Forum der Vereinigung der Sternfreunde

Latest spectrum. The difference from the mean of past results is within the scatter of individual observations. (Is it real or a systematic error, possibly a problem with flats?)

Regards
Robin

Hi Robin, great that you continue!!!

Is it really scatter of individual observations? How do you know? Did you justify it by comparing individual spectra during the first period? And if it is flat fielding: How can you justify this and explain why the CIV line shows no scatter at all? And why didn't you identify the line variability as real?

I believe the latter. Attached is a spectal series of a very similar binary system gamma Velorum which I obtained in Chile (the dashed lines are 1 sigma variability values over the lines). Keep in mind that both lines are coming from different regions and behave depending on their excitation energies (CIV is a line of the shock cone and CIII probably from the WR star atmosphere with stochastic variability).

I recommend to send your data to Tony Moffat.

Cheers, Thomas

Hi Thomas,

Thank you for your confidence. Of course my results tell me the differences are real but I need to be sure that I am producing good quality results. This is the first time I have looked for differences in spectra of just 5%. I use Buil's Halogen lamp method for flats but I am not sure that the illumination is perfectly uniform. The LHIRES also lets in stray light.

I have attached all my results at high resolution in this region to date. I was wrong. The latest results are the lowest to date between 5800-5940A

I measured a standard star on each night to get the instrument response so I will look at these next as a cross check to see if there is any systematic errors from night to night.

Best Wishes
Robin

Great work!

Obviously the lowest continuum levels are not varying at all, as expected. In addition I see an intensity increase in both lines from the blue to the pink spectrum. The respective increase is in good agreement with the line strength which hints to a systematic effect (double variability in a double intensity line). However, if we assume a systematic effect then you should be able to explain why the blue line wing of CIV shows an excess variability compared to its red wing. I believe that the systematic glitch can be seen in the narrow absorptions around 5890. But that is significantly less then inside the CIV line. I still believe its real variability and Tony will tell you what you really see. It should be some stochastic stuff from wind clumping of the WR component. Talk to him. But I would be very surprised if your stuff is not real!

I am very eager to talk to you in Heidelberg.

Nice weekend, Thomas

Hi WR140 team,

I have just had some very useful feedback from Tony (below). WR140 is reported stable outside periastron so looks like I have a bit more work to do try to improve repeatability.

Robin

--------------------------------------------------
Hi Robin:

Looks like a residual dark-current or rectification problem to me. Normally this star's WR wind lines
do not change with time (except in the 2-3 months around periastron, and even then only up to ~10%
in some lines). I've seen this before in various spectra in various stars from various observatories,
and it invariably is caused by instrumental effects. One has to be sure to observe always with the same parameters, allow for dark current if necessary and do the same normalization each time. It's a bit tricky
but doable. Another thing we often do is to recalibrate (actually tweak up) the wavelength calibration
by using IS lines in the spectrum itself. This can often improve the RV scatter by a factor of up to 2.

Hope this is useful!

Cheers and lots of courage!

Tony

-------------------------------------------------

Hello Robin,

2 months ago I learned that my flats through the tele and the Lhires are wrong, because about 30% of the intensity was stray light. So I left this method and use only biases for the correction. The dark and cosmics correction automatically is done during the sky background correction .

If you want you can sent me your rough pictures (by email) and I reduct them for comparison purposes.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

Thank you Lothar. I will do that. But first, now I know the spectra should be the same, I am looking for possible sources of small errors (<+-5%).

I have already found two.

1. I measure a standard star each night to calculate instrument response. All the results use Vega except 29 Dec 07 which used Altair, I see a small (few %) difference in the Altair calculated response. I need to understand why.

2. For the latest result I havestarted using a remote computer in the observatory. It an old slow computer and it is having to work hard guiding and downloading the images. The slower downloads are showing some amplifier glow in some images. The background sky subtraction should get rid of it but there may be a problem.

There is lots to check when we get down to this level. If the variations at periastron are only 10% we will need to be consistent at the +-2% level at least to see anything. This is going to be tough!

Cheers
Robin

OK I understand this one. I changed from an ordinary lamp to a Halogen lamp for the flats after 29 Dec. This of course affected the measured instrument response but will not change the final result.

I have corrected for this and plotted the differences between instrument responses for each date relative to the mean. (attached) The differences are only +-1.5% This test most parts of the acquisition and data reduction (for a bright star) so I am happy

Next I need to find if there are any problems with dark/background subtraction with the much fainter WR140 results


Robin

Hi Lothar!
Here I am surprised! The flat eliminates local response variations to aplly a proper rectification of the continuum. I do not believe it to be a good idea to neglect the flat.

Cheers, Thomas

Hi Robin!
Oh yes, it will be tough indeed. Especially because 3% wavelength dependend deviations are not acceptable. The procedure, including flats, needs to be repeatable. So, after learning your very last response functions there is no wonder that we see line variations. This whole business needs to be clarified for our WR campaign.

Keep going my friend and send your input. We need it for Tenerife!

Bye, Thomas

But not big enough on their own to explain all the difference (+-4% between 29jan and 15 apr)

The hunt continues!

Robin

Hallo,

es kommen erste signifikante blobs (variabel ?) zum Vorschein: vgl. die Strukturen um 5700 A (Spektrum bitte auf dem Bildschirm vergrößern).

Technische Details der Aufnahmen:
C14, Lhires III, 40 um slit, 1200 g/mm grating, 4x20min. = 80 min Belichtungszeit, nicht normiert, kalibriert mit 4 Neonlinien am rechten Rand -> Kalibrierung ungenau (da kommt es im Moment nicht darauf an). S/N = 170 [5970:5990].
FWHM der Na-Linien (D, 5890 und 5896) 1.52 A -> R >= 3875.

Spektrum stelle ich in unsere Datenbank ein.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

Super, Lothar, das sind zweifelsohne erste deutliche Variationen, bin mir da ganz sicher, aber - wie sagt Thomas immer?- :"eines ist mir sonnenklar, nämlich das, dass ich fast nichts weiß"..... hast Du das mit SMS ausgewertet?

Berthold

Hi Berthold,
nicht mit sms. Das kam nicht damit zu recht (es liest die erste Aufnahme nicht richtig ein, da ist noch ein bug oder Bedienungsfehler vorhanden). Ich habe es mit den Skripten von Otmar ausgewertet (extflat, calneon und extract). Abzug von cosmics, dark und Himmelshintergrund innerhalb des Skripts. Kein flatfielding. Keine Kontinuumnormierung.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

nice indeed!

Berthold