V1369 Cen

Information about outbursts of eruptive stars, Be activity, ...
Ken Harrison
Posts: 197
Joined: Thu Oct 25, 2012 5:45 am
Location: St Leonards, Australia
Contact:

Re: V1369 Cen

Post by Ken Harrison »

Robin,
The data I presented is not comparable with Terry's.
The composite graph has a 100 unit off-set applied between days to improve the clarity.
BASS (by default) generally shows the ADU values as the Y-axis. Obviously these can vary with conditions etc.

I do have a question re presentation...I don't apply flats/ instrument corrections at the moment. If I apply a flat then we end up with a spectrum, normalised to the flat spectrum (Appenzeller, p166) How is this then "corrected" - or is the assumption that the "flat spectrum" is flat- across the wavelengths imaged?
"Astronomical Spectroscopy - The Final Frontier" - to boldly go where few amateurs have gone before....
"Imaging Sunlight - Using a digital Spectroheliograph" - Springer
http://www.astronomicalspectroscopy.com
Robin Leadbeater
Posts: 1926
Joined: Mon Sep 26, 2011 4:41 pm
Contact:

Re: V1369 Cen

Post by Robin Leadbeater »

Hi Ken,

As you say values in ADU have no physical significance. In this case it is standard practise to "normalise" (actually scale) the spectra to 1 at some specified wavelength in the continuum (as Terry's second plot shows) . Your spectra from different days can then be compared directly.

Since you have not corrected for instrument response, it would probably be advisable to take this a step further and normalise the spectrum to the continuum (ie make the continuum 1 everywhere by dividing by the smoothed continuum.) They could then be compared directly with other similarly normalised spectra from other sources.

Spectroscopic flats are fascinating things as they contain both the usual spacial (eg dust, vignetting, pixel/pixel) variations and wavelength dependent (instrument optics/camera response, flat lamp spectrum) information. If the wavelength range is small, eg at high resolution where the shape of the continuum is not of interest it is often sufficient just to normalise the spectrum to the continuum, effectively removing the effect of the instrument response, flat lamp spectrum etc but it is more usual however to make an "instrument response" correction using a reference star which takes care of all these wavelength dependent effects (actually mainly a correction for the flat lamp spectrum as the act of dividing by the flat removes the true instrument response)

A hot star with a known spectrum is used as a reference (eg for low resolution work like this, a Miles standard or failing that a star of known spectral type and low interstellar reddening using the Pickles spectrum as an approximation) The reference star is recorded on the same night under the same conditions and the result divided by the actual spectrum for that star. The result is then used to correct the target spectrum. (This normally also includes correcting for atmospheric extinction, either by choosing a reference star at the same altitude or by measuring the reference star at different altitudes and making a correction to the altitude of the target.

Provided you can get a good reliable measure of the actual signal from the target and reference stars (not trivial with a narrow slit spectrograph) you can even go a step further and measure the actual rather than relative flux in the target spectrum.

That is it in a nutshell of course. Christian's ISIS tutorials and additional information here are is good place to see these process steps in action.
http://www.astrosurf.com/buil/isis/isis_en.htm

Cheers
Robin
LHIRES III #29 ATIK314 ALPY 600/200 ATIK428 Star Analyser 100/200 C11 EQ6
http://www.threehillsobservatory.co.uk
Ken Harrison
Posts: 197
Joined: Thu Oct 25, 2012 5:45 am
Location: St Leonards, Australia
Contact:

Re: V1369 Cen

Post by Ken Harrison »

Robin,
Thanks.
I understand the normalisation etc, but:
Quote
it is more usual however to make an "instrument response" correction using a reference star which takes care of all these wavelength dependent effects (actually mainly a correction for the flat lamp spectrum as the act of dividing by the flat removes the true instrument response)

