Any cheap optical spectrum analyzers available?

[DJ]

Wondering if any chinese mfg have given a shot at a budget michelson interferometer OSA.

Always wanted one to have a look at spectral qualities of various visible to near IR sources.
Does not have to be terribly accurate. Figure they have been around for some time, and someone somewhere usually attempts an inexpensive knockoff. Any out there or other approaches?

[PedroDaGr8]

Wondering if any chinese mfg have given a shot at a budget michelson interferometer OSA.

Always wanted one to have a look at spectral qualities of various visible to near IR sources.
Does not have to be terribly accurate. Figure they have been around for some time, and someone somewhere usually attempts an inexpensive knockoff. Any out there or other approaches?

Depends on your definition of budget. Ocean Optics makes a nice miniaturized spectrometer unit for <$1k. You need a separate light source but still the combined cost is way less than the cost of uv/vis spectrophotometer. In general, the people buying spectrometers need precision, accuracy abd reliability and will pay for it. There is little to no hobby market so there is really no low end.

[FrankenPC]

I assume you have checked out Public Lab: http://publiclab.org/wiki/spectrometer

[DJ]

I assume you have checked out Public Lab: http://publiclab.org/wiki/spectrometer

Was not aware of them at all. Looks very interesting.

Thanks! 

[mrflibble]

TSL1410 + ebay-ed holographic diffraction grating does the trick. Just add stm32f4 for the ADCs and such, and you are good to go.

[FrankenPC]

TSL1410 + ebay-ed holographic diffraction grating does the trick. Just add stm32f4 for the ADCs and such, and you are good to go.

That's cool... So, the idea is to calibrate the pixel location on the array with a corresponding wavelength, then use the ADC to calculate the relative strength.  That would take a LOT of calibration!  Especially since there is no color signal out.  It's just intensity.  Seems the method used by Public Lab would be easier simply because you can retrieve the spectra along with intensity.

[DJ]

TSL1410 + ebay-ed holographic diffraction grating does the trick. Just add stm32f4 for the ADCs and such, and you are good to go.

That's an interesting idea.

http://www.digikey.com/product-detail/en/TSL1410R/TSL1410-R-ND

$35US for the sensor.

I also stumbled across a usb spectrometer,  (Czerny-Turner Spectrometer) using a ccd linear array, but ~$2k

http://www.thorlabs.com/newgrouppage9.cfm?objectgroup_id=3482

General arrangement






Nice thing about these are no moving parts, save for some adjustment screws on one of the optical path elements.

If one could find some inexpensive optics, and sort out the path lengths, might be able to 3d print the box and drop everything in. Not as cheap as the kickstarter unit, but potentially better performance (over restricted temperature I
would imagine)

Hmmmm...

[DJ]

TSL1410 + ebay-ed holographic diffraction grating does the trick. Just add stm32f4 for the ADCs and such, and you are good to go.

That's cool... So, the idea is to calibrate the pixel location on the array with a corresponding wavelength, then use the ADC to calculate the relative strength.  That would take a LOT of calibration!  Especially since there is no color signal out.  It's just intensity.  Seems the method used by Public Lab would be easier simply because you can retrieve the spectra along with intensity.

The linear array mentioned is characterized for spectral response from 400 to 1100nm (data sheet fig 3) so a rough calibration is readily characterized.


Basic info and some formulae here

http://www.spectra-magic.de/E-Spectro.htm








http://www.azom.com/article.aspx?ArticleID=10121


and a known unit utilizing a Toshiba TCD1304AP CCD sensor ($10 on ebay, not sure which linear array is simpler wrt support electronics)

http://www.oceanoptics.com/products/usb4000.asp ($3k, roughly)

[mrflibble]

TSL1410 + ebay-ed holographic diffraction grating does the trick. Just add stm32f4 for the ADCs and such, and you are good to go.

That's an interesting idea.

I know it's an interesting idea. It's even a working idea.

http://www.digikey.com/product-detail/en/TSL1410R/TSL1410-R-ND

$35US for the sensor.

Yup, that's the one. Only I think I got mine at mouser at the time. And I notice the TSL1412 has become more affordable by now... So if I were doing it again I would probably get that one.

[DJ]

Yes, $38 for the higher resolution part would make sense.

http://www.ams.com/eng/LinearSensorArray

Anyone up for doing a quick estimate of what a sensor like this could do?

Let's say 400-1100nm, which would spreading 700nm over 1536 elements. Not an optical person, but would guess some spillover into adjacent elements,  even with a monochromatic source. Still, getting within+/- 5nm would be pretty good (?)

Then on to dynamic range. Anyone care to hazard a guess on the general ballpark?

other random questions:

Slit dimensions? 

http://bwtek.com/spectrometer-part-1-the-slit/


Mirror quality and coatings?

Diffraction grate quality & sourcing?

Any existing COTS that can be repurposed/leveraged?

How much better than the (most clever) public lab offerings could be realized for say <$150 - 200US?