samedi 5 janvier 2019

Review of the year 2018

2018 has been the first complete year with the observatory fully functional. It is a good time to review what happend then.


Mid 2017 the observatory was bck to functional with the Losmandy Titan mount refurbished. I have then been able to observe manually through 2017 with very good success.

In september 2017, I invested in some equipment to do remote. This was a key milestone for more productive (and with warmer environment!) observations.

2018 started with a very very bad weather conditions. I was able to observe seriously in june. But a major turn occured in july as I developped a script for more automated observation including software control jump from one target to the next and of course all acquisition and autoguiding. At the beginning I gave th script a list of target to do. In september, I added the ability to select the next target based on its distance to current telescope position in the sky, the BeSS Be Star Spectra database current priority and my own priority (some bonus for some targets).




During 2018, I have observed 123 nights (partial or total), with more than 615 hours of total exposure time. My project (and only one - almost mono-maniac!) is the spectrscopy of Be stars. I thus acquired 536 spectra of Be stars (more than 12000 orders), reaching a total of 1361 spectra in BeSS and close to 30000 orders.

I reached #1 in number of orders in BeSS and top 3 for the number of spectra of Halpha in BeSS:



Based on all those observations, here is the updated graphs showing the observatory performance on Be stars in terms of SNR (Signal to Noise Ratio) versus Visual Magnitude.






Some sheeps around the observatory this last Fall:



dimanche 21 octobre 2018

Gemini CR2032 battery change & Gemini.Net configuration

I changed the battery CR2032 inside the Gemini 1 controler. I noticed it was easy to bent one pin and connect both pins on the 3V side. So be very carefull when putting the battery so that 0V and 3V are well connected:

WRONG: the top pin is below the battery

Make sure the pin is up

OK: one pin is below and one is on the side of the battery



Details of the Gemini.Net configuration including Gemini Modeling Parameters


ATTENTION: in Prism, connect Gemini.Net through Ascom generic HUB.


Last but not least, I have added a dew heater around the C11 front glass to avoid dew which is now coming after midnight; it is visible in the image below as the bright band around the telescope.



samedi 15 septembre 2018

Top 3 BeSS observer

I have been running my automatic script for some time.now. This helped me to reach 3rd position in the BeSS Halpha observer list and acquiring up to 25000 spectra/orders!



Tonight - september 15th 2018 - I am testing a new approach for my script: each target is automatically selected from a subset of BeSS stars, using BeSS/ARASBeAm priority (ie: date of last observation versus ideal observing period), my own priority (my top 10 targets) and distance to telescope.

dimanche 26 août 2018

V442 And ATEL & BeSS spectra

On august 14th, 2018, I found that V442 And Be star went into a new outburst. I published an ATEL (Astronomical Telegram) with the help of Steve Shore:

http://www.astronomerstelegram.org/?read=11966


Here is copy of that telegram:

A new outburst of the Be star V442 And = HD6226

ATel #11966; Olivier Thizy (Observatoire de la Belle Etoile, Revel, France)
on 18 Aug 2018; 14:22 UT

Credential Certification: S. N. Shore (shore@df.unipi.it)

Subjects: Optical, Request for Observations, Star, Variables

V442 And (= HD6226) has been monitored as part of an ongoing effort to understand its activity cycles. Since Jun. 2018 the star has been under scrutiny in an attempt to detect the initiation of an outburst. This Be star has undergone several spectroscopically active intervals in the past few years. Equipment used is a T0.28m telescope with R=10000 echelle spectrograph, Atik 460ex CCD camera, PRISM v10 software for acquisition and ISIS v5.8.0 for spectra reduction. No substantial spectral changes were detected between Jun 27.089 and Aug 11.009 (33 spectra). However, the Halpha line showed a very significant increase in emission on Aug 14.997 (exposure time of 4800 sec, S/N > 50). Halpha line has shown drastic V/R changes every night since then. Emission was not visible on Hbeta line on Aug 14.997 spectrum but was detected in emission with a 6000 sec spectrum taken on Aug t 15.966 as a single emission on the blueward of the absorption line. The Aug 15.966 spectrum is also showing a broad, weak blueshifted emission on Hgamma. Further spectroscopic observations are strongly encouraged, especially across a broad wavelength range and at high resolution. Spectra are available at the BeSS Be Star Spectra database and further spectra submitted there as well.

