AAVSONet telescope W28 at Astrokolkhoz has a new capability: a slitless grating to help identify/characterize new and transient objects.
But we need help. We need people that are experienced in spectral classification and analysis to lend their expertise to the images we collect.
More information about the grating can be found at
http://www.patonhawksley.co.uk/staranalyserusermanual.html
At this early stage in the game I've imaged a few fields to veryify proper operation. But I need guidance as to the best calibration targets (or spectral types) I should image in order to establish a calibration library. Or is that a bad approach?
And perhaps in the future we may establish a program of recurring coverage...tailored to the capabilities of this system.
Thanks in advance.
Have been using a Star Analyser for low-resolution work for the past several years. I will be happy to help and offer suggestions.
Jeff
Hopkins Phoenix Observatory
(187283)
Counting Photons
Phoenix, Arizona USA
www.hposoft.com/Astro/astro.html
http://www.hposoft.com/SB/SpectroBook.html
2012 ASP Amateur Achievement Award
What test/calibration objects do you recommend we image?
We've already imaged a few objects - attached. (Uncalibrated)
What can you determine from them? Dispersion? Resolution? Limiting magnitude for a single 120-second exposure? Other?
Star-Analyzer-01.FIT is centered at 19 51 36 +08 41 00 (near Altair) image scale is 1.08 arcsec/pixel. Image features at least one very red star with many bands in the spectrum.
Star-Analyzer-02.FIT - Beta Lyr.
Star-Analyzer-03.FIT - M57, 60 seconds
Star-Analyzer-04.FIT - M57, 120 seconds
Star-Analyzer-05.FIT - Bright M class star (Sheliak?) with RA drive off. (Grating is oriented to project the spectrum N/S)
Star-Analyzer-06.FIT - V1676 Cyg (SAO 69402) Wolf Rayet star
Thank you in advance.
Thanks, I downloaded them and will look them over and comment.
Jeff
Hi Tom,
I've had a Star Analyser for about one year. I haven't done anything fancy with it; I simply loaded it into my filter wheel. I get a dispersion of 23 Angstroms per pixel, which is very low, but the Star Analyser does an OK job with strong-lined objects (e.g., bright supernovae, novae, M-class stars) for the purposes of rudimentary classification. Indeed, there's a WR star that was discovered using a Star Analyser in an ultralow-dispersion configuration. But I must emphasize that the SA will only reveal the strongest spectral lines. For example, in my experience with observing ordinary OBAFGKM stars, only spectral classes A, M, and (sometimes) K are immediately recognizable. A higher dispersion would improve this situation.
If you want to try a new target, I'd recommend P Cygni and the WR star at the center of NGC 6888 in Cygnus. These were two of my first targets, and they have extremely strong emission lines. If you can detect them, then you'll have a better idea of what your rig is capable of.
It would be great if the AAVSO begins to expand into spectroscopy.
Colin
Hi Colin,
Just to be usre you understand, a high dispersion i NOT good. There is a big difference between dispersion and resolution. Resolution is what you want to be higher. Tom's spectra showed around 35 angstroms per pixel. The Star Analyser has a resolution of around 100 and can be increased slightly wiht a prism in front of (grism arrangement).
Your dispersion is much better. I get around 25 angstroms per pixel if I put the SA in the nose piece of the DSI Pro II. This allows both the zero order and first order spectra to be seen. ideally I believe the optimum dispersion for the SA is around 16. Going beyond that will not help the resolution, but can produce a larger first order spectrum which is somewhat easier to work with. I put the SA in a star diagonal and get around 5 angstroms/pixel. It produces a nice close to full window spectrum with the added feature of allowing rotaion of the camera to get the image horizontal.
Jeff
Hi Jeff,
Thanks for catching that. I guess I wasn't paying close enough attention to my choice of words, but you're correct.
Colin
Jeff, Tom
If you guys go offline with this exchange, would you please iinclude me (jroe at jamesroe.com) in the loop. I am very interested in this. Thanks
Hi Tom,
I looked at the spectra and have some comments. I used RSpec to process the images and spectra.
1. The spectra need to be rotated 270 degrees. As seen in the fits image they are vertical with the zero order spectrum on the bottom. The zero oder spectrum (star image) must be to the left and the first order spectrum to the right. That means the longer wavelenths are to the right (red).
Be aware that the Star Analyser will produce first order spectra on both sides of the zero order. The grating is blazed to make one first order brighter than the other. It appears you have the correct one, just rotated.
2. The best image was for beta Lyrae. Exposure was very good, but image size way way too small. In addition to rotating the SA you need to move it away from the CCD. Spacers are available and tables for calculations. There are two positions you should use. The first includes the zero order at the left edge of hte image and the first order ending at the right edge. The second position has just the first order spectrum filling going from the left edge to the right edge. These positions do not need to be exact.
You might be interested in my book "Small Telescope Astronomical Spectroscopy." I gave both Arne and Rebecca copies at the last SAS meeting.
I've attached a file with what I did and with some additional comments.
Jeff
There is no room inside the filter wheel to change spacing. (And I had to grind away metal on the filter cell to get it to fit in the tight spot of a SBIG CFW-10)
Based on existing spacing/disperson/resolution....what can this grating do?
Thank you in advance.
Hi Tom,
With your set up using a filter wheel with the SA in it, I agree there probably is not much you can do. Other than doing some experimenting and getting familiar with the spectrum processing, I don't think you will find what you are doing to be very productive.
