SPD curves for standard light sources

tedfine

Newly Enlightened
Joined
Apr 13, 2018
Messages
4
I'm looking for suggestions for commonly available (like from Home Depot) and inexpensive lamps that can be used as a standard calibration source for an inexpensive, low resolution, slit-less spectrometer. Ideally the lamp would be small (like an LED) and easy to wire up to a 9 volt battery (preferred) or 110 VAC.

We need this lamp so students can do two types of calibration:


  1. Relative flux calibration (to remove non-linear instrument sensitivity across wavelengths) : The light source needs to have a visible continuum rather than just a few broad emission features.
  2. Wavelength calibration: The light source needs to have one fairly defined emission feature.

I don't need a high degree of accuracy. Students will be capturing spectra and determining their approximate color temperature.

I've thought about using the flashlight (torch) on an iPhone, since they're, um, pretty common.

I know that some lamps change in SPD over time, but I don't know to what degree this is the case with a cell phone or other possible standard lamp.

Any suggestions on lamps?

While researching this a few years ago, I found a site (GE?) that provided SPD graphs for a wide range lamps. I am unable to find this site again. The graph below (from another thread on this board) looks like it is from that site.

This is the kind of graph and standard lamp I'm looking for. I can smooth the features for #1 above. And it has some features with a known wavelength.

reveal.jpg
 

maukka

Enlightened
Joined
Dec 22, 2015
Messages
641
Location
Finland
I can't recommend any common light source found from hardware stores on that side of the pond, but here's measurements for iPhone 6 and iPhone 6s. They're pretty similar. I wouldn't assume similarities between different phone generations though.

Y0AfWxJ.png


8jj04Qc.png


Here's the 6s data in numerical form at 10nm intervals:

3800,002983
3900,003852
4000,005529
4100,029855
4200,154438
4300,467904
4400,894212
4501,000000
4600,447946
4700,168201
4800,090034
4900,108603
5000,230799
5100,448857
5200,659658
5300,796396
5400,874633
5500,913582
5600,932612
5700,919011
5800,869396
5900,801903
6000,726971
6100,648119
6200,567194
6300,487531
6400,410659
6500,339508
6600,276217
6700,217431
6800,169606
6900,132492
7000,104206
7100,081377
7200,065925
7300,057547


ZkycwIF.png
 

ssanasisredna

Enlightened
Joined
Oct 19, 2016
Messages
457
I would go with a UV pumped LED. The phosphor has a well defined output characteristic

I would not use an LED for wavelength calibration. You are better off with a laser as you have a narrow line, and ideally pick up a used HeNe as the wavelength will be more accurate.
 

tedfine

Newly Enlightened
Joined
Apr 13, 2018
Messages
4
Maukka: THANKS so much. GREAT data. Double thanks for the data points, which saves me from having to digitize it myself.

Where did you get the from? Do you know of on-line sources of this kind of information for other light sources?

ssanasisredna: Thanks for the suggestions. Seems like both of the light would be great for my needs, but I'm hoping to find standards that would be available locally. Our target audience is teachers.

Thanks again to you both for your help.
 

iamlucky13

Flashlight Enthusiast
Joined
Oct 11, 2016
Messages
1,139
The graph you show looks like a standard incandescent light bulb, which will produce an almost perfect black body curve around 2700K, plus a comparison of the GE Reveal series incandescents, which is an incandescent with a color filter incorporated in the bulb to make the light appear less yellow-hued.

I don't know if having a spectral peak outside the visible wavelengths will work for you, but you can usually find incandescents by asking around if people you know still have some spares. Or you can still buy halogen incandescent bulbs. I think the major home improvement stores still carry them.

Out of curiosity, I'd like to ask what the spectrometer you're referring to is? It sounds like a step up from the diffraction grating and digital camera project I've seen shared online (which I think might originally have been a NASA educational project)?
 
Last edited:

ssanasisredna

Enlightened
Joined
Oct 19, 2016
Messages
457
The graph you show looks like a standard incandescent light bulb, which will produce an almost perfect black body curve around 2700K, plus a comparison of the GE Reveal series incandescents, which is an incandescent with a color filter incorporated in the bulb to make the light appear less yellow-hued.

I don't know if having a spectral peak outside the visible wavelengths will work for you, but you can usually find incandescents by asking around if people you know still have some spares. Or you can still buy halogen incandescent bulbs. I think the major home improvement stores still carry them.