BeSS Be Star Spectra database: http://basebe.obspm.fr/basebe/Accueil.php?flag_lang=en



For a dozen days, this outburst has been followed by several members of the BeSS database and I plotted all spectra after resampling them (0.05A/pixel), cropping them (6520-6600A) in ISIS and removing the telluric lines in VisualSpec. Then I used my MAtLab 3D graph script to do several plots below.

 3D graph of all spectra since I'm monitoring V442 And in 2018


 Same as above but with the first spectrum (as reference) substracted


2D spectrogram


Same as above but forst spectrum (reference) substracted


 Same as above with Parula colormap - great for a smartphone's background! :-)


vendredi 17 août 2018

V442 And in outburst again!

I have been observing V442 And since I could in june. I saw small changes in its spectrum but then suddently a drastic change occured between the 11th and the 14th of august - an outburst!

I wrote an ATEL about it - thanks to Steve Shore for helping me out:
http://www.astronomerstelegram.org/?read=1196


Here are my own spectra (all in BeSS database) and I encourage everyone to follow this target, specially during such outburst.




Using MatLab 3D script, I was able to do some 3D graphs as well as spectrogram (here they are spectra with the first one substracted, to better show the emission).

The emission is very bright on Halpha, well visible on Hbeta and can be detected on Hgamma. The He I 5876 and line at 4820A do not show emission but clear periodic behaviour (stellar rotation) and at the time of emission I seem to detect  a change in the spectrum too.



















Code for the MatLab Graph3D.m

%
% Graph2D
%
% Plot a graph from a serie of spectra
% X axis: wavelength (in radial velocity)
% Y axis: time, interpolated, in JD
%

% TODO
% ****
% 1/ fold date into phase using Period & HJD0
% 2/ move from JD to HJD
% 3/ correct wavelength from heliocentric RV
% 4/ dsplay dual labels on date axis: HJD-2457000 *and* Year
%

clear

% Create list of FITS files in the directpry & subdirectories
ds = datastore('.\','Type','image','IncludeSubfolders',false,'FileExtensions',{'.fit','.fits','.FIT','.FITS'});

% Global variables
c=300000.0; % light speed in km/s
JD0=2457000.0;

% Key data from target (V442 And) from CDS & litterature
Name='V442 And';
Author='(c) O.Thizy';
RATxt = '01 03 53.3583';
RA = 0.2788;
DecTxt = '+47 38 32.260';
Dec = 0.8315;
SpectralType = 'Be star';
MagB = 6.7900;
MagV = 6.8200;
MagR = 6.8200;
RV = -55.00;
HJD0=2457987.85;
Period=2.61507;

% Select the wavelength to display
Lambda0=4920.0; % Line near Hbeta
%=6678.0; % HeI singlet
%=6563.8; % Halpha
%=6678.0; % HeI singlet
%=4920.0; % Line near Hbeta
%=4861.0; % Hbeta
%=4340.0; % Hgamma
%=5876.0; % HeI triplet
%=4921.0; %HeI singlet

% Define which is the reference spectrum (usually the first one)
REFNum=1;

Titre=[Name ' / ' num2str(Lambda0) 'A / ' Author];

% Define the spectral domain to display
v1=-500.0; % in km/s
v2=+500.0; % in km/s
% Define the beginning of the spectral domain for all your spectra
% and the dispersion
%CRVAL=6507.5; % Valid for my echelle spectra only - Halpha!!!
CRVAL=4825.5; % Valid for my echelle spectra only - Hbeta!!!
%CRVAL=4277.5; % Valid for my echelle spectra only - Hgamma!!!
%CRVAL=5831.5; % Valid for my echelle spectra only - Na doublet!!!
CDELT=0.05; % idem