It is nice to see the AAVSO showing some interest in spectroscopy, however. If you are serious about doing some spectroscopy I suggest you consider a LISA or Lhires III. This would require a dedicated tlescope, however. It takes a while to get everything working and once set up it is best not to distrub the spectrograph. At least during an observing season. As a minimum I would suggest a 10" SCT and at maximum a 16". Beyond a 16" it is a rapidly losing battle and not worth the large apertuure. A 12" SCT like mine can do some serious work with the Lhires III and 2400 l/mm grating on 6th magnitude and brighter stars. The LISA is lower resolution and while it will not produce the detail that the Lhires III does, it does an excellent job, has a much larger spectrum window and requires less exposure time and can image stars several magnitudes fainter.
Please let me know if you would like to put together a proposal for a dedicated spectroscopy system.
Jeff
Can you give me hard numbers that characterize this system?
- resolution?
- resolving power?
- dispersion?
- limiting magnitude for a single 120 second exposure?
- etc. etc.....
... and then some general/overall comments, such as:
- this rig can tell the difference between a slow Fe classical nova, and a fast Fe classical nova ...or...
- this rig can tell the difference between a star with emission lines (type Be, nova, bright CV), and a star without emission lines (ordinary black body spectrum) ...or...
- this rig can tell the difference between type MIII-2V-2.2 and MIII-2V-2.1 (I just made this code up, don't sweat the small stuff) ...or...
- this rig can tell the difference between a type O supergiant and a type M dwarf ...or...
- this rig can't even tell the difference between a mercury vapor lamp, and a low pressure sodium lamp...it's crap.
What can this rig do?
I'm not interested in 'parlor tricks' that are not publication worthy (e.g. measure redshift of brightest quasar....typical student lab exercise). I'm interested in filing a gap between discovery, when the initial report says "awaiting spectral confirmation as to type/ID", and when we can say "W28 low-res slitless spectra show this object to be...."
Thank you in advance.
Jeff,
May I surmise from you comments that the SA is really only good as a learning tool but not really applicable to lower tier research? If that is the case then two questions: 1) Are there any other options for us amatures to get started with (in a serious way)? and 2) Where can I get your book?
Terry
Hi Terry,
The Star Analyser is a great way to get started in spectroscopy. As for doing serious work, the low rsolution makes that difficult. People have measured Quasar red shifts with it and gotten good spectra of supernovae. For the price and ease of use it is hard to beat. Two big avantages are the ability to see the whole visible spectrum and to image faint objects.
A good portion of spectroscopy is in the processing. Once you have an image of a spectrum the fun begins with the processing.There is much to learn and practice before you try to do some serious science. The spectra obtained with the Star Analyser can provide you with what you need to learn the processing. Once you master or are at least comfortable with spectrum processing, prodicing a wavelenght calibrated line profile, correcting for instrumental spectral repsonse, measuring line wavelengths, normalizing the profile, determining equivalent widths, identifying elements, then you are ready to get serious. At that point you need a higher resolution spectrograph, a couple of monochrome CCD cameras (one for guiding and one for imaging) and a good telescope. This is a big commitment, but again it is what will be needed for serious spectroscopy.
You can find more information on my book and ordering details at:
http://www.hposoft.com/SB/SpectroBook.html
Jeff
The gratings rely on the spacing to the CCD to give the plate scale. The SA100 works well around 60-80mm.
There are limitations to the resolution... due to the aberrations etc. The best results seem to be around 30A.
A fast focal ratio down around f5 giives smaller star images and hence brighter and better resolved spectra. Match the camera to the FWHM to get optimum sampling and aim for the maximum exposure you can to improve the SNR.
Many amateurs use the SA100 (and the Rainbow Optics) as an introduction to spectroscopy. It's a great learning tool and get's you involved in the collection of useful data, the "black art" of pre-processing and then the instrument corrections, calibration and finally comparison with other spectral data.
Unfortunately the next step up - to a slit spectroscope can be expensive. SBIG don't supply thie DSS7/ SGS instruments anymore, this now leaves Shelyak (LhiresIII/ LISA/ eShel) or Baader (DADOS) as the main commercial suppliers. To assist amateurs I have designed and built a Littrow spectroscope Kit - Spectra-L200 which can be configured with various gratings - 150/300/600/1200 l/mm. At the moment this kit (which is supplied at cost) is only available to members of the Yahoo Group
Astronomical Spectroscopy:http://tech.groups.yahoo.com/group/astronomical_spectroscopy/
If there was genuine interest from the AAVSO to "standardise" on a slit spectroscope - Spectra-L200, I would be more than prepared to discuss the options.
Hi Tom,
I cannot give you precise data as I don't know much about your telescope and camera. However, I think you are doomed by using the SA in a filter wheel. The spectrum of beta Lyrae you sent me had a dispersion of nearly 35 angstroms/pixel. I typical get around 5 with my 8" LX90 and DSI Pro II camera. Even at 5 angstroms per pixel the resolution is not very high. With my Lhrtes III I typicall get .1 angstrom per pixel. That's with a 2400 l/mm grating. The SA is only 100 l/mm.
Dispersion (A/pixel) = 10000* pixel size (um) / [grating
lines/mm * grating to CCD distance (mm)]
e.g. for the Star Analyser 100 at 35 mm distance from a
Toucam Pro webcam, the dispersion would be 10000 * 5.6
/ [ 100 * 35] = 16 A/pixel
Resolution of the SA is ~100 while my Lhires III is ~18,000.
For sensitivity with an 8" telescope and DSI Pro camera people have gotten good spectra of stars down to 13 magnitude. Several people have produced excellent spectra of 12 magnitude supernovae.
Again, with your setup you most likely cannot do much. If set up properly it would have much more potenital. I have used the SA on my 8" and can easily see the stellar hygrogen Balmer lines and sodium D lines (cannot resolve the two D lines, but the combination is easily seen).
