How to focus a spectrometer?

[SteveBz]

Hi People,

Happy new year.

As we have been clouded over for at least the last month, this has to be a pretty theoretical question ☹️

I've at last worked out how I can mount my homemade spectrometer, it's a bit like the Star Analyser + Prism.  It's a 500mm grating cut down to size with a 30 degrees wedge prism, both mounted in a 31.7mm tube which screws into a nosepiece.

I've take a 70mm refractor and piggy-back mounted it on my main scope with a camera on both scopes.  So I can slew and plate-solve to a star based on my main scope, carefully align the piggy-back scope with the first one.  I imagine this 'll take a few tries.  Then I can mount the grism, but it won't be exactly in focus because of the amount of glass in the prism.  

   

In a visual scenario, I'd use Bahtinov mask or HFD/FWHM measures.  What measures or tools are there to measure focus on a spectrum?  Am I just going to have to rely on my eyes?  The main scope has autofocus, but the piggy-back mount doesn't as you can see.  Secondly, It's just about in focus optically now and there's not much outward travel left.  I could potentially use a focal reducer to bring the focuser in a bit.  Once I've got the hang of this I'll think of how to do a slit.  I imagine I may have to fork out on an Alpy 600 or something.

All ideas welcome!!

Thanks

Regards

Steve.

[robin_astro]

Hi Steve,

To focus you chose a target with nice strong features and focus on these in the spectrum, not on the zero order.  I suspect though you will have difficulties getting a sharp spectrum with this high dispersion grating and thick prism because the beam is converging through the grating and prism giving severe aberrations. (The Star Analyser deliberately uses a low dispersion grating to minimise this problem) You could adapt it to a parallel beam though using an afocal type setup, similar to this I developed using an SA200 (a sort of poor mans ALPY)

http://www.threehillsobservatory.co.uk/astro/spectroscopy_19.htm

Cheers

Robin

[SteveBz]

1 minute ago, robin_astro said:

Hi Steve,

To focus you chose a target with nice strong features and focus on these in the spectrum, not on the zero order.  I suspect though you will have difficulties getting a sharp spectrum with this high dispersion grating and thick prism because the beam is converging through the grating and prism giving severe aberrations. (The Star Analyser deliberately uses a low dispersion grating to minimise this problem) You could adapt it to a parallel beam though using an afocal type setup, similar to this I developed using an SA200 (a sort of poor mans ALPY)

http://www.threehillsobservatory.co.uk/astro/spectroscopy_19.htm

Cheers

Robin

Hi Robin,

Thanks for that.  If it proves too hard, I may go for the star analyser + prism.  The same question arises though.  Do I have to focus by eye?  Maybe I could (programmatically) draw an imaginary line along the spectrum and then use a contrast measure, like variance, for my focus measure.    If I can maximise variance, I can maximise contrast and presumably maximise the focus of the spectrum.  Is that how it works?  OpenCV gives a wealth of tools to do this sort of thing.

Thanks again.

Regards,

Steve.

[SteveBz]

11 minutes ago, robin_astro said:

Hi Steve,

To focus you chose a target with nice strong features and focus on these in the spectrum, not on the zero order.  I suspect though you will have difficulties getting a sharp spectrum with this high dispersion grating and thick prism because the beam is converging through the grating and prism giving severe aberrations. (The Star Analyser deliberately uses a low dispersion grating to minimise this problem) You could adapt it to a parallel beam though using an afocal type setup, similar to this I developed using an SA200 (a sort of poor mans ALPY)

http://www.threehillsobservatory.co.uk/astro/spectroscopy_19.htm

Cheers

Robin

Hi Again,

Actually, I wanted to ask, why did you not go the whole hog and devise a set of components, like optional slit, optional collimator, optional focus lens, all in the same Star Analyser format, that you could just buy and screw together?

Regards

S.