% Calculate the index position of the spectral domain within the
% interpolated spectrum
x1 = int32((Lambda0 + v1*Lambda0/c - CRVAL)/CDELT);
x2 = int32((Lambda0 + v2*Lambda0/c - CRVAL)/CDELT);
dX=int32(x2-x1)+1; % Spectrum now will go through (1:dX)
nX = int32(1000);
nY = 100;


REF = fitsread(ds.Files{REFNum}); % Reference spectrum vector (for substrated graph)

% NAXIS=63281; % Nb of Elements
MATRIX=zeros(length(ds.Files),dX);
MATRIX_SUB=zeros(length(ds.Files),dX);
MATRIX_PHASE=zeros(length(ds.Files),dX);
%OBS_DATE=zeros(length(ds.Files));

for i = 1:length(ds.Files)
    fName = char(ds.Files(i));
    H=fitsinfo(ds.Files{i});
    Data = fitsread(ds.Files{i}); % spectrum vector
       
    % searching for key header values
    if max(size(H.Contents)) == 1
        kwd = H.PrimaryData.Keywords ;
    elseif max(size(H.Contents)) > 3
        kwd = [H.PrimaryData.Keywords;H.Image(max(size(H.Contents))-2).Keywords];
    else
        kwd = H.PrimaryData.Keywords ;
    end

    HeaderMax = max(size(kwd));

    for L=1:HeaderMax
        k = kwd {L,1};
        kc = kwd {L,2};
        kt = kwd {L,3};
   
        switch k
            % CALIBRATION & REDUCTION data
            case 'NAXIS1'
                NAXIS1 = kwd {L,2};
            case 'CRVAL1'
                CRVAL1 = kwd {L,2};
            case 'CDELT1'
                CDELT1 = kwd {L,2};

            case 'JD-MID'
                JD = kwd {L,2};
               
            case 'EXPTIME2' % Total exposure duration (including in-between read-out time)
                ExpTimeTotal = kwd {L,2};
           
            case 'DATE-OBS'
                if length(kwd {L,2}) == 17
                    ObsDate = datetime(kwd {L,2},'InputFormat','uuuuMMdd''T''HH:mm:ss','TimeZone','UTC');
                elseif length(kwd {L,2}) == 18
                    ObsDate = datetime(kwd {L,2},'InputFormat','uuuuMMdd''T:''HH:mm:ss','TimeZone','UTC');
                elseif length(kwd {L,2}) == 19
                    if strcmp(kwd {L,2}(14),'-')
                        ObsDate = datetime(kwd {L,2},'InputFormat','uuuu-MM-dd''T''HH-mm:ss','TimeZone','UTC');
                    else
                        ObsDate = datetime(kwd {L,2},'InputFormat','uuuu-MM-dd''T''HH:mm:ss','TimeZone','UTC');
                    end
                elseif  length(kwd {L,2}) == 20
                    ObsDate = datetime(kwd {L,2},'InputFormat','uuuu-MM-dd''T''HH:mm:ss.','TimeZone','UTC');
                elseif  length(kwd {L,2}) == 21
                    ObsDate = datetime(kwd {L,2},'InputFormat','uuuu-MM-dd''T''HH:mm:ss.S','TimeZone','UTC');
                elseif  length(kwd {L,2}) == 22
                    ObsDate = datetime(kwd {L,2},'InputFormat','uuuu-MM-dd''T''HH:mm:ss.SS','TimeZone','UTC');
                elseif  length(kwd {L,2}) == 30
                    ObsDate = datetime(kwd {L,2},'InputFormat','uuuu-MM-dd''T''HH:mm:ss.SSSSSSSSSS','TimeZone','UTC');
                elseif  length(kwd {L,2}) == 31
                    ObsDate = datetime(kwd {L,2},'InputFormat','uuuu-MM-dd''T''HH:mm:ss.SSSSSSSSSSS','TimeZone','UTC');
                elseif  length(kwd {L,2}) == 32
                    ObsDate = datetime(kwd {L,2},'InputFormat','uuuu-MM-dd''T''HH:mm:ss.SSSSSSSSSSSS','TimeZone','UTC');
                elseif  length(kwd {L,2}) >= 23
                    ObsDate = datetime(kwd {L,2},'InputFormat','uuuu-MM-dd''T''HH:mm:ss.SSS','TimeZone','UTC');
                elseif  isempty(kwd {L,2})
                    ObsDate = kwd {L,2};
             end           
        end
  
