Affiliation
American Association of Variable Star Observers (AAVSO)
Tue, 11/25/2014 - 20:21
It is recommended that you find a nice A0 star for calibration of your grating. Vega seems to be the first choice. But if that's not available I've found this list of A stars that you might try: http://www.solstation.com/stars3/100-as.htm .
We might use this topic for refining this list and posting which stars have worked well.
George
A-type stars work well.
If you can do fainter, I suggest targets with sharp emission lines because the contrast is often more between an emission peak and the continuum than between the absorption line and the continuum. Keep in mind that for emission line targets the emission lines are often brighter than their V or R magnitude will imply because most of the flux emitted by the object is coming out over very narrow bands of wavelength.
If you use a planetary nebula, pick one that has a relatively small angular size so it does not blur the lines in the slitless spectrum. You can find a list of planetary nebula here:
http://www.atlasoftheuniverse.com/plannebs.html
A list of the lines in planetary nebular is here (but keep in mind some of them will blend together at low grating resolutions):
http://laserstars.org/data/nebula/identification.html
You can also use an AGN (quasars, active galactic nuclei). But if you use an AGN keep in mind you must look up and apply the cosmological redshift (NED list them http://ned.ipac.caltech.edu) to the Balmer emission lines in the spectrum. A list of the brightest AGN are here:
http://spider.seds.org/spider/Misc/qso.html
RTFM :-) There is a full list of bright stars by by spectra type (and lots of other useful information in the SA manual which I posted a link to
Most planetary nebulae are too diffuse to do decent spectroscopy on using the SA unless you use it in front of a camera or add a slit
http://www.threehillsobservatory.co.uk/astro/spectroscopy_18.htm
bottom of page
The SA is for star-like objects.
If you are looking for emission line targets then:-
I recommend WR stars (also listed in the SA manual) - wonderful targets for this of of setup with nice intense wide emission lines. They look lovely in colour too with a DSLR
Also PCygni which shows He and H in emission and Gamma Cas (H alpha emission) are a nice bright easy beginners emision line targets
but I agree on starting with a brightish main sequence Av star. You will need this to do your wavelength calibration. If you have trouble with this then go to an M star which will show easy to capture broad molecular absorption bands
Robin
Just a quick word of caution about Wolf-Rayet stars. Their spectra can have variablility... but to be fair the probably can be said for AGN. For both most changes are not going to greatly impact a low resolution spectrum like you get from a grating. I'd say that one should keep it in mind just in case either a WR or an AGN gives an odd or unexpected result for your calibration.
Another challenging thing about WR stars is that their spectra can be much more complex than an AGN or PN. Particularly at low resolution closely spaced spectral lines can blend together and influence each other. Again, I would put this in the catagory of be aware of it in case a WR gives an odd or unexpected result for a wavelength calibration.
Sorry I misunderstood. I thought we were looking for interesting objects as first targets. I would not recommend WR stars for calibration. AGN are not useful for calibration either as they are faint, embedded in galaxies and most of those that are condensed enough to give narrow lines will have sigificant redshifts.
For calibration in wavelength (and in flux) Balmer lines in a main sequence A star plus the zero order is all you need - tried and tested :-)
Quite right. When in doubt, RTFM!
The manuals for the SA devices are here: http://www.patonhawksley.co.uk/resources.html
George
IMHO keep things as
KISS.....
IMHO keep things as simple as possible at the beginning - there will be plenty of challenges as you move forward.
Just practice with a brighter A type star until you have confirmed your dispersion, resolution and focusing techniques......
See the second half of this presentation for the outline of how to produce reduced spectra from Star Analyser images. (Note the comment on the planned AAVSO database from 6 years ago !)
http://www.threehillsobservatory.co.uk/astro/spectroscopy_10.htm
slides 31-48
The software used was IRIS and Visual Spec but the steps are the same regardless of the software
I prepared this presentation quite a while ago now but the procedure currenty used has not changed siginificantly except for the use of actual standard star spectra eg from the Miles database rather than generic ones from the Pickles library which gives a more accurate instrument response correction. Atmospheric extinction ismost easily approximately corrected for by chosing standard stars close in altitude to the target which is sufficient for most applications.
Robin
Over the past couple of years using stars in the Miles database of spectra as calibration standards has proved very effective and has become pretty much the de facto standard procedure for amateurs producing low resolution spectra for Pro-Am work as in this recent very successful international campaign for example.
http://www.astrosurf.com/aras/novae/Nova2013Del.html
These stars also work well with the Star Analyser and the spectra of stars from the database suitable for calibration (A/B stars) are now included in the ISIS data reduction program. Francois Teyssier has also produced a larger list of main sequence A/B stars with low interstellar extinction which can be used in conjunction with generic Pickles spectra where a convenient Miles star may not be available. More details on the ARAS website and forum
Robin
Some links to calibration stars (for both wavelength and flux) currently used by amateurs for low resolution Pro-Am work :-
Paolo Beradi's database of MILES stars with an Excel front end to select stars nearest the target
http://www.spectro-aras.com/forum/viewtopic.php?f=8&t=941
A similar but longer list by Francois Teyssier of low IS extinction A/B stars (used in conjunction with Pickles generic spectra)
http://www.spectro-aras.com/forum/viewtopic.php?f=8&t=680#p2582
The big advantage for people taking up this subject now it is mature is most of the ground work has already been done :-)
What would be useful though is if spectroscopists with frequent photometric skies could produce actual flux calibrated spectra of some of stars from Francois Teyssier's list to supplement the MILES spectra rather than having to rely on them being typical of the Pickles standards. Similar to the database of comparison stars for photometry generated by AAVSO for example
Robin
Most probably, one could use A- or late B-type supergiants as well. They have prominent and quite sharp Halpha lines, not too many other stronger lines to blend hydrogen. Some of them are pretty bright too (Deneb... ;-)), but unfortunately most of them are also low amplitude variable stars with some radial velocity variations. Those changes in radial velocity should be quite insignificant when thinking about typical spectral resolution of both SA-200 and SA-100.