[robin_astro]

If you have the zero order in the same field you can focus on that, then bring the spectrum to best focus either by eye or use a real time spectrum display tool like RSpec looking for the deepest lines. If you then  note how much you have to move the focuser you can reproduce this on other targets. (Note the focus will vary along the spectrum, particularly with your achromatic refractor so it will be a compromise

 

Cheers

Robin

[SteveBz]

Hi Robin,

6 minutes ago, robin_astro said:

Note the focus will vary along the spectrum, particularly with your achromatic refractor so it will be a compromise

I wondered if that was the case.  It it so with SA and Alpy too?

4 minutes ago, robin_astro said:

real time spectrum display tool like RSpec looking for the deepest lines.

I didn't know RSpec was real-time.  I'll need to check if it runs on Linux.

S

[robin_astro]

2 minutes ago, SteveBz said:

Hi Again,

Actually, I wanted to ask, why did you not go the whole hog and devise a set of components, like optional slit, optional collimator, optional focus lens, all in the same Star Analyser format, that you could just buy and screw together?

Regards

S.

The design there was just a bit of fun to demonstrate how spectrographs work It would not be commercially viable for Paton Hawklsely who are a grating manufacturer and the performance would not be as good as an alpy which is purpose designed with special opticsso they supply shelyak with gratings and let them make spectrographs .  To be  practical, slit spectrographs also need a guider which significantly adds to thew complexity

Cheers

Robin

[robin_astro]

the simple SA configuration varies in focus because of the angle of dispersion which means the distance from grating to sensor varies. The Alpy  in focus along the spectrum because of the optics. My junk box design is also in focus within the limits of the camera lens performance.  (field flatness, chromatic aberration)

Cheers

Robin

[SteveBz]

1 minute ago, robin_astro said:

The design there was just a bit of fun to demonstrate how spectrographs work It would not be commercially viable for Paton Hawklsely who are a grating manufacturer and the performance would not be as good as an alpy which is purpose designed with special opticsso they supply shelyak with gratings and let them make spectrographs .  To be  practical, slit spectrographs also need a guider which significantly adds to thew complexity

Cheers

Robin

So I imagined I could guide my main scope which has an OAG and that would have the same effect.  Essentially my Celestron Newtonian 8" would become the guidescope.

S

[SteveBz]

1 minute ago, robin_astro said:

the simple SA configuration varies in focus because of the angle of dispersion which means the distance from grating to sensor varies.

If I buy the SA100 plus the prism, will the prism correct the focus?  That's what I was hoping.

S.

[robin_astro]

1 minute ago, SteveBz said:

If I buy the SA100 plus the prism, will the prism correct the focus?  That's what I was hoping.

S.

The prism gives only a marginal improvement for an SA100, more for an SA200 but brings its own issues (eg non liner calibration) and most users don't use one. It reduces  the aberrations a bit giving a sharper spectrum and less change along the spectrum but the focus in the spectrum is still different compared with the zero order.  Generally you can see the best focus by eye, helped perhaps by deliberately trailing the spectrum. See Christian Buil's pages for example for tips on how to get best results from the Star Analyser.

http://www.astrosurf.com/buil/staranalyser/obs.htm

http://www.astrosurf.com/buil/staranalyser3/userguide.htm

 At the end of the day though these simple setups are about good fun to learn on without spending too much before moving on to a "proper" spectrograph of good design either commercial like the ALPY or a good home build design like the lowspec.  (It is  easy to design a spectrograph on paper but there are lots of important subtleties and to produce one that actually works well in practise is much more difficult)

Cheers

Robin

[robin_astro]

19 minutes ago, SteveBz said:

So I imagined I could guide my main scope which has an OAG and that would have the same effect.  Essentially my Celestron Newtonian 8" would become the guidescope.

S

Slit spectrographs are very difficult to use except on a few bright targets without a  built in (mirror slit) guider The mirror slit guider in a slit spectrograph does more than just keep the star on the slit.

