Page 23 of 26

Re: PNV J20233073+2046041 mag 6.8

Posted: Sat Sep 21, 2013 12:11 pm
by Olivier GARDE
Here's the main evolution of some few lines between the 9th to 21 th september
Voici les évolutions majeurs de quelques raies de la nova entre le 9 et le 21 septembre.

Image

Re: PNV J20233073+2046041 mag 6.8

Posted: Sat Sep 21, 2013 12:50 pm
by Christian Buil
5-days evolutions in flux unit. Helium lines are coming !
(warning, log scale in intensity). Olivier soectra shows details.

Image

The Sept. 20 spectrum is taken from the center of Paris (Valerie's Saint Charles Observatory) with an Alpy and a C9.25

Measured magnitude for Sep. 20.87 :

B = 8.063
V = 7.991
R = 6.454

Christian

Re: PNV J20233073+2046041 mag 6.8

Posted: Tue Sep 24, 2013 9:51 am
by Jacques Montier
Bonjour à tous,

La raie Halpha de la nuit du 23 septembre.
C11 + Lhires 2400 t/mm ; résolution 13825.
_nova_del_20130923_829_Jacques_Montier.png
_nova_del_20130923_829_Jacques_Montier.png (5.11 KiB) Viewed 12805 times
Son évolution depuis le 20 août.
evolution_Halpha.png
evolution_Halpha.png (12.93 KiB) Viewed 12804 times

Re: PNV J20233073+2046041 mag 6.8

Posted: Wed Sep 25, 2013 2:06 am
by Christian Buil
Très bonne séquence haute résolution Jacques !

De mon coté, la nuit dernière, un spectre Alpy 600 fait en deux parties :

- partie visible 49 poses de 4 secondes
- partie infrarouge (avec filre RG630), 6 poses de 180 secondes.

La fusion des deux parties est faite avec la commande L_MERGE de ISIS et le tout
est étalonné en flux absolu :

A Vis + IR spectrum taken into part (49 x 4 s. for visible and 6 x 180 s for infrared (an order RG630 filter is added))
+ flux calibrated :


Image

Une première tentative pour étalonner en flux absolue un spectre eShel (en bleu) avec l'aide d'un spectre Alpy 600 - en rouge
(voir le graphe précédent). Bien sur les spectres sont faits à peu de temps d'intervalle :

Flux calibration of an eShel spectrum by using an Alpy 600 spectrophotometric spectrum:

Image

De nouvelles mesures de magnitudes :

Image

Je soupsonne que la raie Halpha commence à baisser d'éclat (en intensité absolue).

Christian

Re: PNV J20233073+2046041 mag 6.8

Posted: Wed Sep 25, 2013 3:16 pm
by Christian Buil
Forts chanegement autour de Hbeta en 24 heures (spectres échelles étalonnés en flux avec
des données provenant d'un Alpy600) :

Image

Du coté Halpha l'aile rouge est très stable alors que l'aile bleu s'écroule très vite
(le fait d'être en flux absolu permet de bien mesurer ces phénomènes dans être perturbé
par une référence flottante (le continuum) :

Image

(noter que les raies telluriques ont été retirés avec l'outil H2O de ISIS)

Enfin la raie de l'hélium à 5875 A, stable en comparaison :

Image

La baisse d'éclat s'accelère un peu (mesures spectrophotoémtriques Alpy) :

Image

Christian

Variations HBeta avec l'Alpy

Posted: Thu Sep 26, 2013 7:13 am
by Dubreuil Pierre
Bonjour Christian

J'ai pris les mêmes instants avec l'Alpy , on distingue aussi ces variations !
L'Alpy fonctionne super bien :-)
_nova_del_HBeta_pierre dubreuil.png
_nova_del_HBeta_pierre dubreuil.png (12.07 KiB) Viewed 12771 times
cordialement

Pierre

flux calibrated animation

Posted: Fri Sep 27, 2013 6:53 pm
by Martin Dubs
Hello all,

a while ago I suggested to look at H-alpha intensity in flux calibrated spectra. This has been done in the meantime by Robin. Christian has advanced the technique of absolute flux calibration to a new standard. If no flux calibrated spectra are available one can do an absolute flux calibration if the magnitude mV is known.
The procedure is the reverse as described by Christian for the determination of magnitude from a flux calibrated spectrum:
http://www.astrosurf.com/buil/calibrati ... ration.htm
With the help of ISIS the calibration is actually quite simple.
- Load a spectrum calibrated in relative flux.If a reference star nearby has been used for the instrument response this eliminates also the atmospheric extinction.

