Determining the exact colour of a LED

[Circlotron]

Is there some kind of relatively inexpensive hand held device that can be aimed at a LED and will somehow show the spectral output? A friend is going to have some kind of "colour therapy" as he called it and wants to know exactly what will be beamed at him. That's all I know.

[railrun]

Define inexpensive.

[Fraser]

If this is medical kit, it will likely have both a patent and a medical certification as a treatment system. There may also be a specification on the manufacturers web site. That is the easy and free way to find out what spectrum of light is used.

As an example.....

Light boxes for the treatment of SAD (Winter Blues) used to use full spectrum white light. It was then realised that only the blue area of the visible spectrum was required. Philips introduced some excellent SAD light boxes that use an array of very bright blue LED’s. From what I understand they are not your average cheap blue LED, they are high performance with an output spectrum and intensity suited to treating SAD efficiently. Shift workers also use such units to reset their body clocks. There is plenty on the internet about the spectrum of light needed for SAD boxes so I would expect there to be similar for any other legitimate light therapy device, no matter what it’s intended therapy use.

Light spectroscopy is not normally ‘cheap’ but people have produced very basic Visible Light Spectrometers from nothing more complex than a housing, section of CD (as a diffraction grating) and a camera equipped mobile phone. An app displays the spectral result. How useful such a DIY device would be in your friends application is an unknown.

Just search for such free designs with Google. I found several straight away with a simple Google Search. Just search for mobile phone spectrometer.

https://publiclab.org/notes/omarormachea/10-20-2017/smartphone-spectrometer

http://www.upb.edu/en/contenido/smartphone-spectrometer

http://physicsbuzz.physicscentral.com/2013/02/turn-your-phone-into-spectrometer-for.html

http://scheeline.scs.illinois.edu/~asweb/CPS/

Fraser

P.S. this is not Thermal Imaging related, it is visible light spectrum technology

[Ultrapurple]

I have a couple of those visible light grating-based devices. The key part is the light entry slit - if you buy a kit with a ready made slit (normally made from razor blades) you're likely to get better results.

The cheap devices aren't calibrated but you can achieve a certain degree of calibration using lasers as a reference. Cheap red, green, blue and violet laser (pointers) give known reference peaks and you can interpolate from there. I have no idea what the repeatability is like; you may have to calibrate in-situ for each measurement.

And yes, I agree with Fraser that this isn't a thermal imaging topic. The test gear section might be a better place to pose the question - someone may even have an optical spectrum analyser readily to hand.

[kripton2035]

simple module to do this kind of measurments, with some récent AMS chip.
https://www.futurashop.it/breakout-board-spettrometro-AS72651-7100-breakout021

[Kjelt]

What I remember from four years back is that there were spectrometers with a software package to calibrate projectors , tvs and monitors that are affordable ($200).
However when the new led backlights and Led light sources for projectors arrived four years ago , those failed.
New spectrometers with special sensors were developed and those were very expensive (over a $1000).
So my guess is that there is nothing accurate and affordable on the market today, the question remains how accurate it should be for this purpose.

[Twoflower]

What I remember from four years back is that there were spectrometers with a software package to calibrate projectors , tvs and monitors that are affordable ($200).
However when the new led backlights and Led light sources for projectors arrived four years ago , those failed.
New spectrometers with special sensors were developed and those were very expensive (over a $1000).
So my guess is that there is nothing accurate and affordable on the market today, the question remains how accurate it should be for this purpose.

I think the 'cheap' display calibrators don't actually are spectometers, they are colorimeters. They have only a RGB sensor that mimicries the human eye. The more expensive spectometers are actually 'full bandwith' sensors.

[Kjelt]

I think the 'cheap' display calibrators don't actually are spectometers, they are colorimeters. They have only a RGB sensor that mimicries the human eye. The more expensive spectometers are actually 'full bandwith' sensors. 

Makes sense 

[Berni]

That depends how far you want to go.

If you just want to find the general color distribution you can just use a phone camera and calibrate it against a known white color temperature and then just take a photo of a white sheet of paper under the light.