As the target spectrum and the flat spectrum both contain an "instrument response", when divided we just end up with the "actual" spectrum, normalised to what ever the flat lamp spectrum was. Yes?
My question was...how is this "flat response" actual removed from the target spectrum? Do I generate an instrument response (Miles etc) then divide the original flat by the instrument response?
Just curious.....
Yes, the time will come soon when we can do a back to back comparison between BASS and the other available processing software.........
"Astronomical Spectroscopy - The Final Frontier" - to boldly go where few amateurs have gone before....
"Imaging Sunlight - Using a digital Spectroheliograph" - Springer
http://www.astronomicalspectroscopy.com
Ken Harrison
Posts: 197
Joined: Thu Oct 25, 2012 5:45 am
Location: St Leonards, Australia
Contact:

Re: V1369 Cen

Post by Ken Harrison »

http://gabalou.canalblog.com/archives/2 ... 07263.html
This shows the type of "flat spectrum" I'm thinking/ concerned about......
"Astronomical Spectroscopy - The Final Frontier" - to boldly go where few amateurs have gone before....
"Imaging Sunlight - Using a digital Spectroheliograph" - Springer
http://www.astronomicalspectroscopy.com
Malcolm Locke
Posts: 25
Joined: Mon Dec 16, 2013 11:27 am

Re: V1369 Cen

Post by Malcolm Locke »

Belated SA100 results from the 3rd and 4th of Jan attached. Hb line is very faint on the 3rd.
v1369_cen.20140103.png
v1369_cen.20140103.png (50.49 KiB) Viewed 12976 times
v1369_cen.20140104.png
v1369_cen.20140104.png (51.06 KiB) Viewed 12976 times
Malcolm Locke
Posts: 25
Joined: Mon Dec 16, 2013 11:27 am

Re: V1369 Cen

Post by Malcolm Locke »

Steve has recommended we plot against velocity rather than wavelength to reveal structures and changes.

Here is an overplot of Ha, Hb, Hg, Hd, FeII 4923, FeII 5018 and FeII 5169 from Terry Bohlsen's spectra of Dec 28th.
v1369_cen.velocities.20131228.tbohlsen.png
v1369_cen.velocities.20131228.tbohlsen.png (16.37 KiB) Viewed 12973 times
And the same treatment from 9 days later on the 5th Jan.
v1369_cen.velocities.20140105.tbohlsen.png
v1369_cen.velocities.20140105.tbohlsen.png (15.28 KiB) Viewed 12973 times
There is a considerable increase in velocity for the two absorption features between the two dates. There also seems to be a slight trend of lower velocity for lower wavelength.

Here is the Ha line with the absorption minima velocities marked. The 25th Dec:
v1369_cen.ha_velocity.20131228.tbohlsen.png
v1369_cen.ha_velocity.20131228.tbohlsen.png (9.7 KiB) Viewed 12973 times
And the 5th Jan:
v1369_cen.ha_velocity.20140105.tbohlsen.png
v1369_cen.ha_velocity.20140105.tbohlsen.png (9.45 KiB) Viewed 12973 times
Thanks to Terry for submitting this data to the ARAS database.
Robin Leadbeater
Posts: 1926
Joined: Mon Sep 26, 2011 4:41 pm
Contact:

Re: V1369 Cen

Post by Robin Leadbeater »

Ken Harrison wrote:Robin,

As the target spectrum and the flat spectrum both contain an "instrument response", when divided we just end up with the "actual" spectrum, normalised to what ever the flat lamp spectrum was. Yes?
My question was...how is this "flat response" actual removed from the target spectrum? Do I generate an instrument response (Miles etc) then divide the original flat by the instrument response?
Just curious.....
There no need to correct the flat for the lamp spectrum. Provided both the target and reference spectrum are flat corrected they both contain the lamp spectrum effects. All wavelength dependent effects of the flat lamp source are therefore removed when you divide the target spectrum by the "instrument response".

Note that although not really necessary the flat can be normalised in the X direction removing the broad variation in the wavelength axis, (or as is the case in ISIS the flat is divided by the nominal black body curve for the lamp temperature.) This is just a processing aid though which roughly removes the effect of the flat lamp spectrum to reduce the dynamic range of the flat correction calculations during processing. It does not remove the need to do a true instrument response correction using a reference star. (Note it is important that the flats used for target and reference are processed in the same way using the same normalisation and the lamp spectrum had not changed.)

