Forum der Vereinigung der Sternfreunde

Forum of the German Amateur Astronomy Association
It is currently 21. November 2019, 19:06:10 PM

All times are UTC+02:00




Post new topic  Reply to topic  [ 126 posts ]  Go to page Previous 1 2 3
Author Message
 Post subject:
PostPosted: 03. September 2013, 11:07:12 AM 
Offline
Meister
User avatar

Joined: 19. August 2006, 11:45:20 AM
Posts: 2726
Location: 66907 Glan-Münchweiler
Hallo,

hier das weitgehend vom Wasserdampf befreite Spektrum der Nova vom 2.9.2013 um h_alpha.

cheers berthold


Attachments:
Nova_del_ 2.815_09_2013_noh20(sto).fit [11.25 KiB]
Downloaded 189 times
nova_del_ 20130902_noh2o.png
nova_del_ 20130902_noh2o.png [ 6.74 KiB | Viewed 7125 times ]
Top
   
PostPosted: 12. September 2013, 22:08:41 PM 
Offline
Meister
User avatar

Joined: 20. March 2009, 20:34:04 PM
Posts: 257
Location: Unterallgäu bei MM
Hallo,

endlich habe ich mal ein paar Beobachtungen aus dem Blau-Violetten Bereich fertig auswerten können und die folgende gif-Animation daraus gemacht.


Volle Größe, bitte ins Bild klicken!


Image



Viele Grüße,
Torsten


Top
   
 Post subject:
PostPosted: 18. September 2013, 12:49:04 PM 
Offline
Meister
User avatar

Joined: 25. October 2006, 23:43:13 PM
Posts: 740
Location: Cumbria England
Hi Thomas,


Quote:


Absolute flux calibration (as opposed to relative flux calibration) is much tougher and apart from some recent attempts by Christian Buil, I do not know any amateurs doing this. .... and so on
Quote:
Correct! If Christian will ever succsessfully perform I will spend some beer.
Christian is now publishing fully flux calibrated Nova Del spectra using eShel and Alpy 600 spectrographs
http://www.astrosurf.com/aras/Aras_Data ... l-2013.htm
so you might have to start serving the beer :lol:

His methods are described here
http://www.astrosurf.com/buil/calibrati ... ion_en.htm

Cheers
Robin


Top
   
 Post subject:
PostPosted: 18. September 2013, 19:09:16 PM 
Offline
Meister
User avatar

Joined: 31. July 2006, 16:43:32 PM
Posts: 3400
Location: Köln
Hi Robin, you will get a beer from me in any case, even if you are or I am right. I enjoy this discussion.

I checked Christian’s methods. Looks nice and convincing. The crucial point can be found in Step 6: “The spectra of both the target and reference stars should be measured using the same method.”

That is what I always said. You have to observe both stars under the very same conditions. This is basically not possible without a photometric sky. I do not know the conditions at his observing site. But normaly photometric sky conditions can only be found at first-class sites. If the telescope is moved between the stars already slightly different sky conditions change the result (the more the different the conditions – imagine a 90° change). It might well be that both methods match very well but this does not answer how accurate the results really are.

Maybe some other experts can comment this issue because I am interested in additional arguments. And maybe I am even wrong but if these methods are the key for doing such absolute measurement why doesn’t he publish them in a peer reviewed journal? Or has somebody else already done it? I do not know.

The presented procedures do not convince me. I rest my case.

Cheers, Thomas


Top
   
 Post subject:
PostPosted: 20. September 2013, 12:58:42 PM 
Offline
Meister
User avatar

Joined: 25. October 2006, 23:43:13 PM
Posts: 740
Location: Cumbria England
Quote:
It might well be that both methods match very well but this does not answer how accurate the results really are.
Hi Thomas,

Well any measurement (even professional ones) will have an uncertainty of course (and before and after checks on stability of the reference star for example could be used to confirm stability of the atmosphere) but results so far look encouraging, for example comparing AAVSO photometric magnitudes with the magnitudes calculated from the flux calibrated spectra for nova Del 2013
http://www.astrosurf.com/buil/nova_del2 ... hot_v3.png
(buil data)
http://tinyurl.com/p73hzxz
(aavso data)

