build is the physical shape of the device. The first was very light and narrow making it way to unstable to sit on a bench and record reliable data. The lessons learned from the first construction will be applied to hopefully produce a working spectrometer for under $10, a significant reduction of cost in an already inexpensive design.