- Multiply it with Bessell V response function. This closely resembles the function of a Johnson V filter which is used by AAVSO

- Integrate the relative flux with the FWHM tool in ISIS. The spectral range of the spectrum should be from 4700 to 7300 A.
I0 = integral[Bessell_V(lambda)*F0(lambda)*dlambda] is the integrated relative flux:
calculating I0
calculating I0
flux calculation3.PNG (71.09 KiB) Viewed 12744 times
- Calculate the calibration factor with the following equation:
F(lambda) = F0(lambda)/integral[Bessell_V(lambda)*F0(lambda)*dlambda]*10^[- 0.4*(mV + 13.70)]
Here F0 is the relative flux
The integral has been computed in the step before
mV is the measured V magnitude taken from e.g. AAVSO database or an independent measurement
13.70 is a constant to convert the magnitude to the flux in the Landolt system. It was determined by comparing the flux and magnitude of stars in the Calspec database. It is valid for all spectral classes.

- Reload the original spectrum and multiply it with the calibration factor C = 1/I0*10^[- 0.4*(mV + 13.70)]

- Save the spectrum with _flux appended.

It would be nice if this sequence of operations could be implemented in ISIS (There is a flux calculation from magnitude V, by taking the value of the spectrum at 5556 A, but this is only valid for regular spectra, not for Novae or emission line spectra etc) :idea:

For the following animation, spectra of ALPY and LISA spectrographs from the of the Nova database:
http://www.astrosurf.com/aras/Aras_Data ... l-2013.htm
have been used, with spectra from Olivier Thizy, who recorded the first days continuously, supplemented with spectra from Jim Edlin and Francois Teyssier for the gaps, with spacings of one day in the beginning and 2 to 3 days in the later phases.
Notice that spectra taken with different instruments, different observers, different weather conditions have been used, so some imperfections are visible. But despite this it shows that the method works. The magnitudes mV have been obtained by eye from averaging the AAVSO observations over one day bins, rounded to 1/10 mag. Of course the absolute flux is only as good as these measurements. The relatively large scatter of individual AAVSO measurements may come from the fact that not everybody uses the correct CCD filter combinations, which is not so important for regular stars but especially important in the case of strong emission lines like H-alpha.
Of course it would be an advantage if all measurements would be taken with the same equipment, same reference star and same photometric equipment with correct spectral response.
These files were renamed and animated with the ISIS animation tool.
The use of a log scale has the following advantages:
- It allows to show large intensity variations over three orders of magnitude (especially if spectra will be added with a decrease of Nova intensity).
- Both the continuum and the H-alpha peak are shown with good detail.
- a constant S/N shows as a constant amplitude, irrespective of intensity of the signal.
- the ratio of line intensity is shown as a constant difference in height. This is particularly well visible for the Balmer series in the later stages, but also for other line groups.
animated sequence of flux calibrated spectra of Nova Del 2013, log scale, corrected!!!
animated sequence of flux calibrated spectra of Nova Del 2013, log scale, corrected!!!
@NovaDel_long2_.gif (576.22 KiB) Viewed 12728 times
The log scale was produced by modifying std.gnu:
set ylabel "flux [erg/cm^2/sec/A]" # added MD
set grid
set mxtics 5
set mytics 5 # added MD
set grid ytics mytics xtics # added MD
set logscale y # added MD

The details of the spectra and the actual calculation of C can be found in the attached ZIP file.
Once the spectra are flux calibrated, the physics of the Nova can be studied by plottting the flux of different lines as a function of time etc.
The good thing: everything except the calculation of the calibration factor can be done in ISIS, even the log scaling of the flux. Thanks, Christian :)

Regards, Martin

Re: PNV J20233073+2046041 mag 6.8

Posted: Sat Sep 28, 2013 2:12 am
by Francois Teyssier
Ouuaahh !
That's great and very interesting.
But, there's an issue with the scale : the current value of continuum is 10e-12 ers/s/cm2/A (ordre de grandeur)

François

Note également Martin, que je ne suis toujours pas d'accord sur la non-prise en compte des raies d'émission (biais de 20% environ), mais ceci est un autre sujet, qui n'enlève rien à la remarquable qualité de ton travail

Re: PNV J20233073+2046041 mag 6.8

Posted: Sat Sep 28, 2013 7:45 am
by Martin Dubs
Hello Francois,

thanks for pointing out this error to me. What a silly mistake :(
The equation given in the text is correct, but in the EXCEL sheet I misplaced a bracket, so the results were off by a factor 10^(-.6*13.70) = 1/1.6596e8. I have corrected the image in the meantime.
Just to make sure I compared my flux calculation, comparing your measurement of sep20, assuming m8.1, with a direct flux measurement by Christian:
comparison of different flux calibration methods
comparison of different flux calibration methods
comparison Teyssier0920Buil0920_flux.PNG (55.41 KiB) Viewed 12728 times
The difference in the two curves comes probably from an inaccurate estimate of the AAVSO magnitude.
Of course I should have made that comparison before sending the last post ;) .
Wrt to your statement about the inclusion or noninclusion of emission lines I still stick to my opinion. The measured V magnitudes include the emission lines, so I think they also have to be included in the flux calculation. My integral over the weighted relative flux corresponds to a measurement with a Bessell V response (detector sensitivity * filter transmission). What I am not sure about it is what observers at AAVSO actually use for their measurements.

