Forum der Vereinigung der Sternfreunde

Very well!

Hello Thomas and Knud,

the 10s-Regulus image has an max. intensity (after bias correction) of 10000 ADU.
1 picture of that is too low signal. I recommend 40 s or 30 s exposure time and >= 5 images to get a good S/N.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

Aren't there three Regulus spectra?

Hello,

in the nova spectrum images if find only a very weak continuum (about 60 - 80 ADU in the brightest orders in the averaged and dark corrected image). Without any structures (emission lines).
The sky spectrum is empty (perhaps very weak shadows of ThAr (memory effect ?)).

I don't have used a echelle data reduction pipeline. Only averaged the images (nova and sky), dark corrected and looked at the resulting picture. Also taken column profiles. See the graph.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

I see only 1 Regulus image and 2 ThAr on the server.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

Hmmm.... we startet at about 35 degrees (large airmass) and the spectra have been weak indeed. This target is faint for the spectrograph. Maybe we better leave this business and stick on our prime targets.

Thomas

Oh indeed, we did only one spectrum. SORRY!

Thomas,

I think you should improve your standards a little bit. We are only amateurs, but should try to improve constantly. Now that the process is working you could do some fine tuning.
My questions
- where are you?
- how far is the beach?
- do you own the telescope?
- excellent S/N for a 4 inch telescope!

Why is there a fits header in every image?

Regards,

Martin

Hi Martin, we installed an automatic routine. So, we can spend our nights in the clubs of Puerto de la Cruz just at the beach together with the canarian girls. Great fun! Yes, I just bought the telescope and, hence, I now have another 80cm but close to the beach clubs. The 4 inch in the header is for the guider scope. We take data with it and watch the stars by eye together with Angelina, Juanita, Pila and Mathilda. Even more fun.

I think, we give respective information in the image titles and everybody knows the Teide telescope (5 inch), altitude (280m) and the location (beach club). Please understand, the girls keep us busy so that individual header editing should not be necessary.

Cheriohoho, Thomas

ATTENTION!

It seems that we will be clouded out tonight. Additional support observations are highly welcome tonight.

Hello everybody,

sorry to bother you once again with the FITS header
After I read that, I assumed that the times in the fits header are wrong before may 27.
WRONG!
The computer knows about the difference of UT and CEST and writes the correct timestamp, both UT and local time, see below. This is true even before may 27. The times indicated in my spectra as
_zetaql_20130524_990_Berthold Stober_full
_yyyymmdd.ddd...
with the decimal fraction of the day are therefore correct. The time is calculated from the information in the FITS header.
I tried to calculate the elevation of zet Aql for that spectrum, assuming the time was CEST and arrived at +3°, so I knew that something was wrong, as the elevation limits are set at around 30°. With the correct time everything works out fine.
So it is always good to do a cross check from time to time.

Regards, Martin

Thomas,

since clouds come in and disturb your party you may have time to study the following post by Christian Buil:
http://www.spectro-aras.com/forum/viewt ... 2272#p2272
a subject which has made me headaches with my response function. With a fiber at long focal length the problem of spectral response becomes even trickier. Fortunately we do not have to produce photometric calibrated spectra! But you see some effects also in the correction of the blaze function. Therefore once again my recommendation to use a calibration star nearby such as zet Aql. Of course it is easier to do Regulus early at night, but not for the whole campaign duration. Switching the calibration star seems risky to me, adding mpossibly another unknown.

Clear skies and turn on the autopilot, Martin

Yes, Martin, I understand this problem. Thank your for this remark! After cancelling the Nova we might focus on the second telluric spectrum from zeta Aql in the morning. We try it tonight.

The clouds probably will go away as far as I can see now. La Palma is already free of clouds and Venus/Mercury are visible again.

Thomas

the spectra of zet_aquilae 130603 are overexposed

berthold

at time it seems impossible to add pictures here...........