Cheers
Robin


Top
   
 Post subject:
PostPosted: 20. September 2013, 13:47:28 PM 
Offline
Meister
User avatar

Joined: 31. July 2006, 16:43:32 PM
Posts: 3400
Location: Köln
Hi Robin! The B and V data of Buil are without any spread. What is the error bar? The AAVSO spread is about 1/5mag. How are these measurements done? I am not really sure what you mean here.
Quote:
Well any measurement (even professional ones) will have an uncertainty of course...
That is what I want to say. If no error bar is provided each result is useless. We agree that such measurements are more accurate at professional sites. And we both agree that such tests are encouraging. But what does "encouraging" mean in detail. The Ansatz is interesting but I disagree with the idea that observers at usual amateur sites can deliver high level accurate data comparable to pro sites, especially if this level is never discussed (no error bars = no result). And real flux measurements with ground black bodies are too difficult. This all together is the reason why I wonder why amateurs often blindly use the response function without taking the physics of the target stars (e.g., line widths) nor the measurement goals (e.g., continuum estimation) into consideration. For estimating the target effective temperature on should use relatively accurate line strengths and not a relatively unreliable continuum.

To summarize: Using standard stars are a reasonable approach for getting an idea about the absolute flux. But not taking local sky conditions into consideration degenerates the task. Not delivering reasonable and justified error bars makes the result entirely useless. Unless you are satisfied with rule-of-thump results. Is this the reason why I can't find any amateur publication about this idea...

Again, I am open for any other comments from experienced observers, especially from professionals. The truth can't be found by votings.

Cheers, Thomas


Top
   
 Post subject:
PostPosted: 20. September 2013, 19:57:39 PM 
Offline
Meister
User avatar

Joined: 25. October 2006, 23:43:13 PM
Posts: 740
Location: Cumbria England
Hi Thomas,

I agree about the widespread absence of error bars in amateur work but that is a general problem not specific to the measurement of absolute flux.

Buil's measurements (based on integrating the flux in his spectra within the photometric filter passbands) fall within the spread of AAVSO photometric results and closely follows the same trend. It also correctly predicts the relative V and B magnitude, suggesting that any errors due to variations in atmospheric extinction are not severe. If we reject the hypothesis that Buil is measuring absolute flux to a least around +-10% then we must also reject the value of the photometric measurements submitted to the AAVSO.


Cheers
Robin


Top
   
 Post subject:
PostPosted: 21. September 2013, 01:05:56 AM 
Offline
Meister
User avatar

Joined: 25. October 2006, 23:43:13 PM
Posts: 740
Location: Cumbria England
Quote:
If we reject the hypothesis that Buil is measuring absolute flux to a least around +-10% then we must also reject the value of the photometric measurements submitted to the AAVSO.
In case you are wondering where the +-10% figure came from it is from the residuals of the Buil figures compared with the interpolated trend values of the AAVSO data. It is an only a quick estimate by eye but I suspect it is not far off and gives a reasonable estimate of the uncertainty in Buil's data. I will leave it as an exercise for the reader to calculate the actual figure ;-)

Cheers
Robin


Top
   
 Post subject:
PostPosted: 21. September 2013, 18:20:01 PM 
Offline
Meister
User avatar

Joined: 31. July 2006, 16:43:32 PM
Posts: 3400
Location: Köln
Hi Robin!
Quote:
I agree about the widespread absence of error bars in amateur work but that is a general problem not specific to the measurement of absolute flux.
Yes, this is lastingly annoying.
Quote:
Buil's measurements (based on integrating the flux in his spectra within the photometric filter passbands) fall within the spread of AAVSO photometric results and closely follows the same trend.
Correct! But if Buil does neither deliver error bars his measurements neither can be justified. The spread in the AAVSO data are not a great thing anyway.

Missing errorbars is the overall problem and is maybe the reason for lasting misunderstanding of my statement. Everybody can deliver absolute flux values. They are probably highly inacurate be they are delivered. Great! But does that make sense? The question should be: What do I want to estimate and what accuracy is necessary. But this is mostly not done. Instead I often see a lot of whishfull thinking.