Regards, Martin

Re: PNV J20233073+2046041 mag 6.8

Posted: Sat Sep 28, 2013 9:10 am
by Francois Teyssier
This really excellent Martin.
For tomorrow report !

A propos de la prise en compte des raies d'émission dans la calibration en flux absolu.

Je te propose une "expérience de la pensée"

Un spectre en intensité relative de notre nova A a été normalisé à 1 pour une longueur d'onde donnée, par exemple 6100 A. Nous choisissons un point du "continuum" (ce qui est une difficulté, car le "continuum" est fortement altéré par les raies)
La calibration en flux absolu consiste à déterminer quelle est l'intensité de ce point 6100 en unités physiques (nous utilisons erg/cm2/s/A)
Pour cela, dans la méthode photométrique, nous utilisons la mesure de la bande verte en magnitude (VA), qui est convertie en flux absolu.

Hypothèse : Imaginons une nova B qui ait le même continuum, MAIS AUCUNE RAIE dans la bande verte.
Quelle va être sa magnitude ? La magnitude VB va être un peu plus grande car elle ne correspond qu'au flux du continuum reçu à travers le filtre vert : il manque le flux émis par les raies en émission. VB > VA
Or l'intensité absolue du continuum de nova B au point 6100 est la même que celui de nova A.
A l'issue de notre travail de calibration absolue les continuums doivent être au même niveau d'intensité, ce qui n'est possible que si l'on corrige la magnitude VA des raies d'émission avec VA' = VA + VA_raies.
q.e.d.

Dans notre cas, la contribution des raies d'émissions à la magnitude V (càd à travers le filtre V) est d'environ 20%, ce qui n'est pas négligeable (0.2 magnitudes environ)


Je reprends les références que j'ai déjà cité. Je suppose que si Skopal s'y est repris à deux reprises (2003 et 2007) c'est a dû avoir du mal à se faire entendre ;)

Correction of UBV for emission lines, Skopal, 2003 http://adsabs.harvard.edu/abs/2003BaltA..12..604S

On the effect of emission lines on UBVR photometry, Skopal, 2007, http://adsabs.harvard.edu/abs/2007NewA...12..597S

In many astrophysical applications photometric measurements through the standard U, B, V, R filters are used to analyze radiation in the continuum from stellar
objects
. For example, a diagnostic by the (U − B, B − V )-diagram is frequently applied to compare the observed colour indices to those of the continuum radiation. However, photometric magnitudes represent integrated fluxes that include both the continuum and the line spectrum.
Therefore, a correction for lines has to be applied to obtain photometric flux-points of the true continuum.

The effect on the U, B, V magnitudes of the removal of absorption lines from the spectrum was already introduced by Sandage and Eggen (1959). They found that
the absorption lines cause fainter magnitudes at all wavelengths.
On the other hand the presence of emission lines in the spectral region of the photometric passbands leads
to brighter magnitudes than those of the continuum
. However, it is difficult to quantify this effect. A strong variation of the emission spectrum (e.g. due to outbursts and/or
orbital motion), large differences between individual objects and the complex profile of the true continuum of some interacting binaries (e.g. symbiotic stars, classical
novae) preclude a simple solution. Therefore this problem has been approached only individually and without giving a concept for a general application. Probably a first
more thorough approach was described by Menzies et al. (1982) who corrected their V magnitudes for strong emission lines in the spectrum of AR Pav. Other authors briefly
reported just their results on the emission lines effect to evaluate the continuum level in the optical region (e.g. Fernandez-Castro et al. , 1995; Tomov et al. , 2003)
. Recently Skopal (2003) suggested an accurate calculation of ∆m corrections for the UBV passbands

Corrections ∆m can be used to convert arbitrary flux
units of an emission-line spectrum to fluxes in absolute
units with the aid of the simultaneous UBV R photometry.
The following steps are relevant.
(i) Appropriate ∆m can be estimated according to
Eq. (6) to obtain magnitudes of the line-removed continuum.




Description of the method by Menzies for AR Pav
Tomov.PNG
Tomov.PNG (80.88 KiB) Viewed 12725 times
Je ne peux pas en dire plus.

La méthode décrite par Skopal, utilisée par Tomov, Menzies ... est décrite par un exemple de traitement de nova del 2013 sur :
http://www.astronomie-amateur.fr/feuill ... ement.html

François