First it allows you to see the spectrograph slit and the surrounding field so you can actually place the star on the slit (which is typically only ~20 microns wide) Without this, until the star lands on the slit you see nothing in the spectrograph so you are completely blind

It then allows you to focus the star on the slit so you get the maximum amount of light into the spectrograph

Finally it allows you to keep the star  on the slit in long exposures which might be an hour or more in total on faint on objects without any issues like flexure or field rotation etc

I would say in the past this was one of the most overlooked item in amateur spectroscopy resulting in amateur built spectrographs collecting dust in cupboards. The introduction of the mirror slit guider into spectrographs for the amateur was, along with the CCD camera  probably the most significant development contributing to the advancement of amateur spectroscopy

Cheers

Robin 

[SteveBz]

1 minute ago, robin_astro said:

Slit spectrographs are very difficult to use except on a few bright targets without a  built in (mirror slit) guider The mirror slit guider in a slit spectrograph does more than just keep the star on the slit.

First it allows you to see the spectrograph slit and the surrounding field so you can actually place the star on the slit (which is typically only ~20 microns wide) Without this, until the star lands on the slit you see nothing in the spectrograph so you are completely blind

It then allows you to focus the star on the slit so you get the maximum amount of light into the spectrograph

Finally it allows you to keep the star  on the slit in long exposures which might be an hour or more in total on faint on objects without any issues like flexure or field rotation etc

I would say in the past this was one of the most overlooked item in amateur spectroscopy resulting in amateur built spectrographs collecting dust in cupboards. The introduction of the mirror slit guider into spectrographs for the amateur was, along with the CCD camera  probably the most significant development contributing to the advancement of amateur spectroscopy

Cheers

Robin 

Great answer.  Thanks for that.  Really you're saying that these slit-less ones are fun to get going, but until you have the full Alpy (all modules), Dados or I guess the SX one, you won't get the best results. In which case, I guess the Dados is best value for money.  Is that your perspective too?

S

[robin_astro]

There are alternative solutions to calibration so you can get away without the ALPY calibration module, to begin with at least. Baader do not supply one for the DADOS (well a crude neon lamp which is not sufficient) I would definitely not buy an ALPY (or any spectrograph) without a guider module though. There is no simple answer to which spectrograph is best though. It depends on what you want to do with it. Each can do things another cannot.  I have yet to see any serious results from the SX spectrograph. 

Cheers

Robin

[SteveBz]

On 03/01/2021 at 18:21, robin_astro said:

There are alternative solutions to calibration so you can get away without the ALPY calibration module, to begin with at least. Baader do not supply one for the DADOS (well a crude neon lamp which is not sufficient) I would definitely not buy an ALPY (or any spectrograph) without a guider module though. There is no simple answer to which spectrograph is best though. It depends on what you want to do with it. Each can do things another cannot.  I have yet to see any serious results from the SX spectrograph. 

Cheers

Robin

Hi Robin,

I've just been looking at your website, and it seems to me that you were happy with the SA100 for many years and you took a lot of spectra with it.  That looks like great news.  And now, even though you have the amazing LHIRES III, most of your spectra are currently produced with the Alpy 600 with guider - I guess because it's less fiddly to set up and less heavy.  Is that pretty much right?  So I'm beginning to feel that a simple grating in the converging beam is a fine way to start off, and will probably last me a while.  If my DIY version works all well and good, if not then the SA100 + prism will fill the gap.  If I then decide to move to a slit-based spectrograph, then it must be slit-based with a mirror-based guider to position the star on the slit.  In this case the mirror becomes the slit.  Because the object is now the slit at a few cm and not the star at infinity it now needs at least one new lens to focus the slit on the sensor, and preferably a second to collimate the beam between the slit and the grating. 

The last fixing bolts arrived from Amazon Prime yesterday (how did we ever live without it), and I'm just waiting for the clouds to clear, which looks like there may be a window at the weekend.   Maybe I'll practice my day-time focusing in the meantime.

Kind regards,

Steve.