We need errorbars and can then say if the mesurements are ok for the respective goal.

Cheers, Thomas


Top
   
 Post subject:
PostPosted: 23. September 2013, 15:18:22 PM 
Offline
Meister
User avatar

Joined: 19. August 2006, 11:45:20 AM
Posts: 2726
Location: 66907 Glan-Münchweiler
Hi,

here the spectrum of the nova from 130922. Not H20 dried!

cheers
berthold


Attachments:
novadel_130922_norm.fit [11.25 KiB]
Downloaded 171 times
novadel_130922_.png
novadel_130922_.png [ 13.12 KiB | Viewed 6746 times ]
Top
   
PostPosted: 01. October 2013, 02:40:45 AM 
Offline
Nutzer

Joined: 11. December 2012, 19:53:03 PM
Posts: 29
Hi Thomas,

You write:

... I fear the difficulties getting absolute fluxes is highly underestimated in the amateur domain....

... If Christian will ever succsessfully perform I will spend some beer. ...

Attached, comparison of Alpy 600 data and a NOT spectrum of Nova Del 2013, and the comment of Steve Shore.

We have now capacity to calibrate spectrum in absolute flux with a robust (and not very complex) protocol. Cross check with standard stars and some amateurs (François Teyssier with LISA, ...) demonstrate the "professional" quality of the results. We can also extract very consistant BVR magnitude from the actual spectrophotometric data (+/- 0.01 or better). eShel spectra can be also calibrated in flux density with the aim of Alpy 600 data).

Attached also a typical time evolution of the nova in erg/cms2/s/A (flux density) taken with Alpy 600 (equiped with a photimetric slit), eShel, ...

Details about spectrophotometric methods are here:

http://www.astrosurf.com/buil/calibrati ... ion_en.htm

Christian Buil


Attachments:
File comment: Comparison of NOT telescope spectrum (professionnal) and Castanet obs. spectrum (amateur)
NOT_Castanet.png
NOT_Castanet.png [ 30.48 KiB | Viewed 6669 times ]
File comment: Steve Shore comment
NOT_Castanet2.png
NOT_Castanet2.png [ 21.07 KiB | Viewed 6669 times ]
spectrophot2.png
spectrophot2.png [ 79.45 KiB | Viewed 6669 times ]
File comment: eShel flux calibrated spectra (short time evolution)
spectrophot.png
spectrophot.png [ 10.59 KiB | Viewed 6669 times ]
Top
   
PostPosted: 01. October 2013, 03:12:24 AM 
Offline
Nutzer

Joined: 11. December 2012, 19:53:03 PM
Posts: 29
Sorry Thomas, I had not noticed your discussion with Robin.

Attached, a typical example for check the internal consistancy of the measures.
For the date of Sept 4 and Sept 5 the the nova is very stable (see for example AAVSO data). Of course the two spectrum are
indepently taken and no correction is applied. It's two distincts observations.

Concerning abolute values, I have observed some Stsci Calspec standards,
and I recoved the absolue flux with a 10% error or less.

Concerning magnitude estimations look here:

http://www.astrosurf.com/buil/nova_del2 ... ometry.htm

and compare values for a given night (independant observations, of course) and evaluate the dispersion.

The typical time between target objects observation and reference star observation is 1 minute. The air mass is similar. And... I observe from a suburban observatory.

Christian


Attachments:
File comment: Two independant measures give an idea of precision.
spectrophot3.png
spectrophot3.png [ 8.83 KiB | Viewed 6668 times ]
Top
   
 Post subject:
PostPosted: 03. October 2013, 11:44:54 AM 
Offline
Meister
User avatar

Joined: 31. July 2006, 16:43:32 PM
Posts: 3400
Location: Köln
Hi Christian,

Your effort is promising and interesting. Congratulations for the positive reaction from Steve Shore.