    end
    % Lambda = CRVAL1 + CDELT1 * x
    % x = (Lambda - CRVAL1)/CDELT1
    % v = (Lambda-Lambda0)/Lambda0*c
    % Lambda = Lambda0 + v*Lambda0/c

    %Xx1 = int32((Lambda0 + v1*Lambda0/c - CRVAL1)/CDELT1);
    %Xx2 = int32((Lambda0 + v2*Lambda0/c - CRVAL1)/CDELT1);
    % dXx = int32(Xx2-Xx1)+1;
    % Xnew=linspace(1,dXx,nX);

    OBS_DATE(i)=(JD-JD0);
    % CalcJD = juliandate(ObsDate + seconds(ExpTimeTotal / 2))
    [HJD,ObjVel]=otz_hjd(JD,[RA Dec]);
    OBS_PHASE(i)=(HJD-JD0)/Period-fix((HJD-JD0)/Period);

    %    MATRIX(i,1:dX)=Data(x1:x2);
    MATRIX_SUB(i,1:dX)=Data(x1:x2)/mean(Data(x2-50:x2-40))-REF(x1:x2)/mean(REF(x2-50:x2-40));
    MATRIX(i,1:dX)=Data(x1:x2)/mean(Data(x2-50:x2-40));

end % end of loop for each file/spectrum

% pcolor(MATRIX), shading interp;
   
T=1:length(ds.Files);
Tnew=linspace(1,length(ds.Files),nY);
%X=double(1:dX);
%Vnew=linspace(double(1),double(dX),double(nX));

Lambda=CRVAL1+double((x1:x2))*CDELT1;
Velocity=(Lambda-Lambda0)/Lambda0*c;
   
Y=linspace(OBS_DATE(1),OBS_DATE(length(ds.Files)),nY);
Y_PHASE=linspace(0,1,nY);

IMAGE=interp1(T,MATRIX,Tnew);
IMAGE_SUB=interp1(T,MATRIX_SUB,Tnew);

%IMAGE2=interp2(T,X,MATRIX(:,x1:x2),Tnew,Vnew);

T=1:length(ds.Files);
Tnew=linspace(0,1,nY);

IMAGE_PHASE=interp1(T,MATRIX,Tnew);


% At the end, create & display:
%
% 2D image in Velocity Vs JD-JD0
% 2D image in Velocity Vs Phase [0:1]
% 2D image + 3D graph (spectra) in Velocity Vs JD-JD0
% 2D image + 3D graph (spectra) in Velocity Vs Phase [0:1]
% Same with REF spectra substracted
%

%pcolor(Velocity,Y,IMAGE),shading interp,colormap gray;
%pcolor(IMAGE),shading interp,colormap gray;
%figure;
%surf(Velocity,Y,IMAGE),shading interp,colormap parula,colorbar,ylabel('JD-2458000'),xlabel('velocity (km/s)');
% colormap jet, colormap gray, colormap parula
% Use plot for 1D, pcolor for 2D and surf for 3D, contourf for art !
%contourf(Velocity,Y,IMAGE),shading interp,colormap gray,colorbar,ylabel('JD-2458000'),xlabel('velocity (km/s)');

% Note: whos VARIABLE gives structure information on the variable !

