I recorded a spectrum of Cor Caroli (to be used as standard) around H-beta using a LHIRES III with 300l/mm grating.
Since Neon doesn't have usable lines in that range I used an Osram ST111 glow starter (Ar/H/ traces of He) as a calibration source which was mounted in a flip mirror in front of the spectrograph.
Apparently this geometry introduces distortions:
Looking at the raw image of Cor Caroli I find the H-beta line approximately at 582/243 which is in good agreement with the calibration image which also has this line at these coordinates. However, looking at H-gamma the Cor Caroli image has it at 132/243 whereas the line in the calibration image is at 139/243. There *is* a line at 132/243, however, that is not H-beta since the bright line at 145/243 is Ar II 4348.06. Calibrating the calibration image with 4277.53, 4348.06, 4545.05, 4764.87, 4965.08, and 5495.87 using a second order polynomial yields a good fit.
Would anybody have an idea what is going on here?
For reference:
Raw calibration image: https://www.dropbox.com/s/vybom473o11maf3/Calib_Cor_Caroli_Light_30_secs_001.fits?dl=0
Raw Cor Caroli image: https://www.dropbox.com/s/ze2cfkvwna3dygg/Cor_Caroli_Light_003.fits?dl=0
I do not trust Hydrogen lines in lamp starter spectra, they don't always appear and become weaker as the lamp ages. Having said that, calibrating your spectrum using the same Ar lines as you used does give me an ~ 8 Angstrom error in your star H gamma line which appears to be at 4332 A compared with the H beta line which is close to correct at 4860 A.
There was a recent thread on the astronomicalspectroscopy group from somebody who had a similar problem with a lowspec using a similar lamp injection system
https://groups.io/g/astronomicalspectroscopy/message/18145
I made some suggestions there but I am not sure if it was resolved
Since you have a LHIRES, presumably with the old neon lamp, an upgrade to the new RELCO lamp could solve the problem (I don't see any calibration errors with my updated instrument) Alternatively if you take longer exposures with the built in neon lamp you may find some weak Ne or Ar lines you can use as I did here for example before I updated my LHIRES
http://www.threehillsobservatory.co.uk/astro/spectroscopy_12.htm
Cheers
Robin
Thank you Robin,
that confirms that I'm not hunting ghosts but seeing a real effect.
Yes, my LHIRES is from 2007, i.e. it has the old Ne calibration lamp. Life happened, so the spectrograph sat in its box without being used for quite some time.
Using longer exposures on the built-in calibration source is certainly an idea worth looking into. I could then do direct comparisons between calibrations from the built-in source vs. the starter.
One other thing which comes to mind: I didn't pay too much attention to the distance between spectrograph and camera sensor (I'm using an ST10XME). The setup I'm using brings the sensor to a distance of approx. 38mm to the spectrometer. Could that be the root cause of the distortion? The manual mentions that the spectrometer supports DSLRs which would imply a 'standard' distance of 55mm and thinking of the geometry of a Littrow spectrograph this could well be critical.
Thanks & regards,
Marcus
Hi Marcus
"One other thing which comes to mind: I didn't pay too much attention to the distance between spectrograph and camera sensor (I'm using an ST10XME). The setup I'm using brings the sensor to a distance of approx. 38mm to the spectrometer. Could that be the root cause of the distortion? "
The distance to the camera sensor is indeed critical for best performance as it defines the collimation of the beam at the grating. This should ideally be correct within a mm or two (my kit build LHIRES came with a set of adapters for various cameras.) Mis-collimation can produce aberrations including distortions of the lines near the edges of the field so might possibly cause calibration errors. Having said the lines in the lamp spectrum image don't look too bad to me. The star spectrum is very broad though, as though the telescope focus is significantly off. I don't think I would not expect that to cause a calibration error though.
Christian Buil has some notes on the effect of mis-collimation and how to precisely tune the LHIRES on his website here
http://astrosurf.com/aras/lhires_eval/optique.htm
and the drawings showing the correct distance of the sensor from spectrograph and the various adapters which came with the kit built version can be found here
http://www.astrosurf.com/thizy/lhires3/index-en.html
specifically this one showing a distance of 54.85m from the rear flange so your so 38mm would be a long way out.
http://www.astrosurf.com/thizy/lhires3/d_Bague-T%20Montage%20CCD-3%20-%20Feuille.jpg
Cheers
Robin
Hi Robin,
the broad spectrum is due to me struggling with guiding. I find it rather difficult most of the time to find a guiding star with my C11. The fact that I am observing from a city doesn't exactly help either. The visual limit on a good day is maybe 4m.
Next thing to try is setting the camera sensor at the correct distance and acquiring spectra of a star and both the internal lamp and the starter. I'll post the outcome - this might take some time with the November cloud soup slowly moving in ;-)
Cheers,
Marcus
Which orientation is the slit with respect to right ascension? I have my L200 set up with right ascension at right angles to the slit and declination axis is parallel to the slit. most guiding inputs should be in right ascension, due to periodic error, unless if your polar alignment is poor in which case there will be declination inputs in one direction. Keeping at least the dec guiding inputs minimized can help to keep your spectrum trace as narrow as possible, minimizing the number of pixels used, and thereby minimizing all non-target noise sources (read, dark shot, sky background shot).