Nice (and maybe even the only "true spectral classification") reference can be found from http://ned.ipac.caltech.edu/level5/Gray/Gray_contents.html, look at luminocity effects at A0 or B5.
Also, as an example, an old(ish) catalog of supergiant stars (probably not the only one) is accessible via Vizier: http://vizier.cfa.harvard.edu/viz-bin/VizieR-3?-source=V/17A&-out.max=50&-out.form=HTML%20Table&-out.add=_r&-out.add=_RAJ,_DEJ&-sort=_r&-oc.form=sexa
Try to search using e.g Vmag: < 6 and Sp: A? I?
From results, click on sequence number and from top of the individual star info table select "Simbad" for more information (e.g. for measured radial velocity).
One might try different streetlamps as well, LPS and Hg lines may be quite useful. Even pro's use them sometimes to fix absolute radial velocity scale.
I have also a question. Has anyone here tried the SA200 with an Apogee FW50 and Apogee Alta U42 camera? I'm mainly interested if SA200 fits into the filter wheel or not.
Best wishes,
Tõnis
Hi Tõnis,
Actually with the Star Analyser, the Balmer lines show up better in main sequence stars than in supergiants. This is because the resolution (~40A at best) is more than the line width so the narrower lines in supergiants appear less intense even though they may have similar Equivalent Width . Beginners are sometimes puzzled to find how much weaker the lines appear in Star Analyser spectra of Deneb for example compared with Vega.
Interesingly I am currently taking part in a Pro-Am campaign measuring Deneb RV variations using high resolution to a precision better than 1km/s. The variations in Deneb are typically +- 5km/s so well below the detectable limit of the Star Analyser. (~5A or ~300km/s with care)
Robin
good to know that.
Hi Robin,
good to know that. I played around with a Grating to sensor calculator (http://www.patonhawksley.co.uk/calculator/) and found that ~30mm grating to sensor distance and 9um pixels are indeed not very far what you said. And when being conservative about sampling - ca 40 A per ca 2.+ pixels it is...
Just because of curiosity - do you know what is the typical wavelength determination uncertainty for system with SA100/200 and stars as calibrators? I have quite a bit of experience with higher resolutions (R=1000 ... 8000) and I have found that if our spectrograph is well focused, using a reasonably low order dispersion function yields typically 1/20 .. 1/40 pixel RMS. For comparison, I measured just today 15 R=2500 spectra of an eclipsing binary and got the 1 sigma wavelength scatter from spectra to spectra (when measuring a strong and narrow interstellar line) about +-1.1 km/s. Resolution in average was about 0.46 A/pix, at 6613A it is about 20 km/s => RMS was ~1/20 pixel..
Tõnis
Hi Tõnis,
With the grating in the converging beam you cannot simply calculate the resolution based on dispersion and nyquist sampling or star image/slit size like with a classical collimated spectrograph design because the converging beam produces aberrations. This means in practise whatever you do the resolution will be around 40A at best even if you keep increasing the dispersion by increasing the spacing or by increasing the line density of the grating. The calculator warns you if you do not have enough dispersion to overcome the limitations of star image or pixel size to give a reasonable resolution or if you are using too much dispersion which will just give a fainter spectrum but no increase in resolution because of the converging beam aberrations.
There is more background about the limits to resolution considered by the calculator here
http://www.threehillsobservatory.co.uk/astro/spectroscopy_16.htm
and a theoretical treatment of the converging beam configuration by Christian Buil here
http://www.astrosurf.com/buil/us/spe1/spectro1.htm
Cheers
Robin
The gratings work in
Tonis,
The gratings work in "slitless mode", so the spectrum formed is defined by the size and shape of the target.
They really need "star like" objects.
Street lights would have to be really far away to present a small enought "star" to the grating.
Just go for an A type star....then look at other "opportunities"
Ken, thanks for clarifying, that is an very important point indeed, no matter what kind of "local" sources to use.
Wavelength uncertainty is a pretty big subject depending on many factors including the instrument stability, target properties and measurement techniques so it is difficult to generalise.
In the amateur domain:-
Christian Buil is measuring uncertainties at the 20m/sec (yes metres!) 3 sigma level for exoplanet detection using an R ~50000 fibre fed echelle spectrograph and cross correlation techniques.
http://www.astrosurf.com/buil/exoplanet2/51peg.htm
With my R ~17000 telescope mounted LHIRES spectrograph the uncertainty is ~ 1km/s 3 sigma at best but significantly worse at ~3km/s if I use the internal calibrator, limited by spectrograph stability.
http://spektroskopieforum.vdsastro.de/viewtopic.php?f=28&t=4217#p26185
Slitless systems like the Star Analyser are more problematic as you do not have a stable reference point like slit/ fibre spectrographs. The figure I suggested is 5A uncertainty (~1/10 resolution) and even this would need a good technique and a non linear calibration. Realistically the Star Analyser is not a device for precision wavelength measurement.
Robin
(edited to add link references)