There are three considerations when choosing a flat lamp source. Is the spectrum stable with time? (It needs to be constant between the flats taken for both target and reference) Does it have enough flux at all wavelengths covered? and is it free of fast changes in output with wavelength which would be difficult to follow when producing the instrument response? Thierry's el panel looks ok in this respect except at the IR where the output drops to zero. Halogen lamps are also ok but can give problems at UV where the output is low. (It is important to generate a flat with very low noise across the wavelength range to avoid adding noise in the flat correction)

All this begs the question - if all the flat field defects we are trying to eliminate are in both the target and the reference raw spectra then why bother to do a flat at all since they will divide out? Indeed, if we could guarantee that the two spectra covered exactly the same pixels and if we could produce a satisfactory instrument response including small scale variations (ie without smoothing which masks any small scale defects) there would be no need to use flats. This is difficult/impossible to achieve perfectly in practise though. I have however resorted to doing this with some success on occasion where I have had problems with the quality of the flat correction. Indeed I have seen references to "Star Flats" in professional literature where a bright hot star is scanned down the slit to produce a flat from starlight.

Cheers
Robin
LHIRES III #29 ATIK314 ALPY 600/200 ATIK428 Star Analyser 100/200 C11 EQ6
http://www.threehillsobservatory.co.uk
Terry Bohlsen
Posts: 191
Joined: Thu Sep 29, 2011 7:40 am

Re: V1369 Cen

Post by Terry Bohlsen »

They are great graphs Malcolm. I don't know how to make similar ones as I aven't nutted through IRAF.
Very good.
Terry
Terry Bohlsen
Armidale NSW
Australia
Terry Bohlsen
Posts: 191
Joined: Thu Sep 29, 2011 7:40 am

Re: V1369 Cen

Post by Terry Bohlsen »

Robin
I agree about the use of flats. I have always used them when I was using my ST8 camera as there was a sinusoidal shaped variation over the frame that was hard to remove with an instrument response division.
With my new Atik camera the flat frame is very uniform and I have found that the flat frame adds noise rather than reducing it. I use the tungsten lamp in my LISA to make flat frames. It is brightest in the green and dim in the blue and UV. To stop the green area saturating I have to have low signal in the blue and this is not ideal.
I have stopped using flats with the new camera. There are no dust bunnies on my camera to flat frame out.
With the old camera I would still use flats.

Terry
Terry Bohlsen
Armidale NSW
Australia
Tasso Napoleao
Posts: 21
Joined: Wed Dec 11, 2013 5:47 pm
Location: São Paulo, Brazil

Re: V1369 Cen

Post by Tasso Napoleao »

Hi friends,

Weather has not been favourable for us over the last weeks (summer is usually very hot and wet in this part of Brazil). Below please find what could be done in the period:
131221_novacentauri_600s.png
131221_novacentauri_600s.png (20.38 KiB) Viewed 12936 times
140103_novacentauri_600s.png
140103_novacentauri_600s.png (20.25 KiB) Viewed 12936 times
Spectra of V1369 Cen taken on Dec 21, 2013 and Jan 03, 2014, by Rogerio Marcon (acq´n) and Tasso Napoleao (red´n). Exposure 600s. Observing site: Campinas, SP, Brazil (22.792271 S, 47.070032 W, height 597m). Spectroscope R=11 000 (grating 2400 l/mm), Newtonian scope 315mm, Audine camera (KAF 1602), all equipment built by R. Marcon. Pre-treatment and reduction made with Maxim DL and VSpec 4.1.2.

Our Jan 3 profile shows a kind of P-Cygni dip which appears to be the same already visible on Ken´s and Terry´s spectra taken on late December / early January.

H-Alpha has apparently been weakening gradually since the last spectra we had taken (Dec 19).

Cheers,
Tasso.
Post Reply