If you used a dedicated telescope, even an 8", you could do supernova spectroscopy. You might even piggyback the system on your larger scope. This is an area that would be very useful as new supernovae need immediate and continuous observations.
Bottom line, if you want to do serious spectroscopy you must use the equipment properly and for the best quality, use a higher resolution spectrograph.
Please let me know if you have any other questions.
Jeff
Hmmmm, actual
Jeff,
Hmmmm, actual numbers are always eye opening. Would using a 2x barlow half the A/P resolution or is that going down the wrong track???
http://www.aavso.org/astrokolkhoz-observatory
look at the spec's for telescope W28
What other information do you need about this scope?
I uploaded 60 and 120-second exposures of the M57 region. Lots of stars. Care to comment on limiting magnitude based on those images?
The above does not help answer my question: Can this rig ID new/outbursting objects? Yes? No? Maybe? It depends?
(I have no experience with spectra, but a casual glance at the spectra in Star-Analyzer-01.FIT shows what appears to be a very late type star with lots of bands...next to some earlier spectral class star...just right of center If someone with experience were to look at that image...I bet they could tell much more about the spectral classes present. Yes?)
This page from Michael Richmond shows some basic examples of broad spectral classes:
http://spiff.rit.edu/classes/phys440/lectures/spec_class/spec_class.html
and this page shows some 'templates'
http://www.sdss.org/dr5/algorithms/spectemplates/index.html
Can W28/star analyzer deliver sufficiently detailed spectra of new/outbursting objects...that can be compared to templates of standard/calibration objects...so that an identification can be made? Yes? No? Maybe? It depends?
Or is there a better way to maximze the performance of W28/star analyzer?
Hi Tom,
With your setup ypu cannot do anything meaningful for spectrometry.
I suggest you do some research on what this is all about and try again, I have given you the best of what I know.
The files you sent me are worse than the worse amatuer spectroscic files I have seen. You obviously have no idea what you are doing.
Before we waste any more time on this you need to do some studying. The "rig" you have is worthless for spectroscopy!
Stick with photometry or invest some time in learning spectroscopy.
Jeff
I have to strenuously
I have to strenuously disagree that the star analyzer is not good for "serious work." Obviously it lacks the resolution needed for most work involving radial velocity and detailed profile measurements but few comercially available instruments have high enough resolution to be relevant in those areas in a professional sense. And this is not going to be able to do the kind of projects that Jeff has undertaken with Eps Aur and Betelguese.
However, this grating is good for anything that a professional would use an objective prism for. Just a few examples: spectral classification in a star cluster for an HR diagram, monitoring the gross characteristics of Be star emission, monitoring emmission lines for novae or supernovae, gross measurements of spectral energy distribution, monitoring AGN variability, etc. Pros routinely use this level of resolution for almost all extra-galactic work, spectral classification, or any faint target.
Tom may have to experiment with exposure times to get better results. Keep in mind with a spectrum you are spreading the light out over more pixels. So it may be necessary to overexpose the zero-order image of the star to get good signal-to-noise in the first order spectrum. Don't worry too much about wavelength calibration to start with. The resolution is too low for accurate wavelength calibration to really matter. And Rspec does a pretty good job with that in this situation on the fly.
I did not say the SA cannot be used for serious work.
Tom wanted my opinion on his setup (rig). The way he has it set up in my opinion it will not be productive. I looked closely at the images he sent and commented. With his setup he cannot change the configuration. As such unless a different arangement is used I think it will be an exercise in frustration.
If you want to do astronomical spectroscopy there are some basics that should be followed. Having a good telescope system is just one.
I have seen serious work done with a SA on supernovae. With Tom's Rig I do not think he would be able to match that. If he follows my suggestions he probably can do some good work, but again, not with what he has setup now.
BTW, I use RSpec most of the time for the processing. I also have used VSpec and ISIS.
If the AAVSO ever gets serious about spectroscopy I would be delighted to help.
Jeff
To get a better handle on what this rig can do, I need a list of calibration/template targets so that we can build a library of reference spectra...so that we are ready to quickly analyze the next transient/outburst.
Or is that not the best approach with this rig? I'm open to suggestions.
I can set up the automation software to 'bracket' exposures...and with enough experience I can set up some rules of thumb when planning future spectral data collection plans.
Thank you in advance.
Hi Tom,
Most targets will not show any changes with your rig. With the scale of your images most of the spectra will look the same even when there are significant differences. I still think the SA would be most productive for supernova work, but not with the setup you have now.
Having a good telescope and location and one that can be used remotely and automatically would be a great asset for spectroscopy, but you will need a dedicated system. As I mentioned before, even a piggyback dedicated system would work. An 8" scope would allow many project with the SA.
I know the SA looks like a filter and can be made to fit in a filter wheel and many people may be tempted to do that, but it is not a filter and cannot be effectively used as such. The SA is a spectrograph and to be used effectively you MUST configure it properly.
Jeff
Hi John,
I didn't know it was your post. What do you think of Tom's setup? Do you have suggestions? Did you look at his images?
It would be great if you could provide guidance for him to put together a system that would be of value. Also it would help if you could provide some detailed projects and help with them. There is a lot of money invested in equipment already. It would be nice if the AAVSO could invest in say a LISA or Lhires III with suitable CCD cameras. With the proper equipment and guidance, I think the AAVSO could make a very significant contribution to Astronomical Spectroscopy. I would love to see that.