Presently I work with a number of astronomers for the European Commission in Brussels and have the opportunity to discuss your procedure. My lasting concern is confirmed by these discussions and by Noel Richardson in Montreal. All colleagues tell me that no flux calibration is reliable unless the sky is perfectly clear and stable. That is what I learned, too. Note, that Steve Shore also mentions the different sky conditions. According to my colleagues (my knowledge is limited here) 10% flux accuracy is not good enough for scientific observations. It should be below 1% level. Noel says that “you need to derive coefficients for your observing site so that you can perform these corrections with only one standard in a night. That can be avoided with a standard star near in sky to your target, but even that can cause problems”. He refers to his 2010 paper on eta Car (http://iopscience.iop.org/1538-3881/139 ... 4_1534.pdf) or 2011 paper on P Cyg (http://iopscience.iop.org/1538-3881/141 ... _4_120.pdf) where he describes the use of “a light curve in conjunction with the analysis, and then simple corrections can be made to the measured emission line fluxes”. You write “The typical time between target objects observation and reference star observation is 1 minute“. So, the time between the observations obviously does matter. That means you understand the sky problem. I suggest you consult more than one professional expert in this field - or even better, you write a paper and submit it to a refereed journal (A&A or ApJ). Then you can obtain real proof for this procedure and certainly get an even better idea about its limitations. If your method claims to be of professional quality, then it MUST be tested in that very community and not only by a small amateur group.

Cheers, Thomas

PS: This thread is actually not the right one for the present discussion. You can also describe your method in the ConVento thread. Some pros are connected via email and might comment.


Top
   
 Post subject:
PostPosted: 03. October 2013, 18:59:57 PM 
Offline
Meister
User avatar

Joined: 25. October 2006, 23:43:13 PM
Posts: 740
Location: Cumbria England
Quote:
According to my colleagues (my knowledge is limited here) 10% flux accuracy is not good enough for scientific observations. It should be below 1% level.
Hi Thomas,

That is an interesting (and I must say rather unscientific) statement. It sounds more like a wish to me. (If a car manufacturer asked me what specification of car they should design I would ask for one which can go 300km/hr, 100km on a litre of fuel but cost only 1000 euro :lol: )

If it were the case that all measurements of astronomical flux need to be made to better than 1% absolute accuracy or 0.01 magnitude (remember we are talking absolute accuracy here not uncertainty in the measurement) the AAVSO for example would have to delete >95% of its database tomorrow (as would quite a few professional ground based observatories I suspect)

The reality of course is that the actual levels of required accuracy (and precision) depend entirely on the reason the measurement was made. If a certain quality of measurement is needed to prove or disprove a particular hypothesis then this is the standard required, no more, no less. Few astronomical papers rely on 1% measurement accuracy to reach their conclusions fortunately :D


Cheers
Robin


Top
   
 Post subject:
PostPosted: 04. October 2013, 09:57:12 AM 
Offline
Meister
User avatar

Joined: 31. July 2006, 16:43:32 PM
Posts: 3400
Location: Köln
Again, Robin, I am not an expert and it might be that 1% accuracy is too harsh for massivce star research. As I already said, it is a matter of the target and physics you want to work on. I am sceptic about the overall task. Again, a refereed paper (or a presentation and discussion on a pro conference, if you wish) could settle the thing. That's it from my side.

Cheers, Thomas


Top
   
PostPosted: 05. October 2013, 12:05:30 PM 
Offline
Nutzer

Joined: 11. December 2012, 19:53:03 PM
Posts: 29
Thomas, the techique I use is not MY technique. This is a simple, logical and nearly standard method, used in professional observatories in the field of spectrophometry , and now by amateurs . Really nothing new and there is no object to plublish in A&A, except for inform amateurs community about the capacity of course.

The only rafinnement I added is to perform an ABABA protocol, with A, the observation of the reference star , and B, the target observation. I can consider also the differential air mass between A and B ( bias correction).

The most important is probably the use of a special wide slit to reduce the seeing and atmospheric chromatism effects. Really sorry Thomas, but all professional telescopes are not equipped with an atmospheric dispersive correction (ADC). I've seen some strange things in some professional databases :?

Finally, I made several checks on Landolt standards (see an example on my page) and I can find the true magnitudes (BRV) with an RMS error of 0.03 to 0.02 typically from my amateur observatory by using a spectrograph with some attention (numerous mean - important also for reduce scintillation error, wide slit, color equation, ...). You are free not to believe of course. And I admit with you that the sources of errors are potentially numerous (not only atmosphere). This is why I prefer to perform a check on the sky (crosscheck with standard photometric objects) rather than an error calculation, which itself has its own error!

