Alpy and Isis calibration lines

[JP Nougayrede]

Bonjour, Hello,

Je me suis fait une annotation des raies d'intérêt de la lampe de calibration (Relco 480?) du module de calibration Alpy. Je le poste ici à toute fin utile pour les spectros...
L'image (crop) est une médiane de cinq spectres de 30 sec (prétraités) avec une Atik 314, le spectre déduit avec VSpec.
Pour l'annotation je me suis basé sur l'atlas de Richard Walker (table 107) et celui d'Olivier Garde (Lampe du Lisa). En bleu les raies qui sont utilisées par Isis pour la calibration automatique, déduit en lisant les logs du logiciel.
Faites moi savoir si vous repérez des erreurs, et n'hésitez pas à utiliser, modifier etc pour votre usage.

Bon ciel !

I made this annotation of the lines of interest of the calibration lamp (Relco 480?) of the Alpy calibration module. I post it here for any purpose that might be useful for spectros ...
The image (crop) is a median of five 30 sec spectra with an Atik 314, the spectrum was deduced with VSpec.
For the annotation I used the atlas of Richard Walker (table 107) and that of Olivier Garde (Lamp of the Lisa). In blue, the lines that are used by Isis for the automatic calibration, deduced by reading the logs of the software.
Let me know if you spot any errors, and feel free to use, modify etc for your use.

Jean-Philippe

[Francois Teyssier]

Merci Jean-Philippe
Bien utile

François

[Olivier GARDE]

J'avais fait il y a quelques années le même travail sur la lampe Ar/Ne du LISA :
http://www.spectro-aras.com/forum/viewt ... f=8&t=1091
Il serait intéressant de comparer les 2 spectres obtenues, car il y a des différences entre 2 lampes Ar/ne ne provenant pas du même fournisseur ou de la même série.

[Tony Rodda]

Hello,

I used the Ar/He chart posted by Olivier for use with the ALPY calibration module but I found that the Relco 480 (posted by Jean-Philippe) produces better lines at short wavelengths.

Is it possible to use both for calibration purposes? That is, a blend of the Relco and Calibration module lines.

I use ISIS. Is it "Method 3" in the tutorial?

Regards

Tony

[Peter Somogyi]

Tony,

The Alpy 600 is capable of receiving light without focus loss down till 3550A, whilst LISA seem to have the light fall apart somewhere below 4000A.
The other difference is, LISA resolves more lines - not necessarily all suitable for Alpy.

I personally experienced a "good" filelist-based calibration with this (pixel size for an unbinned ATIK 414):

4
3.4789260638297872340425531914893
3946.097
4510.733
4764.865
5400.5618
5764.42
5852.4879
5852.49
5944.83
6266.4950
6506.5281
6598.9529
6752.834
7067.218
7147.0406
7272.936
7383.95

However, this method suffers from a perfect UV calibration - just very close down till Balmer end -, but may be an alternative when ref.star's UV calibration error was too high.
Ideally you want to use the UV ref.star calibration assistant, but I think enforcing the spline fit for the UV part introduces errors on the otherwise linear green-red regions. Currently I accept it as a limitation of this widely preferred process. I also found the red-green region can be calibrated with a much less RMS (like 0.01) at order 2, if you just calibrate that part.

For a verification - as Jean-Philippe did, it's a wise idea to extract the spectrum of the lamp itself (by uncheck "Remove background", and use HFWM tool in ISIS to measure e.g. H-alpha center or any other known line) and also looking at the ref.star both. This is a much better verification than just looking at the RMS.

What I have on mind, to make both UV and other part calibration happy, IMHO the best was to do 2 separate calibrations and make a relax function in between...but this needs wide acceptance and somebody taking up this development.

Cheers,
Peter

PS.: I've found a feature in ISIS file mode, when you write a wavelength line twice or more, seems alike ISIS taking it as weighted... at least it seems so, somebody pls confirm.

EDIT: I often refer mistakenly to spline fit (seen elsewhere)... however, ISIS seem to use polynomial fit. Sorry for my wordings.

[Tony Rodda]

Peter,
Many thanks for your explanation. I will try your method.
Kind Regards
Tony