Calibration ISIS dans le bleu - probleme

Design, construction, tuning of spectroscopes
Information and discussion about softwares (telescope remote, autoguiding, acquisition, spectral processing ...)

Calibration ISIS dans le bleu - probleme

Postby etienne bertrand » Mon Jun 21, 2021 2:20 pm

Bonjour,

J'utilise depuis pas mal de temps avec ISIS pour la calibration ; la loi de dispersion qui consiste à prendre une étoile A/B pour calibrer dans le bleu (UV) sur les raies de Balmer + boitier de calibration avec une lampe argon/néon.

En essayant par hasard de comparer le spectre du Soleil depuis la méthode de la loi de dispersion d'un coté, puis du module de calibration de l'autre juste avec la lampe argon/néon je trouve en résultat que le spectre calibré uniquement avec le module de calibration est plus précis dans le bleu que le spectre calibré avec la loi de dispersion (étoile UV + lampe argon/néon).

Comment cela se fait ?

Ci-dessous les deux spectres dans l'UV.
Attachments
soleil_modulecalib-loidedispertion_c.jpg
etienne bertrand
 
Posts: 996
Joined: Mon Aug 31, 2015 11:26 am

Re: Calibration ISIS dans le bleu - probleme

Postby etienne bertrand » Mon Jun 21, 2021 2:58 pm

J'ai essayé avec le logiciel de Benji SpcAudace et là c'est parfait dans le bleu.

Spc Audace marche très bien le seul soucis pour moi c'est quand je fais des spectres avec très très peu de signal Spc Audace me laisse des cosmiques, mais sinon ça marche très bien.
ISIS ne pourrait pas utiliser la même recherche que Spc Audace pour calibrer dans le bleu ?
Attachments
calibre_spcaudace_soleile_G2V.png
etienne bertrand
 
Posts: 996
Joined: Mon Aug 31, 2015 11:26 am

Re: Calibration ISIS dans le bleu - probleme

Postby Peter Somogyi » Mon Jun 21, 2021 3:52 pm

Etienne,

At the time when UV is correct, can you also please check the green and IR-end regions? (preferably with a target star, using eShel spectra for a comparison)
Green range is hard to check without target star, because calibration lines are weak and blended... so we won't even get a proper RMS of the fit...

I've made ISIS correct in the UV with my Alpy 600 (with trial & error lucky choice of lines and order), but still see a systematic skew of 0.2-0.3A in the green (requires emission of a real star e.g. symbiotic with HeI lines to be measured via EW tool e.g. in ISIS). This skew is lower than the true RMS of the fit, however I just know it is systematic and not a noise-like (contrary to other spectroscopes practice with shorter wavelength ranges).
Problem is, fitting a single polynomial curve is not quite good for this kind of spectrograph (high gradient of fitting curve in the blue and deviating more in UV, but close to linear in green-red range, then IR part going away again... splines or distinct curve fits could work much better). We'd need proper software support for a perfect use, however we're just observing amateurs and not perfectionist developers...

Balmer end of another star could also make the UV more correct (having not to extrapolate for only CaII lines, but also till Balmer end down to 3600A where I still get signal.... your graph starts only from 3800A, and I know getting calibration line below 4000A is very tricky). But with 1 single polynom, making UV perfect causes green part go away again. Spline fit would be better.

So the luck which lines are picked is playing a lot here (even wrongly, e.g. chosing blended ones). I don't know which method SpcAudace is using and which of the many methods you picked from ISIS (I use file method, specifying the lines and the order). Automatic methods will just do something, with a matter of luck how the spectrum is accepted at consumer end.

Peter
Peter Somogyi
 
Posts: 419
Joined: Sun Jul 13, 2014 8:56 am


Return to Spectroscopes and Softwares

Who is online

Users browsing this forum: No registered users and 8 guests

x