(one of my tasks in my professional job is to define and verify the calibration procedures for instruments on board satellites (French Space Agency) - mainly for Earth observation. From my long experience, "true" absolute accuracy of 1% is still a very difficult task (the relative measures are another matter), and always expensive. My first instinct is caution in this domain).

Christian


Top
   
 Post subject:
PostPosted: 05. October 2013, 15:23:01 PM 
Offline
Meister
User avatar

Joined: 31. July 2006, 16:43:32 PM
Posts: 3400
Location: Köln
Hi Christian,
Quote:
the techique I use is not MY technique. This is a simple, logical and nearly standard method, used in professional observatories in the field of spectrophometry
but they don’t do it from their hometown balcony, right?
Quote:
I've seen some strange things in some professional databases
Keep cool, that applies also to amateur databases as BESS… :)

I already said everything what I wanted to address. If you want to proceed with it, I am fine, of course. It is not my turn to keep you and others away from it. If you contribute such absolute flux data to a professional project – great! I presume, the respective referee will then talk to you, anyway.

Cherio, Thomas


Top
   
 Post subject:
PostPosted: 06. October 2013, 12:33:28 PM 
Offline
Dauernutzer
User avatar

Joined: 06. January 2009, 20:18:29 PM
Posts: 86
Location: Neckargemünd
Hallo

jetzt bin ich endlich auch mal dazu gekommen meine Spektren vom 16./17.8. auszuwerten.

Viele Grüße
Roland


Attachments:
spektrum_nova_del_16_170813_2330_0200.pdf [64.03 KiB]
Downloaded 161 times
Top
   
 Post subject:
PostPosted: 22. January 2014, 12:40:11 PM 
Offline
Meister
User avatar

Joined: 19. August 2006, 11:45:20 AM
Posts: 2726
Location: 66907 Glan-Münchweiler
Hallo,

ich habe Schwierigkeiten mit der Kallibrierung dieses Spektrums. Eigentlich müsste die korrekt sein. Ich meine, dass die Wasserlinien das auch bestätigen, bin abret nicht völlig sicher.

Aber das maximum der h alpha linie ist zu weit im Roten. Das verstehe ich nicht; irgendwas stimmt da nicht. Vielleicht kann der eine oder andere mal mit danach schauen.
es ist natürlich schlecht, dass ich hier keine *.fit datei anhängen kann

berthold


Attachments:
Nova_del100913.png
Nova_del100913.png [ 75.59 KiB | Viewed 5287 times ]
Top
   
 Post subject:
PostPosted: 22. January 2014, 16:12:47 PM 
Offline
Meister
User avatar

Joined: 01. August 2006, 16:08:56 PM
Posts: 3819
Location: 68163 Mannheim
Quote:
es ist natürlich schlecht, dass ich hier keine *.fit datei anhängen kann
Eigentlich sollte ein .fit Datei anhängbar sein, wenn sie 500 kB nicht überschreitet.
kannst die Datei auch zippen und die .zip dann anhängen.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/


Top
   
 Post subject:
PostPosted: 22. January 2014, 18:43:32 PM 
Offline
Meister
User avatar

Joined: 19. August 2006, 11:45:20 AM
Posts: 2726
Location: 66907 Glan-Münchweiler
Hallo,

seltsam, JETZT plötzlich kann ich eine *.fit Datei anhängen! Also hänge ich das reduzierte und kallibrierte Spektrum mal an. Dazu noch ein Bild mit ein paar Daten. Leider wird das wohl nicht reichen, mir mit meinen Kallibrierproblem zu helfen. Aber ich werde Lothar mal mein Neon Spektrum und ein Rohspektrum von der Nova schicken. Vielleicht kannst Du mal sehen, was da los ist, ich vermute ja, dass ich was falsch gemacht habe. Finde halt den Fehler wieder nicht.