If you want to actually see the spectrum all you need is a diffraction grating that you can look trough. You can get such manual passive spectrometers that put the grating in a box and provide a scale to read off the wavelength. You can probably take a photo trough it with a phone too. Tho this limits you to visible light since your eyes won't see any spectral lines out of the visible spectrum and the phones camera will have filters to nearly eliminate those wavelengths (Tho are still visible if they are very very bright)

So if you want to see the spectrum all the way from UV to IR you do need to buy an actual spectrometer. There are some little USB ones out there, but they do start to cost a good bit. Inside its again just a difraction grating that splits the light into components, but then a linear CMOS/CCD sensor is used to capture the light, because these have no filters they can see UV and IR just fine and are calibrated to show the correct intensity at every wavelength.

Next step up are proper lab grade optical spectrometers that have additional mechanical moving optical parts inside to get higher resolution and accuracy of the measurement, they take up a whole small table and need 2 people to move and have a price tag with 5 digits.

[Vipitis]

Maybe something like this will tell you enough. https://www.amazon.de/Astromedia-Handspektroskop-Bausatz/dp/B008I1SPWS
you can find similar things for less than 8€ elsewhere - I would consider that cheap.

But being on the topic of Thermal Imaging. Would it be possible to use a prism or grating in LWIR as well? Made from fitting medium of course? You would have to divide by your self sensors spectral response, but those are known most of the time.

[Berni]

Yep that's the sort of manual passive spectrometer i had in mind.

You could certainly make a spectrometer for long wave infrared too, tho at that point you can't use CMOS/CCD image sensors to detect the light as they are not sensitive to it. Pure silicon and germanium become transparent at those wavelengths, this makes them great for making infrared lenses, not so great for detecting it cause it just passes trough. A sutiable detector would likely be a microbolometer array such as is used in thermal cameras, a vidicon tube (CRT like device that works as a camera), or a suitable photomultiplier tube on a moving sled.

Tho not that you would see a whole lot of interesting things in long wave infrared. Anything close to room temperature glows brightly in it and emitter lamps designed for this wavelength are mostly just glorified heating elements. LEDs are incapable of producing this long wavelength of IR (Well apart from black body radiation from the LED being warm)

Similar story for going into the short wavelength UV. The longer wavelength UV works fine with normal lenses, but the shorter wavelength high energy UV needs optics made from special materials. But at least detecting UV is easy because it not only has enough energy to easily knock off electrons in pretty much all semiconductors, but it even has enough energy to rip apart molecules. This high energy UV is what gives you sun burns. The highest energy UV even rips oxygen in the air apart and creates ozone, you really don't want that shining on you, go any shorter in wavelength and you get Xrays and you really don't want to be exposed to those. But LEDs are unable to produce short wavelength UV

[richnormand]

Try this cellphone camera based for $10us

https://www.ebay.com/itm/DIY-Smartphone-Spectrometer-Grating-Transmission/283371515159?hash=item41fa422d17:g:tfoAAOSwXeJYHgk0:rk:6:pf:0

Use a regular fluorescent tube. In addition to the broadband emission of the phosphor you will also see the sharp plasma discharge emission lines for wavelength calibration.
Several examples on the net.

Also see the references in here
https://store.publiclab.org/collections/featured-kits/products/papercraft-spectrometer-intro-kit?variant=6100877213724

[Stray Electron]

Is there some kind of relatively inexpensive hand held device that can be aimed at a LED and will somehow show the spectral output? A friend is going to have some kind of "colour therapy" as he called it and wants to know exactly what will be beamed at him. That's all I know.

  If your friend is being treated by a legitimate doctor or therapist and not some quack then they have a legal and ethical obligation to fully explain what they're using to treat him. I would ask them instead of asking everyone to make guesses.  The light source that they intend to use is most likely required to be FDA (in the US) approved so the wavelength, output levels and other data should be tested and known.  If it's not an FDA approved device then I smell Quackery.

[Fraser]

@Stray Electron

+1

[Circlotron]

P.S. this is not Thermal Imaging related, it is visible light spectrum technology

Yeah, I see that now. The forum subtitle says "Everything to do with thermal imaging and cameras" so I got the idea it was predominantly thermal imaging stuff but also optical + electronic in general.

[David Hess]

Is there some kind of relatively inexpensive hand held device that can be aimed at a LED and will somehow show the spectral output?

A prism, ruler, calculator, and mark 1 eyeball are cheap.