All Over But the Funding: Open Hardware Spectrometer Kit

New submitter mybluevan writes "The Public Laboratory for Open Technology and Science is putting together an open hardware spectrometer kit on Kickstarter. The kits are built using an HD webcam, discarded DVD, and a couple other odd bits. They've also put together a kit for your smart phone and open-source software for desktop, Android, and iOS. Need to analyze the contents of your coffee, the output of your new grow lights, or a distant star on a budget? Just build your own spectrometer, or pick up the limited edition steampunk version." Besides making cool hardware, they'd like to "build a Wikipedia-style library of open source spectra, and to refine and improve sample collection and analysis techniques. We imagine a kind of 'SHAZAM for materials' which can help to investigate chemical spills, diagnose crop diseases, identify contaminants in household products, and even analyze olive oil, coffee, and homebrew beer."

Pretty cool ...

[RockDoctor]

That's an interesting combination of technologies and materials. The basic idea should be workable, but how workable they'd be for material identification given the prevalences of IR filters on (consumer grade) CCD sensors and their relatively low sensitivity ....Identifying spectral lines on a 10-bit sensor is difficult enough. Trying to do it on the 7-8 bits that you get from a (consumer grade) sensor ... is going to be more difficult.

But ... that's an interesting idea. I think that's worth a six-pack worth of funding.

Did I get a first post? Are the trolls and the GNAA spammers asleep? Or, preferably, dieing slowly?

Re:I freeken love this.

[Anonymous Coward]

What is interesting about talking to researchers about spectrometers, is there is quite a diverse spread of priorities and needs depending on what they are doing. The work I did required a significant amount of time calibrating several aspects of the spectrometer (intensity and response calibration being typically much more annoying than just the position of lines). I would discuss calibration methods with other researchers at conferences, and be thinking, "How primitive, you would barely be able pass an undergrad lab with calibration like that ... oh, that is what is being used for... yeah, there isn't much point in putting more effort into." Yet at the same conference I would pump into people on the other end, who likewise were thinking along the lines of "Oh, you only spend a few hours calibrating a day and didn't take into account all these shifts and broadenings... oh, those are are a tiny fraction of a pixel on your detector."

There should be room for fun at all levels of capability (ok... more advanced spectrometers may be more tedium than fun...). If this project isn't enough, adding another zero on the price starts to get into various commercial USB spectrometers available. Those are pretty compact and self-contained, although software almost universally sucks. The only negative potentially being a slightly larger influx of crackpots whos discoveries are actually bad calibrations... I've had more than one cold contact along those lines where their spectra were clearly components of air with a slight offset due to bad calibration.

First picture on their page failed the idiot test

[thegarbz]

Really the very first picture they showed of the spectra and the resulting plot on the very top of their kickstarter page has the graph backwards compared to the visible spectra.

With such an epic facepalm I'm surprised they know what a spectrometer is.

Re:No Shazam at these wavelengths

[Anonymous Coward]

Not nearly enough.
Near IR (900-1300 nm) is the best you can hope for with a camera chip as a detector. Yes, you can sort of/kind of do a bit of "fingerprinting" there, but 10 nm spectral resolution won't cut it for this kind of application. On a side note, I seriously doubt they can get 10 nm from a CD as a diffraction grating.
To go for the traditional IR fingerprint region, they will need to look at the region between ~6000 and 20000 nm, which is decidedly not something you can do with a webcam chip. You need a *very* special kind of sensor for that. For the reference, a reasonably specced IR spectrometer will cost a tad over $10K. Most of this money will go towards the hardware, that is the sensor and the optical bench.
I wish it was really possible to identify "toxins" with a simple visible light spectrometer. I have one of these el cheap machines in my lab:

http://www.chemglass.com/product_view.asp?pnr=CLS-4047

They are only useful for measuring concentrations of very, very specific compounds under very specific conditions. We also have an $15K IR machine, as well as two mass spectrometers and routine access to 7-8 $1M NMR spectrometers. Sometimes this is enough to analyze practical samples, and sometimes it isn't.
When it is not working out for us, it is not a software problem.
It is clear to me that the people behind this project are, unfortunately, out of their depth.

Re:I freeken love this.

[photonic]

Optics guy here too. I don't know a lot about spectroscopy, but I had to assemble a spectrometer for my thesis project. It was a pretty fancy imaging spectrometer (main element was a concave mirror and grating combined in one) and used a LN-cooled CCD as the sensor. This was not cheap stuff (~5000eu for the spectrometer and probably > 20000eu for the camera), but the operating principle is exactly the same as the DVD + webcam. The resolution was limited to around 1 nm due to the input slit, not sure if they could improve things by using a slit in this home-built device. I had to calibrate it from scratch, which was actually pretty easy: I borrowed some spectral lamps from the 1st year lab course and also used a HeNe-laser we had laying around. Choose a few of the big lines (which should all be known to better than 1 nm) and for each write down the pixel position of the line on the CCD. Perform low order (e.g. quadratic) polynomial fit and you are done calibrating. I don't know if there are some cheap spectral lamps that you could use at home, there is at least the yellow lines from the Sodium (?) street lightning. I agree with others that the resolution of these home built devices is probably too low to identify materials, but it is for sure a fun project.

Re:Funding? They're way over!

[mybluevan]

They've got an interesting stretch goal though: "if we get to $100,000 -- a 'stretch goal'. In the spirit of enabling open collaboration we'd like to make it possible for people to make their own Spectral Workbench-based apps". I'm thinking a Spectral App Platform API would be pretty sweet. "Imagine a "Spectral App Store" where wine enthusiasts can develop their own testing suite for wines, gardeners for grow lamps, and coffee connoisseurs for coffee -- each looking for specific wavelengths or patterns *specific to that application*." They're not that far away from it either.

Speaking of crude detection

[dbIII]

As late as the 1990s I used a portable optical spectroscope with a battery to strike an arc and a book full of photos showing where the lines should be for specific elements for alloy identification. I only used it a couple of times and didn't get the hang of it, but apparently some people got useful results from it to the point of estimating percentages of elements instead of just seeing if they were present.