Public Lab Research note


Homemade spectrometer

by david_uwi | September 30, 2016 12:35 30 Sep 12:35 | #13512 | #13512

380-930 nm resolution about 4 nm. Cost < 50GBP Fully self contained (battery powered). With color display (240x320) and SD card slot. Image shows spectrum of a didymium filter (wavelenght calibration) P1030132.jpg


9 Comments

Very nice, but more detail is needed.

stef

Reply to this comment...


It uses a TSL1402 diode array (I got mine from digikey) it has 256 pixels. A holgraphic diffraction grating (1000 lines) plastic lenses x2 (3 cm focal length) (both form Greenweld). Microcontroller PIC16F1788 the light source is a torch bulb. P1030133.jpg

Reply to this comment...


Here's a real spectrum which I stored on the SD card. It is of potassium permanganate. Because there is a very strong absorbance 500-580 nm and the sample is transparent elsewhere it is a tough spectrum to obtain experimentally due to random scattering off the diffraction grating. There is some internal compensation for the scattering effect but due to limited processing power (the 16F1788 has an 8-bit processor) the correction is far from perfect. It seems I can only upload JPG images so the spectrum has some JPG artifacts as well which were not present in the original GIF format.

PMAN256B.jpg

Reply to this comment...


Hi, this looks really cool! Are you interested in posting build instructions so others can follow your steps?

I'm impressed at the cost -- what's driving the display?

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


The display uses SPI communication (it has an internal ILI9341 driver). The PIC16F1788 has a hardware SPI module. These type of TFT displays can be picked up for <£10 , but they refresh rather slowly due to the slow serial communication, but this is not a problem for this application. The processor was programmed using GCBasic. Apart from the programming it's fairly straightforward stuff. Bulb - slit - followed by a lens placed at 1 focal length should give a parallel beam then through the cuvette and onto the diffraction grating. Then a lens to focus the beam onto the diode array (approx 1 focal length between lens and diode array). Getting the SD card to work was really difficult as there is little (and conflicting) information on the internet.

Reply to this comment...


Hi @david_uwi -- i'd like to help you format your notes into instructions that someone else could follow, would this be of interest?

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Yes I have no problem with that. It is however more involved than other projects I see on this site. It needs a lot of fine soldering (maybe a PCB could be fabricated to make this easier). It also requires a microprocessor and the knowledge and equipment to program it (I can make my software available). And of course some optical alignment is needed to get the light to hit the diode array (preferably in focus) and give the required wavelength range.

Reply to this comment...


@david_uwi That's fine, some of our tools require more technical understanding; the wheestat also requires programming as well as soldering (to my understanding) but if it works, it works. Hopefully over time we as a community can simplify the build, or at the very least learn from it by knowing how its made!

Thank you so much for both posting this and joining our conversation!

Reply to this comment...


is this something that a Raspberry or Arduino could be use to substitute for the processor - I ask due to the following that both processors have. also there "shields" that can be used to where only minimal amount of work is required to have a working circuit.

would you have a schematic that you could post?

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Login to comment.