Was man aber nach meiner Meinung an dem reduzierten Spektrum prüfen kann ist, mit den Wasserlinien zu gucken, ob die Kallibrierung stimmt oder nicht, aber auch das sehe ich nicht genau genug.

berthold


Attachments:
Nova_130910.fit [16.88 KiB]
Downloaded 122 times
Nova_del 130910.png
Nova_del 130910.png [ 16 KiB | Viewed 5268 times ]
Top
   
 Post subject:
PostPosted: 23. January 2014, 13:12:00 PM 
Offline
Meister
User avatar

Joined: 01. August 2006, 16:08:56 PM
Posts: 3819
Location: 68163 Mannheim
Quote:
Aber ich werde Lothar mal mein Neon Spektrum und ein Rohspektrum von der Nova schicken. Vielleicht kannst Du mal sehen, was da los ist, ich vermute ja, dass ich was falsch gemacht habe.
Hallo Berthold,

ja , mach das. ich schaue es mir an.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/


Top
   
 Post subject:
PostPosted: 27. January 2014, 16:01:11 PM 
Offline
Meister
User avatar

Joined: 01. August 2006, 16:08:56 PM
Posts: 3819
Location: 68163 Mannheim
Hallo Berthold,

anbei das mit SMS ausgewertete Spektrum. Als .fit und als .jpg.

Das Spektrum liegt normal, das Maximum liegt etwa bei 6565 Angström.

Ergebnis von center/gauss:
Midas 005> CENTER/GAUS gcursor ? ? e
start end center pixel_value FWHM

6496.601 6630.970 6565.080 34779. 26.121


Attachments:
Spektrum1d.jpg
Spektrum1d.jpg [ 225.77 KiB | Viewed 5110 times ]
Berthold_orig.fits.zip [7.51 KiB]
Downloaded 116 times

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/
Top
   
 Post subject:
PostPosted: 27. January 2014, 20:56:57 PM 
Offline
Meister
User avatar

Joined: 19. August 2006, 11:45:20 AM
Posts: 2726
Location: 66907 Glan-Münchweiler
Ja , vielen Dank Lothar. Könntest Du Dir bitte das Spektrum ansehen, dass ich hier zuvor eingestellt habe ( vom 10.9.13). ? Ist das richtig oder falsch kallibriert?


(Kannst das eventuell nur mit den Wasserlinien überprüfen)
Danke nochmal und herzlichen Gruss
Berthold


Top
   
 Post subject:
PostPosted: 28. January 2014, 09:41:52 AM 
Offline
Meister
User avatar

Joined: 01. August 2006, 16:08:56 PM
Posts: 3819
Location: 68163 Mannheim
Quote:
(Kannst das eventuell nur mit den Wasserlinien überprüfen)
Hallo Berthold,

das sind leider keine Wasserlinien erkennbar. Das Maximum der Emission liegt in diesem Spektrum bei 6571 A. Das scheint mir auch zu hoch zu sein.

Schick mir am besten die Objektaufnahme und das Neon, Dark zu. Ich kalibriere es dann nochmal.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/


Top
   
 Post subject:
PostPosted: 28. January 2014, 09:56:46 AM 
Offline
Meister
User avatar

Joined: 19. August 2006, 11:45:20 AM
Posts: 2726
Location: 66907 Glan-Münchweiler
Hallo Lothar,

sehr seltsam, habe das mit Vspec und ISIS reduziert und immer wieder diese Abweichung gehabt. Genau von diesem Spektrum hast Du ja den Datensatz. Aber ......dankeschön ......lass mal, auf alle Fälle IST meine Kallibrierung eben falsch. Ich mache das nochmal mit nur 2 Linien. Vielleicht habe ich in beiden Prgrammen die gleiche und offensichtlich irgendwie fehlerhafte linien.lst verwandt......die könnte ich Dir ggf. mal zuschicken......mal sehen......aber nochmals "danke".

Herzlicher Gruss

Berthold


Top
   
Display posts from previous:  Sort by  
Post new topic  Reply to topic  [ 126 posts ]  Go to page Previous 1 2 3

All times are UTC+02:00


Who is online

Users browsing this forum: No registered users and 1 guest


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron
Powered by phpBB® Forum Software © phpBB Limited