Hallo Berthold,

you may annex 5 files, each max. 500 kb heavy. "fits" files are forebidden. But you may put on the server forbidden files in form of a ".zip" file < 500 kB.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

yes Lothar you are right, I didn't realize, the size of the files were to big, now it could be ok!

berthold

ATTENTION! As yesterday, it seems that we will be clouded out tonight. Additional support observations are highly welcome tonight.

http://oiswww.eumetsat.org/IPPS/html/la ... Africa.jpg

I cannot even help with the supplementary targets as I have been clouded/rained out for over a week in southern Australia

Cheers, Bernard


From: Thomas Eversberg (fg-spek-convento@vds-astro.de)
Sent: Wednesday, June 05, 2013 7:09 AM
To: fg-spek-convento@vds-astro.de (fg-spek-convento@vds-astro.de)
Subject: [fg-spek] Re: Teide 2013 WR campaign




ATTENTION! As yesterday, it seems that we will be clouded out tonight. Additional support observations are highly welcome tonight.

http://oiswww.eumetsat.org/IPPS/html/la ... Africa.jpg





No virus found in this message.
Checked by AVG - www.avg.com
Version: 2013.0.3343 / Virus Database: 3184/6381 - Release Date: 06/03/13

_________________
Quicquid Nitet Notandum

Here is something important I just learned from our Danish fellow Knud:

In Danmark people are mysterious!

Aha

In any case, we all appreciate your efforts, Bernard! Even my mysterious partner Knud.

...waiting for clear skies.................

Hi all! The night is almost over. We have been clouded out. No spectra from Teide.

Cheers, Thomas

Hello,

here is a first comparaison from WR134 spectra but only in a overview analysis.

berthold

Hallo Berthold,

're pretty well comparable / reproducible.
Can you add the dates of observation to the graph?

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

hello Berthold,

are the spectra heliocentric corrected?

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

Hello Lothar,

spectra are only simply reduced and calibrated. Not heliocentric corrected and not normalized....

Berthold

Hello,

now I have written the dates of observation to each spectrum! I could it do only with midas. As said already , I don't know how to do this in VSPEC....

As further already mentioned, this can only be a first extraction, not normalized and without heliocentric correction.

But I think in only 3 days the differenz of heliocentric speed variation would be rather small.....

berthold

Hello,

now I have reduced the spectra from WR135 in an other overview but already without normalizing and so on........but in this small overview I cannot se any varitions in the spectra........

Berthold

Hello all, hello the Teide team,

There is an easy way to produce such PNG synthesis plot with a simple SpcAudace's function:

spc_multifit2pngdec offset_vertical_entre_profils ?xdeb xfin? ?lambda_reference?

- offset_vertical_entre_profils: vertical offset between each lines profiles;
- xdeb xfin: wavelength's crop limits (optional)
- lambda_reference: line's default wavelength (example Ha:6562.819 A) (optional)
then horizontal axis will then be in radialvelocity unit (km/s)

Lines profile's anotation is automatic.

Example: spc_multifit2pngdec 1 6545 6583
do an 1 unit vertical offset between lines profiles that have been croped btween Lambda 6545 and 6583 A.
spc_multifit2pngdec rescale each spectra's continuum to 1 because it's then easier to apply a vertical offset.

Have in mind that this fucntion works on all fits files present in the working directory.
Copy your selected lines profiles in a new directory and setup this as working directory in Audela's Setup menu.

Here are output examples made for del Sco's 2011 survey.

Clear skies,

_________________
Benji

/* Webpage: http://wsdiscovery.free.fr/astronomie/spectro/ */
*~~ SpcAudace: the easiest spectra reduction software for beginners ~~*
*~ High rate production pipelines: http://wsdiscovery.free.fr/spcaudace/

Hi all! After two service nights and one clouded out we will take spectra again. Seeing is bad (about 1") and the guiding unit FOVIA will operate in Monday at the earliest. Too bad. Alex Oscoz, IAC80 person in charge, confirmed that this happens once in many years. That is real astronomy live... We will proceed anyway, of course, and hope to get at least one FOVIA night before team 2 arrives in Wednesday. Only then can we train them on this technique.

Cheers, Thomas

Hello Thomas and Knud

that's fine.

Perhaps you should test a new USB cabel between the Laptop and the CCD-interface. The fact that a simple moving on this cable yields the fine function of the CCD some days ago is a little mysterious. At any case look to that connection when read oud problems happen.

_________________
Herzliche Grüße / best regards

Lothar

https://lotharschanne.wordpress.com/

Hello,

looking the Thar_spectra from 130601 and comparing it with the Tharspectrum of 130607 I see a big drift 34(!) pixel in x direction and about 11 Pixel in y direction.

That's a lot!

