Star analyser use for measuring camera frequency response

[billhinge]

Apologies if this has been asked before

I have just ordered the SA 100, some filter rings for attaching to camera lenses and a copy of 'Grating Spectroscopes and How to Use Them'

Apart from the astronomy aspect which I am interested in, one other long time question is bugging me - what is my dslr's spectral response to monochromatic light in the range 400-700nm

I have a Gerd Neumann Aurora panel as a potential monochromatic source, would this set up allow me to make a simple measurement ? (simple is key word here)

Are there other factors I need consider ?

[Merlin66]

If you can rig up a slit arrangement... (See Chapter 5) you can use a solar spectrum to help determine the "instrument response" curve with enough resolution to see what the camera does at Ha wavelengths.

Buil has already done this for a Canon....

http://www.astrosurf.com/buil/cameras.htm

[billhinge]

Hi Merlin

Sounds what I'm looking for, many thanks

It was by chance I stumbled on the AS 100, I was looking for lumicon filters when I was googling and I came on the SA 100 and it sounded a pretty interesting branch of the hobby

I did astrophysics many years ago (early 80's) at Leicester university and I remember doing a lab experiment on spectroscopy with one of the early apple computers (pre mac)

[tom--rspec]

An easier (IMO) way to determine your DSLR's spectral response is to take a spectrum of a bright type A star like Vega. You don't even need tracking. See the details at this link.

It's rather remarkable what a DSLR can do spectroscopically. See, for example, what Janet SImpson did with a 30 -second exposure on her DSLR with simple mechanical tracking: link.

Tom

[billhinge]

Hi Tom

Thanks for the link, I had seen it before but perhaps I'm misunderstanding something (the book I ordered wont arrive for a couple of days). Surely many stars will have different light outputs at different wavelengths eg some stars are red, others blue. So wouldn't the recorded spectrum on camera be dependent upon the the stars output at different wavelengths, ie a blue star will record with a brighter image in the blue than red ? Or are you saying that type A calibration stars effectively are white with a balanced output across the visible spectrum (within the measurable sensitivity of the camera)?

[Merlin66]

You are correct about the "shape" of the spectrum recorded.

The instrument response is obtained by dividing the aquired spectrum after calibration by a "standard" spectrum - this is most easily done with an A0 star like Vega. The resulting response curve can then be divided into another (unknown, but calibrated) spectrum to give a realstic Planck/ Black body curve for the star.

HTH

[billhinge]

Thanks Merlin, just had quick flick through the documentation for Visual Spec

[tom--rspec]

Yes, Bill, as Merlin points out, you'd want to do a quick instrument calibration. Check out video 15 on the RSpec site: http://www.rspec-astro.com/more-videos to see how it's done.

Tom

[nytecam]

If you can rig up a slit arrangement... (See Chapter 5) you can use a solar spectrum to help determine the "instrument response" curve with enough resolution to see what the camera does at Ha wavelengths. Buil has already done this for a Canon.... http://www.astrosurf...uil/cameras.htm

Thanks Ken for the link to Buil's work - the dire performance of DSLR v mono CCD is very revealing - Buil then states RGB on Mono reduces effective CCD 'speed' via filtration but I find under test my complementary [CMYG] filtered SX Lodestar-C I use on astro runs @ ~60% of mono Lodestar 'speed' so colour without massive loss of photons.

ps: Bought 1100D @ Jessops yesterday at a modest £273 [incl Canon cash discount] in their current deals

[robin_astro]

An easier (IMO) way to determine your DSLR's spectral response is to take a spectrum of a bright type A star like Vega. You don't even need tracking. See the details at this link.

It's rather remarkable what a DSLR can do spectroscopically. See, for example, what Janet SImpson did with a 30 -second exposure on her DSLR with simple mechanical tracking: link.

Tom

The techiques described in this thread so far will not give you the response of the camera. The instrument response obtained this way includes the effect of all the optics, particularly the response of the grating which is as significant as the camera. It also includes the effects of the atmosphere.