Edited January 5, 2021 by SteveBz

[robin_astro]

Hi Steve,

I developed the Star Analyser in 2004 and convinced Paton Hawksley to manufacture it after experimenting with similar setups as the one you are trying. It is a great simple low cost way to try spectroscopy and you can discover a lot about how to do spectroscopy and what it can tell you before spending too much money. (There are at the last count around 6000 of them worldwide). If you are interested you can listen to me talk about the development of the Star Analyser (and how I modified my ALPY to measure some of the faintest objects ever recorded by an amateur) in this BAA talk.

https://britastro.org/video/11250/12234

You will also find more useful information on the BAA spectroscopy resources page

https://britastro.org/node/19378

The other goto website for everything about spectroscopy for amateurs is Christian Buil's

http://www.astrosurf.com/buil/index.html

I would not say most of my spectra are taken with the ALPY. (I do not put most of my spectra on my website  these days. ) I use them both fairly equally. (eg the LHIRES on bright objects at high resolution on moonlit nights and the ALPY on faint objects at low resolution on dark nights) You can see my current spectra (400 of them of 203 different targets currently) under R Leadbeater in the BAA database. (click the all headers button to see what equipment was used)

https://britastro.org/specdb/

I bought the LHIRES as a kit when it first came out in 2006 and I was a beta tester for the ALPY when it came out in 2013. The LHIRES and ALPY are quite different instruments which fulfil different functions so they complement each other.  With spectroscopy the more you spread the light out the fainter the spectrum so you can either measure bright objects at high resolution or faint objects at low resolution. (High and low resolution spectra tell you different things about the astrophysics) The LHIRES is optimised for the former while the ALPY is a low resolution instrument which can measure much fainter objects.  The ALPY is a more straightforward next step from the Star Analyser and  easier to use than the LHIRES with no adjustments once it is set up, though the processing is somewhat more complex.

Cheers

Robin

Edited January 5, 2021 by robin_astro
typo

[robin_astro]

I still like to use the Star Analyser from time to time, though mainly for fun rather than for Pro-Am science. See here for an example 

and see here on my BAA personal page for more examples of spectra taken by all three instruments

https://britastro.org/observations/user.php?user=146

Cheers

Robin

[SteveBz]

So I did have a go, in the brief couple of hours that it was clear, but I ended up getting frosted over.  Here is an extraordinarily bad picture of me trying to take a spectrum of Alpha-Cas through a layer of frost with a diffraction grating!  I think I need to buy dew heaters by the bucket-load.  Still it was a nice test.  I discovered my telescopes weren't co-aligned, but pretty much everything else worked.  I also couldn't focus with the frost and it got worse as I tried!!

[robin_astro]

41 minutes ago, SteveBz said:

 me trying to take a spectrum of Alpha-Cas through a layer of frost with a diffraction grating! 

K spectral class stars like alpha Cas are not the best choice for a first try as the spectrum  has a myriad of faint  fine lines which are unresolved resolved at low resolution. Delta Cas would be better as it is a main sequence A star with clear Hydrogen Balmer absorption lines, or if your camera has sufficient sensitivity at H alpha, Gamma Cas which is a Be star so shows He alpha in emission. See here for examples of both using an early prototype of the Star Analyser (In Patrick Moore's back garden!)

http://www.threehillsobservatory.co.uk/astro/spectra_12.htm

Cheers

Robin

 

[SteveBz]

On 07/01/2021 at 13:58, robin_astro said:

Delta Cas would be better as it is a main sequence A star with clear Hydrogen Balmer absorption lines, or if your camera has sufficient sensitivity at H alpha, Gamma Cas which is a Be star so shows He alpha in emission. See here for examples of both using an early prototype of the Star Analyser (In Patrick Moore's back garden!)

http://www.threehillsobservatory.co.uk/astro/spectra_12.htm

Cheers

Robin

 

Hi Folks,

So here is Delta-Cas from last night.  I spent a lot of time trying to get into focus.  I think a) either the star is so bright it's not a point, or b) the grating film is curved.  But as you see, the picture has some small spectra in it too:

 

I can't really see any bands.  Here is the analysis from V-Spec:

What do you think the problem is?  This is a 30 sec exposure taken with a Nikon D5000 through a cheap old 70 mm Skywatcher refractor and then processed with vspec.