This cannot be explained with a instability of the eShel. There is something wrong or the eShel was moved. Later I'll try to reduce some spectra of 0607 and 0608........

berthold

Hello,i hope that everything is going well. I just have access to the June 7 raw data and i have some suggestions.
1-Thar are over exposed. 30s is too much at once it is better to do 3X10s or even only one 10s, to avoid memory effect on the first object exposure. Wait 2-3 min before starting object exposition.
2- Tung Flat are not enough expose. Try to get a higher dynamic in the blue order. Go for 10s exposure or multiply the number of individual exposures.
3- Led flat are too noisy. Too less dynamic. well there are not necessary if you do point 2, and instrumental response will be smoother. At the moment they shouldn't be used for data reduction.
5- Bias frame are not really bias meaning a 0s exposure. There are 0,5s
6- Thirty percent of the frames are lost cause of  a readout problem. It could be a good idea to bring a new ccd or/and a new laptop by the next team.
7- Fit header are not filled correctly. Site coordinate, altitude, site name, instrument, Object name are not correct. You have to setup the right values in the acquisition software to make easier the reduction.
8-There is no dark for flat and thar, 6s and 30s


That's not a lot to do to reach perfection cause object exposure looks good, although the readout problem and will make easier and better the data reduction process. You are doing great observations.
Cheers
Thierry



2013/6/9 Berthold Stober <fg-spek-convento@vds-astro.de (fg-spek-convento@vds-astro.de)>

_________________
Thierry Garrel
Observatoire de Fontcaude, Montpellier, France
SpectroBlog http://gabalou.canalblog.com/

Hi Thierry,

Thanks for your input. I don't have an issue with the ThAr as they are now. Co-adding them is actually worse than just tossing the overexposed lines (in my opinion). The flats could perhaps use a little more time.


The headers can be overcome with the file renaming procedure. As I have been working on a reduction pipeline here at UdeM, I have already noticed this and developed some simple software that fixes all the issues you noted, as well as a quick plot of the spectra to check the object names.


For the darks, we prefer just the longer ones. We can subtract a bias from the dark, and then subtract a scaled down version of the dark from the files. So, if we have 900 sec darks, and our exposure is only 30 sec, then we would subtract (30/900)*(Dark-Bias) to get the dark corrected image.


Have you obtained any spectra yourself yet, Thierry?


Cheers,


Noel


On Jun 9, 2013, at 12:07 PM, thierry Garrel <fg-spek-convento@vds-astro.de (fg-spek-convento@vds-astro.de)> wrote:

Hi Thierry!

Thanks a lot for watching and controling our procedures. We highly appreciate that!

1 - I already asked Noel if 30s for ThAr are fine and he agreed. But we will now try 20s. Please check again tomorrow.

2 - I disagree. If we increase Tungsten times it will saturate the red part. For the blue we have LED flats.

3 - Our LED flats have 25000 counts in the peak. We have some flexibility here and could increase exposure times. However, we will apply the average out of 10 to the spectra. That certainly will deliver a good master flat.

5 - Correct! But we can not expose 0 seconds with our software. 0.5s should be fine, anyway (at least I hope so).

6 - We already discussed this sad issue. Noel said that he will be able to correct it later in IRAF (but not the ThAr, they should be done again if necessary). This is a reasonable statement which he confirmed yesterday on my additional request (just to be sure). Even the data images of bad read-out contain all information, they are only split and shiftet somehow. Because of this we should not touch the already working but sensitive system by investing adjustment time and loosing nights.

7 - Yes, we know. Please keep in mind that we are two people in team 2 and the guiding unit FOVIA is out of operation. We do hand guiding by eye. As long as the data file naming is done correctly including object name, date and exposure time, data identification is granted. The rest (site coordinate, altitude, site name and instrument) is known by everybody and can be found in the internet.

8 - Good point! But you talk about the slow data reduction method without bias. The fast method uses a master bias re-scaling the master darks with . This is the main reason why we take biases.

I hope I could clear up some issues. Our problem with performing the observations is twofold:

a) Presently we are only two. We guide the whole night by hand (no problem) and have no time for additional things of minor importance. For this reason we follow the most efficient track: Observations at Teide - data inspection in Europe (You, Martin etc.) - Data reduction in Montréal after the run.

b) With respect to all teams with different spectroscopic experience we should keep everything as simple as possible. We have a running system and pipeline, controlled by other team members in Europe and we better leave it.

I hope this helps.

Salut de Teide, Thomas

Oh, a message overlap...

Noel, should we then increase the flat exp time?

Again: Checking the work of the Teide teams and data by other camapign members plus respective comments is very important for the outcome of this task. Please proceed with it! Thanks a lot for doing so.