The AAVSO studies variable stars. That does not mean only measuring brightness changes. As you know, spectroscopy can be very exciting and rewarding. Just as with photometry, there is much to learn. In my opinion astronomical spectroscopy is tens times more exciting and rewarding than photometry. Just one image can provide a wealth of information. With photometry one must take data over many nights to see changes. Multiple spectroscopy observations can also track changes but on a much more detailed level.
I hope you will get more involved with this and help guide the AAVSO into this new and exciting territory.
Jeff
I cringe a little when people ask me for projects. Because really I think the best advice is to follow your interest and let that grow organically. I'm happy to talk about the things that interst me and I feel I do a lot of that. Hopefully that gives people ideas for projects that really grab their enthusiasm.
I would advise Tom that this is a good first step. He should play around with the capabilities and get to know the "rig" and what it can do. That may lead him to projects that he is interested in or it may lead him to refine his setup. It's the same basic advice that I'd give to someone who just discovered astronomy and wants to buy a telescope. Start small, build on what you are intersted in, and consider upgrading your equipment as you can to keep up with your interest.
I agree that the AAVSO should encourage its members to expand their horizons. We're still figuring out the best way to do that in spectroscopy with limited resources. I think we've had a good approach so far exposing people to the possibilities with seminars and workshops. We should provide an environment that encourages members like Tom to go out and get their feet wet. I think that our members will blaze that path into spectroscopy for us if we encourage and support them.
Hi all,
Intesresting discussion.
Of couse, SA is able to provide interinsting information on new transcients.
Especially, it produces supernovae spectra wich allows inambiguous classification between Type Ia and Type II, due to very broad lines.
About novae near maximum light, SA shows the brisghtest emission lines of Balmer serie. If it's very well tuned, it may reveals the weaker He and N lines on one hand and FeII lines on the other hand which allows to distinguish Fe and He/N types
CV in outburst do not show bright emission lines. It will be very difficult to get the carcateristic features of such a transcient (like, weak emission in the absorption component of Balmer lines, weak HeII ...)
See a page with samples of spectra of transcient objects obtained with a higher resolution spectroscope. http://www.astronomie-amateur.fr/feuilles/Spectroscopie/NewStars.html
In using SA, be careful of false H alpha line wich can be produce by zero order image of a star in the field. To confirm an identification, you have to take another spectrum of the transcient after a slight rotation of the SA.
In any case it's useful to correct the atmospheric and instrumental response in order to get the shape of the continuum, useful for identification. It's not useful to get a library of reference spectra, because atmospherric response will vary from night to night and with the altitude of the target. You have to get de spectrum of a reference star closely to the target. Any type AV star can be used.
Here's an excel spreadsheet to find easily a reference star
http://www.astronomie-amateur.fr/Documents%20Spectro/ReferenceStarFinder.xlsm
SA was perhaps the best way for mateurs to get an identification spectrum of a neaw transcient. But, in addition to very resolution, its interest for imaging spectra of very faint objects is limited by the light polluted sky.
The new low resolution spectrsocop LISA (Shekyak Instruments) has been designed for faint objects spectroscopy. Its resolution (R = 1000) allows inambiguous identification of transcients and measurement of the means caracteristics, especillay FWHM of novae lines. It works at F/D 5 and provide easily spectra of 12th magnitude targets, 12 to 14th with some experience, with small scopes (10 or 12")
With 20" scope correct spectra of 15th or more are obtained. See for instance a spectrum of symbiotic V407 Cyg at mag 16 : http://www.spectro-aras.com/forum/viewtopic.php?f=6&t=404
LISA (or the less luminous LHIRES in low resolution configuration) has produced numerous identification spectra of recent novae. For instance V496 Sct (http://www.aavso.org/publications/alerts/alert412.shtml) ; the very important Nova Mon 2012, third nova with gamma rays (http://www.aavso.org/aavso-alert-notice-467) ; Nova Oph 2012 (CBET#3124) ; Nova Sgr 2012 (CBET#3089) ; Nova Oph 2012 n° 2 (CBET#3124) ; Nova Sgr#3 (CBET#3156) ; Nova Sgr#4 (CBET#3177) ; Nova Car 2012 (CBET#3040)
Spectra and AAVSO light curves of novae are now gathred on an ARAS (Astronomical Ring for Access to Spectroscopy) web page : http://www.astrosurf.com/aras/novae/Novae_Aras.html
We do hope development of spectroscopy amougn amateurs in the next years not only for trabscient identication but also surveys of variable stars such RR Lyrae, Symbiotics, and so on
See for instance CI Cygni long term survey : http://www.astronomie-amateur.fr/feuilles/Spectroscopie/SyS/CICyg_2.html
or the very impressive change of RR lyrae in a few hours : http://www.astrosurf.com/buil/rrlyr2/rrlyr_campaign.htm
Best regards
François Teyssier (TFM)
In the last two nights AAVSONet scope W28, with Star Analyzer grating, has acquired spectra of nine stars. The images have been calibrated (bias/dark/flat) and are available here:
ftp://ftp.aavso.org/public/aavsonet/ktc/
If you sort by file date, you can easily ignore the earlier files that are not spectra.
I need feedback on exposure times, suitability of the object (for initial assessment of this rig's ability to ID new/transient objects), and it may be possible to expand/alter the list of these calibration targets.
W28 did not take spectra of all objects in my initial calibration target list. See attached spreadsheet for the master list, Open Office format. Column W shows which stars were chosen for this first list. Rows 113 - 287 are objects in the Gunn / Stryker 1983 atlas...and I tried to select two objects per spectral class (one early, one late). Rows 290 - 357 are spectral class A0 stars, and I selected a few that are visible at this time of year, and not too far south, with two imaging plans...one is high-airmass (1.8 to 2.0) and one is low airmass (1.0 to 1.2). Finally, rows 360 - 367 are some pathological stars because the Gunn / Stryker atlas contains (exclusively?) non-pathological stars.