% 2D image in Velocity Vs JD-JD0
figure
pcolor(Velocity,Y,IMAGE),shading interp,colormap gray,colorbar,ylabel(['JD-' num2str(JD0)]),xlabel('velocity (km/s)'),title(Titre);

% 2D image in Velocity Vs JD-JD0 with REF spectrum substracted
figure
pcolor(Velocity,Y,IMAGE_SUB),shading interp,colormap gray,colorbar,ylabel(['JD-' num2str(JD0)]),xlabel('velocity (km/s)'),title(Titre);

% Graph 2D with all spectra (surface plot & 3D plot, time shifted
figure;
hold on
pcolor(Velocity,Y,IMAGE),shading interp,colormap parula,colorbar,ylabel(['JD-' num2str(JD0)]),xlabel('velocity (km/s)'),title(Titre);
for i=1:length(OBS_DATE)
    plot3(Velocity,ones(size(Velocity))*OBS_DATE(i),MATRIX(i,:))
end

% Graph 2D with all spectra (surface plot & 3D plot, folded in phased time
%figure
% hold on
% pcolor(Velocity,Y_PHASE,IMAGE_PHASE),shading interp,colormap parula,colorbar,ylabel('Phase'),xlabel('velocity (km/s)'),title(Titre);
%for i=1:length(OBS_DATE)
%    plot3(Velocity,ones(size(Velocity))*OBS_PHASE(i),MATRIX(i,:))
%end





dimanche 12 août 2018

Be Star Spectra call for observations

I recently read a very interesting review article on Be stars dated from 2013 that mention our work: "highly motivated amateur astronomers have begun to contribute to the field with spectroscopic observations of increasing quality, comparable to that ofsmall professional instruments. Spectra are made available to the community via the Be Star Spectra Database".

If you are interested in learning more about those very insteresting stars, check out this detailed review which explains very well the definition of Be stars (and what are NOT Be stars) as well as recent (past decade) learning on those objects:

There is also a more recent article (2017) on 28 Cygni that use several spectra from dozen of amateur taken back in 2015-2016 and while the star has not strongly active, spectra are showing some ineresting feature specially in 2016 with some changes seen. The article decribes very well what can be seen on Halpha spectra (EW, V/R, RV) and their meaning:


During the month of june, more than a hundred spectra were taken by 18 different observers. If you are also interested in helping acquiring Be stars spectra, specially with high resolution spectrograph (R>5000), I would recommend you to start practicing the BeSS file format and spectra acquisition/reduction with some bright targets. gamma Cassiopae is usually a good choice to start with, well visible in the sky. You can also start with beta Lyrae, whose Halpha spectrum is always changing.

Once you have a spectrum in BeSS format, you can contact one of the database administrator to check everything is okay. You can also submit it on BeSS databas (submission of a spectrum is actually your registration). There, your spectrum will be checked in terms of quality and compared to other's in the database:

Then, go to slightly fainter targets. Why not 28 Cygni for exemple, whose behavior is always of interest (see article above!). You can also pick lambda Cygni, omicron Andromedae or phi Andromedae as those three stars are under a special scrutiny at the moment, looking for drastic changes in their Halpha profile, sign of an outburst!

Again, simply submit your spectra on BeSS database whre it will be checked. You can also do some prep analysis by downloading a recent spectrum from BeSS - a fully public access - and comparing to yours. Check specifically for telluric lines which should be at the same position (wavelength calibration quality control!) and shape of the line similar (good check on instrumental response for exemple).
Look for spectra taken with similar equipment than yours and compare acquisition time and overall noise on the continuum - are they similar, better or maybe some hints of improvement?