Hello everybody,

the good news first:
despite the shift of the images the reduction with ISIS still works well. I uploaded some spectra and the corresponding configurations and logfiles for ISIS to the server at 20130607/ISISspectra0607.zip. Notice that you can load these configurations in ISIS and then redo the reduction yourself after adjusting the working directory and filenames of your own. The description how to do it you find in the same folder, look especially at page 11. Notice that you can also find the used instrument response in the calib subfolder (respreg0607.dat).

the bad news:
I have said it before, Thierry has said it, I say it again for the last time:
PLEASE SORT OUT THE BAD FILES right after you have taken the spectra! Is this so difficult?
That way you can repeat missing spectra. From the spectra of zetAql e.g. there was one good image, one overexposed and one totally bad. From this you cannot produce good reference spectra! It would only take a few minutes to take a few more, but you have to look at the spectra from time to time when you take them.
In my opinion I think it is a waste of time to use a camera which is not working 100% of the time. As Thierry noticed, the rate of bad images is about 30%, so you spend hours each night just to produce bullshit and also waste the time of the people who have to sort out the bad files and throw them away. You also have to throw away valuable scientific data.
Amateurs can do better than that.
That the fits headers still have not been changed is deplorable. Of course they can be changed afterwards, but it would take less than five minutes once and for all to do it right before taking images. Also think about other people contributing spectra to the campaign. Then a correct header is very valuable to sort the spectra out from different telescopes, locations and observers. After all, that is why there is a FITS header in the first place. A correct header ensures good scientific data. It is used for heliocentric correction, atmospheric correction etc.

Last I add a spectrum of WR137. More can be found on the FTP server.

Regards, Martin

Hi Martin! No, it is not. But it is also not necessary according to Noel (asked him twice). Why wasting time if it can be readjusted later?
No, only one overexposed, and one is split (see Noels answer). The third one with intermediate counts. We will reduce the exp time to 60s. This simply depends on the seeing conditions and our hand guiding abilities. Theory from your office is sometimes different than the real world. PLEASE LISTEN TO THE PERSON WHO REDUCES IT!
No, see above. I will not accept attaching a new camera and taking the whole loop into risk. Team 1 needed the full two weeks to bring it into operation. THAT would be waste of time. Again, see Noels two ansers.
Yes, from their backyard with a second CCD in the living room. Or at Teide with additional equipment available. We are 9 teams of different experience. The lowest experience drives the whole campaign and not a single team of experts.

Just this evenning we found glitches in Regulus spectra and repeated it. But we can't do all that under time preasure. Please understand, the PIs and me have to decide for reasonable steps towards the whole campaign considering various aspects. I know that we have different alternatives to proceed. It might be that the future will proof us wrong but that is always the risk for the campaign organisator and the PIs. I will take it on my account...

Cheers, Thomas

Thomas,

just a short comment:
What is the scientific value of such an image?
If you do not want to waste your time, I am not going to waste my time. Of course I have to leave it up to you and Noel how to sort this out.

Martin

Zero scientific use, Martin. But do we really talk about this specific trash which does not happen by 30%? I think no. I talk about the specific "split-and-shift" images. How many of these extremes do you find in percent?

During guiding I thought it over again (a nice meditative task). I have two suggestions:

1. The fits header: All files are in Canada. You take them and change the header content with respect to observatory coordinates, height, target name etc. Information is in the file names. Then you send them back to Montreal. Et voila, the header is fine.

2. The bad images: You and Thierry ask the community for a new CCD camera and Thierry brings the camera to Teide (only his team is familiar with the EShel). After consulting the owner of the spectrograph (Berthold Stober) and only after he says YES Thierry changes the setup. NOBODY TOUCHES THE EShel WITHOUT BERTOLD'S CONFIRMATION. I GOT STRICT ADVICES FROM HIM. For this you guarantee that we do not loose more than three nights.

In fact, I do not believe it to be a hardware problem but you may proof me wrong. Here is Noels answer on my last request two days ago:

If you or anybody else deliver a doable and reliable solution for the improvement of our run I will take it into account. But untill now, you only told me what is wrong not how to realistically solve it.

Cheers, Thomas

Martin, I could decifer the file on your screenshot and checked the respective directory. I presume you missed that there are not three Regulus spectra but five. The reason: We realized this bad image not reducable by Noel (plus an additional one). Hence, we did two quick extra images but forgot to delete the bad ones. Whenever we find these strange data, we do additional exposures.