These FITS images may contain negative pixel values, or otherwise have values of BZERO and BSCALE that confuse some 'big-name' software...but they are proper FITS files. You may have to add a constant value to all pixels (e.g. Maxim), or alter BZERO (e.g CCDSoft) to suit your software. Your mileage may vary, depending on the software you use.
The target star will appear in the upper/center portion of the FITS image (approx X/Y location is 370, 90), and the spectrum will extend below (through the center of the image). (This assumes that scope pointing was accurate.)
At least three images were taken of each object. Some objects may have 6 or 10 spectra.
For fainter objects I have limited my exposure time to 60 seconds because of limitations in the drive/autoguider logic. Stacking individual images should help.
This CCD is linear up to about 44,000 ADU. If any spectra have pixels above 44,000, please give me a recommended exposure time and I will cover that target again. (Actually I would prefer that max ADU not exceed about 25,000 because there may be nights of very good seeing, and very good tracking...that will result in sharper images and higher ADU values. Therefore a max ADU of 25,000 leaves me a safety margin.)
Are there other calibration objects I should add? Can we build suitable calibration/reference templates to aid in ID'ing new/transient objects?
Do we have enough images with which to determine the ID capabilities of this rig? (Cirrus moving in, moon still bright. I may not take more data for the next few nights, and then I head out of town for almost a week.)
I have zero experience with spectra. Your experience and feedback will help me get the most performance out of this rig.
Thank you in advance.
Tom,
I downloaded the files and ran them through RSpec.
You ask for comments, advice and suggestions. I suggest your read what I posted before. Your spectra are not good and I'm being kind. They are rotate 270 degrees. Do you know how to fix that? I suggest you investigate.
Again your images are way too small. Most people starting out with the Star Analyser even beginners are able to produce much better spectra.
You created 123 images of 123 poorly taken spectra. If you are not interested in fixing things and doing it right, why waste the time?
Jeff
I looked at a couple of these spectra; they are interesting for a first attempt. TY Dra, for example, shows the nice molecular bands of an M-star. Jeff is right that rotating the grating by 90 or 270 would be a good choice, if that is possible, since the north/south implementation means that a properly exposed spectra will likely have an overexposed/saturated/bled zero-order image that might bleed into the spectra. Rotating by 90 degrees avoids that problem. With a screw-in filter, that is not always possible.
While coarse resolution, the spectra still seem a bit out of focus. You might play with that a bit. You can focus perpendicular to the spectrum - just take a horizontal cut through the peak of the first order and make that as narrow as possible.
I'll leave other details to other readers.
Arne
Hi Arne,
To correct the rotation the SA will need to be rotated relatative to the CCD. This can be a frustrating exercise. That is why it comes with a locking ring. There may not be suffiecent clearance in the filter wheel to use it, however.
The focus did not look bad, but then the spectrum is so small it is hard to tell. I would estimate on a scale of 1 to 10 where 10 being an excellent high resolution spectrum that the Star Analyser can produce a level 3 or 4 spectrum. The configuration Tom is using is doomed to about a level 1 or 2.
It must be pointed out that spectroscopy is is very different than photometry and trying to combine the two is probably a bad idea. Tom produced 123 images. I only looked at a couple. That is a great amount of data. Who will look at all those images? If they were higher quality, it might inspire someone to look further, but until higher quality images are produced I think it will be hard to find anyone willing to spend the time on them. As it was I spent a couple of hours. With spectroscopy it turns out the imaging part is the easy part, but it must be done right. I think to be of value using the AAVSOnet telescope, a proposal must be submitted and specific configuration for the telescope specified. This would mean NOT using a SA in a filter wheel, If hte SA is to be used it must be used by itself and set up properly.
I find it interesting and exciting that you are looking more seriously at spectrometry. Dr. John Martin is helping me with a new spectrosocpic project and Campaign. If you get the spectroscopy forum up I can relate more about it.
I will be interested to see if anyone else comments on Tom's images.
Jeff
Hi Rebecca,
Thanks for creating the forum. I hope it will stimulate more interest in spectroscopy.
If anyone needs help I will be happy to offer my assistance. For those observers wishing to get their feet wet with spectroscopy I suggest considering a simple transmission grating like the Star Analyser or Rainbow optics (~$200). These are like an eyepiece filter, but are very different from a filter. Do not plan on using them in a filter wheel. They must be used by themselves.
To start out exploring what astronomical spectroscopy is all about, I suggest using the spectrograph in an eyepiece (in a telescope). Try different powers, but start with an 18 mm or 25 mm eyepiece. Look at some bright stars. Vega is an excellent choice and easy. Rotate the eyepiece until the spectrum is horizontal.
You will see the star as a bright dot with spectra radiating outward on each side. The star image is the zero order spectra. The first spectrum on either side of the zero order is the first order spectrum. The Star Analyser is blazed to make one side brighter than the other. The brighter first order side is the one to use. Ideally to start you should use a magnification such that the zero order is to the far left and the brighter first order is to the right and extends nearly to the edge of the FOV. Look at some different type bright stars and you should be able to see slight differences in the spectra.
Once you are comfortable with doing the above, put the spectrograph in the nose of a CCD camera and try again. Either color or monochrome will be fine to start. Monochrome is better becuse the same sized CCD chip will have better resolution, sensitivity and not pollute the spectrum with the color filters.