What is asked for here(ie the response of the DSLR sensor is in fact difficult to measure with any accuracy using amateur equipment. Buil has attempted it here for several DSLR by comparing with a pre-calibrated camera for example but only claims 10-15% accuracy

http://www.astrosurf.com/buil/50d/test.htm

There is also more on this problem here on the AAVSO forum

http://www.aavso.org/filter-response-curves

Cheers

Robin

www.threehillsobservatory.co.uk

[robin_astro]

but I find under test my complementary [CMYG] filtered SX Lodestar-C I use on astro runs @ ~60% of mono Lodestar 'speed' so colour without massive loss of photons.

Hi Maurice,

While it is true that CMYG allows more photons to be detected in total and therefore the sensor is more sensitive overal when displaying the result as a mono image, (eg the L channel in an LRGB image) This advantage is not carried over into the colour information. When generating the required RGB chanels for display (or for measurement in the case of spectroscopy/photometry) from the complementary channels produced by the CCD, one has to subtract two counts for each RGB channel. When this process is carried out I think you will find you wind up with the same level of photon noise in each chanel as for an RGB filtered CCD. ie There's no such thing as a free lunch

Cheers

Robin

[robin_astro]

When generating the required RGB chanels for display (or for measurement in the case of spectroscopy/photometry) from the complementary channels produced by the CCD, one has to subtract two counts for each RGB channel.

Actually that should read

"When generating the required RGB chanels for display (or for measurement in the case of photometry)......"

There is an advantage of CMY over RGB for spectroscopy but this is still not a good reason for chosing a colour camera over a mono for spectroscopy ;-)

Robin

[billhinge]

Thanks for the helpful information everyone - I do have a couple of other tools eg rawdigger (http://www.rawdigger.com/news ) which apparently can measure QE for dslrs (and Pixinsight if that helps?)

on a slight aside .... I now have the SA 100 and yet more step up/down rings and a myriad of astro couplers/adapters. Unfortunately a 28mm stop down ring is 0.5mm too small for a 1.25" filter

Any better ideas than blutak to hold the filter onto the 28mm ring so that it may be attached to a dslr & lens similar to the AD-55 http://www.rspec-astro.com/dslr-adapter/ (don't fancy $38 + P&P)

[Merlin66]

Bill,

see if you can by a spare lens cover from your local camera shop, or ebay for $2....

Cut a neat (!) 28mm hole in the centre and screw the grating filter into the cover....job done!

[robin_astro]

Bill,

see if you can by a spare lens cover from your local camera shop, or ebay for $2....

Cut a neat (!) 28mm hole in the centre and screw the grating filter into the cover....job done!

A couple of improvements on this are either

Screw (and glue) one of Paton Hawksley's spacers into the lens cap first that way you can easily rotate the grating or even better, use a rotatable filter cell as used on polarising filter pairs as I have here

http://www.threehillsobservatory.co.uk/astro/spectroscopy_11.htm

This is very useful for lenses (like the standard Cannon zooms) which, annoyingly for this application, rotate the front element as they focus.

Cheers

Robin

[robin_astro]

A couple of improvements on this are either

Screw (and glue) one of Paton Hawksley's spacers into the lens cap first that way you can easily rotate the grating or even better, use a rotatable filter cell as used on polarising filter pairs as I have here

http://www.threehill...troscopy_11.htm

This is very useful for lenses (like the standard Cannon zooms) which, annoyingly for this application, rotate the front element as they focus.

Cheers

Robin

Oh and if selecting a clip on style lens cap, make sure that the clip mechanism clears the region where centre hole will be. Not all designs do unfortunately. I have attached an image of my current version

Cheers

robin

[billhinge]

Thank you all for the good advice, I'll give these a try. In the mean time the book arrived from amazon today so I'll have a good read

[billhinge]

Had an idea, just ordered some polymorph granules to make an adapter, got a step ring that looks like it would mate with a spare 1.25" nosepiece (?) if I make a coupler

(good excuse to try the polymorph since I saw it a year or so back, maplin has it now http://www.maplin.co.uk/polymorph-35511