Please be sure that we do our best to deliver maximum output.

Presently we test for the software and Lothar's controle computer in the dome. Our assumption: When reading out the CCD the computer might be confused if we work paralelly. For that we now do not touch the computer when reading out. In some hours we know a bit more.

Thomas

Hi all! The night is over. For the first time we got spectra of all our targets. Three series of each WR and a single series each of both supplemental targets. This includes one series of Regulus and zeta Aql each.

Most important: Since we did not touch the computer during read-out we did see only one read-out error again. This is out of more than 30 images (failure rate 3%). This is only a preliminary result (bad statistics) but at least hints towards our idea that the computer is not powerful enough. We continue with this tomorrow. Another possibility is the USB cable. We check it tomorrow.

What do we learn from it? We started an intense discussion about the read-out error and I felt the present engagement. Instead of being frustrated we continued to think about the problem. In this sense it is Martin’s good scientific attitude insisting in solving problems. Thank you, Martin, for that. Besides, Filipe Dias sent a hint on the USB cable. We, hence, learn that other colleagues read the stuff we write and also think about that. Many thanks to Faro!

Sun rises, we go to bed. Cherio, Thomas

Has the camera been rotated a little? Have the tungsten and LED flats been affected?

Bernard


From: Berthold Stober (fg-spek-convento@vds-astro.de)
Sent: Sunday, June 09, 2013 8:49 PM
To: fg-spek-convento@vds-astro.de (fg-spek-convento@vds-astro.de)
Subject: [fg-spek] Re: Teide 2013 WR campaign




Hello,

looking the Thar_spectra from 130601 and comparing it with the Tharspectrum of 130607 I see a big drift 34(!) pixel in x direction and about 11 Pixel in y direction.

That's a lot!

This cannot be explained with a instability of the eShel. There is something wrong or the eShel was moved. Later I'll try to reduce some spectra of 0607 and 0608........

berthold





No virus found in this message.
Checked by AVG - www.avg.com
Version: 2013.0.3343 / Virus Database: 3199/6397 - Release Date: 06/09/13

_________________
Quicquid Nitet Notandum

No, Bernard, the spectrograph is untouched.

Thomas

You don't need to touch the spectrograph to move the echelogram on the ccd. If you touch the fiber at the point of injection in the spectro it moves. It moves also with slight variation of focus. To make the diference look at the slant of the tharg lines. Slant are diferents if focus has changed.
T Le 10 juin 2013 17:56, "Thomas Eversberg" <fg-spek-convento@vds-astro.de (fg-spek-convento@vds-astro.de)> a écrit :

_________________
Thierry Garrel
Observatoire de Fontcaude, Montpellier, France
SpectroBlog http://gabalou.canalblog.com/

Yes, I know, Thierry. Because of service nights we had to disconnect the fibers at the telescope, of course. But the EShel and its fiber input are not moved.

Thomas

Hi Noël,
no more spectra over here waiting for clears skies, very rare at the moment. A very rainy year.
When i was talking about thar, it was not about calibration, but memory effect. Our ccd KaF3200 is badly sensible to this effect who blamed the first exposure of each series. Going down under saturation, useless for calibration, will help to avoid memory effect. You can see this effect on each first image of object series. I used ten second exposure for my own eshel calibration, with the same sensor with a very precise calibration in the blue faint orders.
On dark, i prefer too use special series of darks for different exposures, specially if there are short. I'm more confident than using the same master dark for each exposure. Just a different manner, same results.
Did you check the dark, is it necessary to redo them all the time ? Note that cause of the readout problem the pre treatment images does not fulfill the number required each day, i saw for example only 5 offset good on June 7.
Cheers
Thierry

_________________
Thierry Garrel
Observatoire de Fontcaude, Montpellier, France
SpectroBlog http://gabalou.canalblog.com/

Our last night seeing!

I think you are trying to make us the colour of the graph ... green with envy

Bernard


From: Thomas Eversberg (fg-spek-convento@vds-astro.de)
Sent: Tuesday, June 11, 2013 2:16 PM
To: fg-spek-convento@vds-astro.de (fg-spek-convento@vds-astro.de)
Subject: [fg-spek] Re: Teide 2013 WR campaign




Our last night seeing!






No virus found in this message.
Checked by AVG - www.avg.com
Version: 2013.0.3345 / Virus Database: 3199/6399 - Release Date: 06/10/13

_________________
Quicquid Nitet Notandum