Experiment with spacings between the CCD and spectrograph. Experiment with exposures. For an 8" scope Vega needs ~1.0 seconds. Don't worry about over exposing the zero order spectrum. Try to expose the image so that the brightest pixels of the first order are close to, but not exceeding the linearity of the CCD. Typically if you keep the maximum ADU counts for the first order spectrum under 40,000 you should be fine.
It is important to have the spectrum as horizontal as possible with the brighter first order spectrum in view so that the zero order is to the left. This means the first order spectrum will have the shorter (blue) wavelengths to the left and the longer (red) wavelengths to the right. This is very important.
The Star Analyser comes with a locking ring. To make adjustments easier this ring is screwed into the camera's nose piece. The depth is adjusted so that when the Star Analyser is screwed in it will stop when the orientation is correct. This takes a lot of experimenting and can be frustrating, but once it is set you should not need to change it.
This is a start and once you get a good spectrum, I recommend downloading the trial version of RSpec and experimenting with the processing.
Good luck.
Jeff (187283)
Hi Tom,
The old standard for calibration targets was Gunn & Stryker (1983): http://adsabs.harvard.edu/abs/1983ApJS...52..121G
It lists 175 stars covering the whole range of spectral types and luminosity classes. What is nice about it is that they mention the stars that were used, so it is easy for you to set up a queue to observe each one (and they are bright).
The current standard is Pickles (1998), but that is a combined catalog of about 131 stars to obtain an even wider spectral coverage. I think if you start with G&S, you will get what you want. To some extent, you can do some of this in clusters, but it is harder to find the reference spectra in the literature - single stars are the way to go.
Arne
Thank you! That's a start.
Do you recommend I image all that are within reach of my site, or just 20, or what?
If someone were to make a list from the above atlas (star name, RA, and Dec, ASCII text, comma delimted or something similar)...I could import that into the automation software...and then buy that person a beer.
Would there be any benefit in covering objects that are not mentioned in the atlas? Be stars? Wolf Rayet? CV's at quiescence? Maybe a bright nova or supernova if the heavens cooperate? Other? (My assumption here is that the ID tasks for this rig will be for atypical objects that have emission lines or other pathological features...as opposed to blackbody curves)
And I recommend we try to establish a group of people that are willing to analyze the images, give feedback to improve the second coverage of the calibration/atlas objects...and be ready to review images when a transient pops off.
Thanks in advance.
Hi Tom,
You can use VizieR <http://vizier.cfa.harvard.edu/viz-bin/VizieR?-source=III/88> to get all the fields you want in different output formats.
Patrick
The GS set was posted to give you the names of stars that covered a wide range of spectral characteristics. I would suggest that you get data on a good fraction of these, so that you understand how/where your system works and where it fails. Certainly pick a few stars along the Main Sequence, say one from each of the spectral classes, and then some of the pathological cases.
With a diffraction grating with coarse resolution, your best use is going to be in obtaining "classification" spectra of new transients. Are they CVs, novae, supernovae, miras, what? It is often days before an observatory with a good spectrograph can observe and report on most of the transients. Known variables, like Be stars or WR stars, really require higher resolution for your observations to be relevant. However, another highly relevant project is to use the grating for GRB afterglows, obtaining spectra for the first, say, 20 minutes. Exposure times will depend on what you find by observing known brightness objects.
As for groups/teams - why not wait just a bit and see how well the grating performs and learn how to use software. There are several existing groups doing spectroscopy that you can eavesdrop on; if enough interest is shown on this forum, we can create a spectroscopy forum.
Arne
Hi Arne,
I think an AAVSO Spectroscopy Forum would be excellent. There are many members who have an interest in spectroscopy, yet for one reason or anther have not pursued that interest. I would be more than happy to help with it.
As for the Star Analyser, while as John Matrin points out serious spectroscopy is possible with it, however, just like serious observing can be done with a 60 mm refractor, one must know what he or she is doing and that takes some studying and research. As I have pointed out the configuration Tom has with the Star Anaylser in a filter wheel is doomed. If one wishes to use a Star Analyser it must be configured properly. Exposure times are not much of aproblem with the Star Analyser, even with just a DSLR and telephoto lens bright stars are fairly easy. Naturally the fainter stars will require long exposures and best a larger aperture for more light. The biggest advantages of the Star Analyser are cost and easy to use. Easy to use can be a got-ya, however. It is indeed easy to use, but you must use it properly. There are optimum spacings for the Star Analsyer and CCD. The orientation of the grating relative to the CCD pixels is important as the image must produce a spectrum as horizontal as possible.
Again I truly hope the AAVSO will be more involved with spectroscopy.
Jeff
For those that are interested, here are our current AAVSOnet spectroscopic plans.
- Star Analyzer 100. This is mounted on W28 and will be used for faint classification purposes.
- DSS-7. This will be placed on the Cohen/Menke 35cm; primarily for faint classification purposes.
- Optomechanics 10C spectrograph (long slit, like SGS). Morgan 61cm. general purpose
- echelle spectrograph to be built from ANS group. Hawkins Pond 80cm. general purpose.
- Eshell. TM61. radial velocity, bright stars
- Eshell. OC61. radial velocity, bright stars
All of these systems are either already at the telescope site or at HQ (except for the Hawkins Pond system). We expect them to all be operational during 2013.
Software-wise, we are working on our file storage system and considering file standards.
Arne
From a recent post (not me) in another Spectrocopy forum:
--- In RSpec_Real_Time_Spectroscopy@yahoogroups.com, "barker.tim43" <barker.tim43@...> wrote:
>
> Does anyone have any experience using the Star Analyzer SA-100 with an SBIG CFW-8 filter wheel and an SBIG ST-8 camera?