Tx

Steve

[Merlin66]

Steve,

You need to flip the spectrum. We standardise spectra processing with the blue wavelengths to the left.

Very noisy. The major humps and bumps are probably due to the Bayer matrix filters on the camera.

Where did you acquire the 500 l/mm (?) grating?

[robin_astro]

9 hours ago, SteveBz said:

Hi Folks,

What do you think the problem is?  

Steve

Hi Steve,

As I said in my earlier post, high dispersion gratings with thick wedge prisms like this just don't work at all well with converging beam setups. The aberrations are just too bad so the lines are never in focus. To get a converging beam setup to work you need small dispersion angles ie a low lines/mm grating

Cheers

Robin

[robin_astro]

If you want to try using  your 500l/mm grating I can suggest mounting it in front of a DSLR lens and using it to produce spectra of bright stars using the method on my website. (The high dispersion grating works well in this case giving a sharp spectrum  as the beam is very parallel, having come from the distant stars)

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

About 20-40mm focal length lens would be about right, not the 200mm shown there as that is for the 100l/mm Star Analyser. Note though that most cheap gratings are not very efficient so the spectrum may be too faint. (Unlike more expensive blazed  gratings most of the light ends up in the zero  and other orders)

Cheers

Robin

[SteveBz]

8 hours ago, Merlin66 said:

Where did you acquire the 500 l/mm (?) grating?

It's this one:

Diffraction Grating Slide Linear 500 Lines/mm Holographic Physics Spectrum Color | eBay

with a circular wedge prism from Edmund Optics behind it.

[SteveBz]

8 hours ago, robin_astro said:

As I said in my earlier post, high dispersion gratings with thick wedge prisms like this just don't work at all well with converging beam setups. The aberrations are just too bad so the lines are never in focus. To get a converging beam setup to work you need small dispersion angles ie a low lines/mm grating

Hi Robin,

Part of the problem has been me getting used to the vocabulary like 'Converging beam' and 'high dispersion'.  I scan the sentence, think I've got it and then I find I need to revaluate some of the words.

I think all of this is really about where the focal point is and how to get the focal plane lined up with the sensor.  In this case 'high dispersion' means high angle of diffraction such that the adjacent and hypotenuse of the triangle formed by the red ray, the blue ray and the sensor are different lengths.  I thought to fix this with the prism, but really you can never map an arc of a circle onto a straight line.  

Then I think I can find enough extenders, eyepieces and home-made slits to make a collimated beam, slit and focused beam but really you can't.  It never seems to work.  I suddenly had a brainstorm the other day and I thought if I painted a clear-glass filter and then cut a slit with a razer blade, it would work.  Probably so, but there are so many other issues to think through and fix, that it's not enough.

I think, then, maybe, I just need to buy the SA100 + prism this month and get comfortable with it.  If fortune smiles on me this year, maybe I'll go all out for the full Alpy set at over £2k.  

Thanks for the help and advice.  In retrospect I think the prism has not helped in fact it made things worse.  My spectrum of Vega three years ago with just the 500 limes/mm film was better:

I think If I'd just focused this it would have been better than the prism.  But actually the big break through for me was mounting the small refractor on my main scope so that I had goto.  The big problem with the vega pic, was that I had to find vega and then nudge.... nudge.... nudge... until the spectrum was in the right place.  I guess I want 2 things.  1) a spectrum and 2) goto and 3) I don't want to insert/remove/rotate gratings once I'm there.  I want a permanent set up, so that if the clouds part for 30 minutes, I'm in.

Thanks again.

Regards

Steve.