Out of curiosity, I'd like to ask what the spectrometer you're referring to is? It sounds like a step up from the diffraction grating and digital camera project I've seen shared online (which I think might originally have been a NASA educational project)?


That 2700K is "around" 2700K and highly dependent on line voltage (or battery) voltage both which vary greatly and the OP was looking for a spectral reference. You don't get that with an INCAN without being calibrated to a voltage.

w.r.t. "Local" ... Digikey, Mouser, Farnell, and number of places ship overnight or in a few days fairly inexpensively. A UV pumped LED is going to have a relatively consistent spectra somewhat independent of drive current hence why it works well as a somewhat "calibrated" source. Soraa bulbs are getting a bit easier to find these days too.
 

tedfine

Newly Enlightened
Joined
Apr 13, 2018
Messages
4
Out of curiosity, I'd like to ask what the spectrometer you're referring to is? It sounds like a step up from the diffraction grating and digital camera project I've seen shared online (which I think might originally have been a NASA educational project)?

This is a project using a DSLR with an objective grating: https://www.fieldtestedsystems.com/starspectra/. It's a slit-less spectrometer originally designed for astronomy.

However, it can be used on any compact light source, which means street lights at a distance (with dark surroundings), or closer objects in a dark room and/or behind an external slit (e.g. behind a closet door which is open just a crack...)

The software can show a corrected color curve, adjusting for instrument response in real time (video 30 and 15 here: https://www.rspec-astro.com/more-videos.)

AFAIK, most webcams can't be used for this kind of work because they have automatic gamma correction and other automatic exposure corrections that constantly change camera response.

However, since you can turn off these automatic features in a DSLR, a true black body (SPD) curve can be created. However, in order to do that (as shown in the videos), one needs to know the QE of the entire system (not just the CCD, but the lens and the grating too). THAT'S why my original question looking for standard curves, which is used to create a system-specific QE curve.
 

ssanasisredna

Enlightened
Joined
Oct 19, 2016
Messages
457
This is a project using a DSLR with an objective grating: https://www.fieldtestedsystems.com/starspectra/. It's a slit-less spectrometer originally designed for astronomy.

However, it can be used on any compact light source, which means street lights at a distance (with dark surroundings), or closer objects in a dark room and/or behind an external slit (e.g. behind a closet door which is open just a crack...)

The software can show a corrected color curve, adjusting for instrument response in real time (video 30 and 15 here: https://www.rspec-astro.com/more-videos.)

AFAIK, most webcams can't be used for this kind of work because they have automatic gamma correction and other automatic exposure corrections that constantly change camera response.

However, since you can turn off these automatic features in a DSLR, a true black body (SPD) curve can be created. However, in order to do that (as shown in the videos), one needs to know the QE of the entire system (not just the CCD, but the lens and the grating too). THAT'S why my original question looking for standard curves, which is used to create a system-specific QE curve.

Why not just calibrate against a known star spectra and atmosphere density profile for your location and viewing hour? ... assuming a clear sky of course. I have to expect that is going to be as accurate as any uncalibrated source though I still think UV pumped phosphor may be ideal, though the issue will be with multiple phosphor coatings and the relative ratios screwing up the accuracy. I can't remember where now, but someone was selling phosphors online in small quantities. That may be a solution for you.
 

tedfine

Newly Enlightened
Joined
Apr 13, 2018
Messages
4
Why not just calibrate against a known star spectra and atmosphere density profile for your location and viewing hour? ...

Yes, that's a reasonable approach, but requires the student to capture a star's spectrum, which is surprisingly difficult, involving finding the star in the sky, pointing the camera at it, keeping the camera on the star as the sky moves but keeping the star on the same pixels for 3 to 30 seconds (in other words, "tracking"), setting focus, exposure and ISO levels for a very dim object usually not visible in the view finder (only visible in the jpg/raw output files) and perhaps learning how to operate the manufacturer's camera control software.

Phew! That's a lot!:sssh:

If a student is doing astronomical spectroscopy, then he or she will have already learned how to do the above by the time they need to create a system's instrument response (QE).

But if the student is just doing terrestrial observations, it's much easier to use a lamp of a known SPD rather than meeting all the astronomical challenges. :)

Thanks for your suggestion of phosphor lamps.
 
Last edited:
Top