>
Reply
There are users of this arrangement out there. The ST8 it fine but the main drawback is that the close coupled filter wheel is too close to give the correct optimum spacing so the resolution is not as good as can be achieved. The minimum recommended spacing is 40mm with an optimum of around 70mm.
Jeff
[quote=HPO]
>
> Does anyone have any experience using the Star Analyzer SA-100 with an SBIG CFW-8 filter wheel and an SBIG ST-8 camera?
>
Reply
There are users of this arrangement out there. The ST8 it fine but the main drawback is that the close coupled filter wheel is too close to give the correct optimum spacing so the resolution is not as good as can be achieved. The minimum recommended spacing is 40mm with an optimum of around 70mm.
[/quote]
Yes that reply was from me :-)
I also run a Yahoo group for Star Analyser users here
http://tech.groups.yahoo.com/group/staranalyser
No need to trail blaze, (unless you enjoy that of course!) You just need to ask the right people ;-)The Star Analyser has been around for 7 years now with getting on for 2000 of them out there somewhere so how to get the best out of it and what it can and (just as importantly) can't do is pretty well established, though novel applications still pop up from time to time.
Cheers
Robin
www.threehillsobservatory.co.uk
I can't help but note that there has been one other other posts in the past month...
I can only conclude we have as yet not reached the AAVSO members.....
This is probably true.
There is abit more trafic on the ARAS forum.
I think that until fit files of spectra can be submitte to the AAVSO there will be limited interest here.
The T Tauri project is an interesting one and I will start submitting data for it a bit later this year when the weather is better and the stars higher in my sky.
Terry B
[quote=TCB168]
I think that until fit files of spectra can be submitte to the AAVSO there will be limited interest here.
[/quote]
I think it is important to consider that it is not a simple matter for the AAVSO to just start collecting spectra. Going with the spectrum = 1000 pictures analogy, it should be recognized that a spectrum is more complicated to reduce and store. With photometry, the result is reduced to two numbers (an ordered pair of data): a time and a value. With spectroscopy you have a larger array of numbers including time, wavelength, resolution, flux, etc. Taking the AAVSO expertise with photometry and translating it into spectrscopy literally requires the addition of dimensions and a geometric increase in complexity.
Anyone who has browsed the spectra sites for Be stars will quickly realize if people are just uploading spectra without clear guidlines and universal standards then the archive quickly becomes an unwieldly obstreperous mass. We don't want to go there because that is as discouraging as having no archive. It is also worth noting that professionals haven't tackled this yet either. Most archives of spectroscopic data exist as 2D images that the spectroscopist then reduces to their own liking and specification. Really, we'd want something more ready to use than that.
I think the AAVSO is better positioned than most organizations to tackle the issue of putting together a real useable spectral database. I've been thinking about it. Others have been thinking about it. And I think it will happen eventually. Please keep enthusastically pushing us forward, but don't forget its a challenging project and please have some patience.
-John
John is correct in pointing out that spectra must be handled very differently than photometric data. It is a problem and unless there is a specific project requesting specific spectra just archiving the data will probably be of little value.
Taking a spectrum is relatively easy. Just as astrophotography requires some learning, it is really not hard. Finding the star of interest, determining the exposure time and for high-resolution deciding what area of the spectrum to view are about all that is required. But once you have the spectrum what do you do? This is really where the interesting part comes in. Like astrophotography there is some initial processing, e.g., dark frame subtration for long exposures, delimiting the spectrum in the image, rotation if needed, background subtraction, binning if needed. Now a line profile is created. At a minimum the line profile must be wavelength calibrated. Sometimes this can be a real challenge. One way is to take a neon spectrum (or some other elemental spectrum) before and/or after the star spectrum and use the elemental lines for a calibration. Once calibrated the line profile can be examined in detail. Radial velocities can be measured, even line strengths can be determined (equivalent widths) after the profile is normalized (by dividning the profile by the continuum value - a subject in itself).
Spectroscopy is perhaps different than photometry in that the required work for even just one spectrum is many times more involved that just one photometric observation. Unless someone else is specifically interested in what the spectrum shows, it will be up to you to process and analysis the spectrum. If you fnd something of interest and report it, then others may want to zero in and study that too.
A separate item is the learning of spectroscopy. There is a wealth of science in spectroscopy and lots to learn. It can be exciting and fun. Practice is important. In fact I think this is where the AAVSO can really shine. Helping people get started with spectroscopy as they have with photometry could be a great accomplishment for the AAVSO. For now at least I do not see a great value in just archiving spectra or line profiles. Specific projects could certainly archive the results of the projects, however. I find it exciting that after many years of hoping, the AAVSO is at least looking at spectroscopy closer.
Jeff
HPO
[quote=TCB168]
Anyone who has browsed the spectra sites for Be stars will quickly realize if people are just uploading spectra without clear guidlines and universal standards then the archive quickly becomes an unwieldly obstreperous mass. We don't want to go there because that is as discouraging as having no archive. It is also worth noting that professionals haven't tackled this yet either. Most archives of spectroscopic data exist as 2D images that the spectroscopist then reduces to their own liking and specification. Really, we'd want something more ready to use than that.
[/quote]
Hi John,
I am not sure what database you are refering to but it clearly cannot be the BeSS database
http://basebe.obspm.fr/basebe/
which was set up by professionals and is populated by amateur and professional spectra. The exact format of the data is specified in detail and submitted spectra which do not meet this specification are automatically rejected. New obervers have to register and are moderated to ensure the quality of data and the details of all observers equipment is logged.
The success of the database speaks for itself with many peer reviewed papers using the data and dozens of changes of state of these stars having been detected by amateurs which otherwise would have gone unnoticed. (Helped by the ARAS BeAM website which identifies which stars need checking, based on feedback from professionals. It also supports several pro-am campaigns on specific stars.
I agree that taking spectra needs to be concentrated on specific projects (Be star monitoring is one such carefully targeted project) The reasons for this is the limited number of spectroscopists and the low productivity compared with photometry for example.
There are professional spectroscopic databases. The ELODIE archive for example
http://atlas.obs-hp.fr/elodie/
Cheers
Robin
www.threehillsobservatory.co.uk
I think it is fair to say that what suits one professional spectroscopists does not alway suit all professional spectroscopists. And purhaps "obstreperous" should be applied to the professionals and not the archives. :) I'd like to think I have a bit of a sense of humor about this. Spectroscopy has so many parameters that what is best for one purpose is not always best for another. The chanllenge for any spectroscopic archive that seeks to be universal is to provide just enough processing to be useful to anyone without going so far as to become specialized. This is why most pro-spectroscoposts strongly prefer to "roll their own" so-to-speak. Mainly this is because when you do it, you know what you did and when you get beyond the basics some details can matter.
I'd turn this around on you though because if the AAVSO where to create an archive for spectra I think that we'd want to have it also be useful and accessible to ammeteurs. And I don't mean just for uploading data, but also for downloading and using data to do citizen science. Much like we have with the AAVSO photometry archive. How do you feel about the usefulness of existing archives for doing your own science? What kind of tools or features do you wish you had at your fingertips?
[quote=uis01]
It is also worth noting that professionals haven't tackled this yet either. Most archives of spectroscopic data exist as 2D images that the spectroscopist then reduces to their own liking and specification. [/quote]
Hi John,
I am quite surprised by this. Do you have examples? While it might be the case for archived photographic spectra and the original images for electronically imaged spectra may reside somewhere, my experience is that most profesional spectra today are reduced on automatic pipelines and one rarely sees the raw images. Certainly the spectra I have received from various professionals and that I have in turn provided has been reduced as a far as is neccesary (in so far as it depends on the specific circumstances of the observation) and is supplied in the form of flux v wavelength tables either in fits format with the relevant details in the header or as dat files with the information in a readme file. In any case it would be difficult to reduce a 2D image without at least having the accompanying lamp images.
Cheers
Robin
Professionally, I don't often trust the spectroscopic pipelines. There are enough parameters in the reductions that I'd rather do it myself and be aware of everything that has gone into the process. The most notable of these archives that processes the data for you is HST. I haven't published a single thing from the HST STIS that came out of their spectroscopy reduction pipeline. This may be in part because I know too much about how the sausage was made, but I'm pretty sure I'm not alone.
The model I have professional prefered has been the one of Gemini and the Candaian Science Astronomy Data Center where they provide the raw data and a good set of tools. From there it is up to the scientist. Its a question generally of if you trust a black box to do a complex task or if you rather control all the parameters yourself.
Where does that leave the AAVSO-in terms of data collection...
I assumed they wished to collect regular spectra at various resolutions to allow comparison with the standard photometric data and previous spectral data. The degree of analysis will depend on the research being considered....
After basic pre-processing the calibrated spectral profile can be recorded as a fits file and stored, ready to be consulted and used by other researchers - or am I missing the point?
Hi Ken,
As others have pointed out, spectra archiving is much harder than photometric archiving. I think the answer is to work with projects. Specific projects request spectra and that spectra is submitted, processed and analyzed. The data are then stored. Just what point of the processing should the data be stored? I think the answer may be the final analysed data and/or final calibrated line profile. Because the quality of spectra and processing can vary greatly, there would need to be much tighter control than for photometry. I think this should be handled within a given project. A project leader would be responsible for handling and qualifying the data. While important in photometry, quality if far more important than quantity for spectroscopy.
As has been discussed before, I think the Star Analyser or Rainbow Optics spectrographs can be a good learning tool, but most projects will need higher resolution and that takes most people to a much higher level, both financially and equipment wise.
As I have said before, in addition to specific projects i think the AAVSO could play a very big role using low resolution spectrographs to get people interested and experienced with spectroscopy. I know several people who are very interested, but are having a hard time getting started. Mentor programs would be great as well as a possible user guide for those starting out. Spectroscopy can seem overwhelming, but just like photometry, once you get into it, it all makes sense and is not hard, but there is a learning curve. In the past perhaps the biggest hurtle was the software. Now with the RSpec processing software people have something that is not only easy and fun to use, but more importantly it encourages one to experiment without worrying about crashing or getting lost.
Jeff
Spectroscopy is very different than photometry. I think many people are interested, but don't know how to get started. The Star Analyser provides the cheapest and easiest way to learn the basics. Because of it's low resolution it is not likely any revolutionary data can be acquired.
Once the basics are learned it is time ot stepup to a higher resolution spectorgraph. This is an expensive step. There has been a L200 kit available off and on for around $1000. One can always make their own spectrograph, otherwise plan on spending between $4000 and $5000 for a LISA, Lhires III or SBIG unit. An eShel spectrograph will set you back over $16,000. The cost may be a big stumbling block for many. Excellent photometry can be done for a lot less.
There are several projects underway right now that can use both Star Analyser resolution and higher. Anyone interested please contact me directly at phxjeff@hposoft.com. These are easy projects and a chance to contribute real data.
As Dr. Noel Richardson says "A picture is worth a thousand words, but a spectrum is worth a thousand
pictures."
Although I don't have the experience Jeff brings to the table, I've been working with the staranalyzer about a year and will also be happy to participate any I can
Jim