Then, there are plenty of other Be stars to look at. ARASBeAm provides a list, every night, on Be stars to observe by comparing the last spectrum date and the expected observation period. Targets are displayed in red when they should abolutely be observed, yellow when it is recommended to observe them and green when they are not a priority (but you can still observe them). Check out ARASBeAm website:


Be Star Spectra is not only interesting from a scientific point of view (cf the two proposed articles - and more!) but they also help you to progress in general stellar specrocopy. So do not hesitate to join in.

There is just one warning... once you get hooked, you will love it! :-)

samedi 11 août 2018

reading Note: classical Be stars

I have been reading some articles on Be stars and I found out an excellent review by Thomas Rivinius , Alex C. Carciofi & Christophe Martayan: Classical Be Stars, Rapidly Rotating B Stars with Viscous Keplerian Decretion Disks. It was published in 2013.

I also liked the more recent article (2017) by Baader et al: Short-term variability and mass loss in Be stars - III - BRITE and SMEI satellite photometry of 28 Cygni - a very nice review of the non radial pulsations seen in this Be star.

Here are some of my notes taken from the article read - of course all mistakes would be mine! But this explains why I have such interest in Be stars: they are bright but still very mysterious, even if some veils were lifted in the past decade... :-)


Ten years after the excellent review by Porter & Rivinius published in 2003, the authors first clarifies the definition of Be stars: a non-supergiant B star whose spectrum has, or had at some time, one or
more Balmer lines in emission.

This is actually the first time that this definition is clarified as they list stars that are NOT Be stars while they could look like from the definition above:
-Herbig Ae/Be: young object with acretion disk, while classical Be stars have actually decretion disks;-mass transfering binaries: typical Algol or W UMa binaries sometimes show emission lines but this comes from the mass transfered from one star to the other one, different mechanism than classical Be stars;
-B[e] stars also show forbidden lines. There are different types of B[e] stars, some could say that thre is almost one category for each B[e] star! But their mechanism is also different than classical Be stars even if they could show a disk shape circum stellar material;
-magnetic B stars can sometimes show emission line. Circum stellar environment shows variability with exactly the same period as the photosphere, thus making them different to classical Be stars;
-Oe stars can sometimes be considered as the blue extension of Be stars but some spectropolarimetry observation tend to show otherwise;
-A & F shell stars could also be considered as the red extension of Be stars group within the HR diagram but some study show that they could be closer to Herbig Ae/Be stars thus they should not included in the classical Be stars group.

Of course, supergiant B stars such as P Cygni while considered originally as part of the same group were quickly look as different mechanism to eject mass (winds typically), are clearly not Be stars as the definition states.


Variability on all time scale is a clear characteristic of classical Be stars. One can see variations over several decades, with for exemple period of quiescence and no emission at all is mixed with period with visible emission line. But variation of several days, hours or even minutes can also be seen, leading to a wide variety of astrohysical phenomena!

Be star spectra (typically Halpha) can lead to different measurements:
-EW (Equivalent Width) is a measurement of the line strength. Some times, emission can lead to tens of Angstroems of EW.
-V/R is the ratio, when the emission line is double peak, of the Violet peak intensity by the Red peak intensity. Sometimes the continuum is substracted, sometimes even the photospheric absorption profile. So if you do similar measurements, make sure to always detail the definition of your V & R measurements.
Usully variable within weeks to decade, short time variation are usually binarity driven and longuer ones linked to the disk itself.
Variations within few days are closer to the photosphere. Transportation of material from the star to the disk usually takes days or weeks at most. Stellar rotation and typical pulsations period fall into that range too.
Variation of less than half a day are often beta Cep pulsations modes, specially in early type Be stars.

Cyclic asymetry of the emission lines can also be linked to global waves over the disk, very similar to acretion disk phenomenon seen in young objects. But remember that for Be stars, disk are decretion ones.


Of course, Be stars are observed with other technics than spectroscopy, such as:
-(spectro)Polarimetry: light emitted from the star isn't polarized but when the light is scattered by the free electrons in the ionized disk, it becomes polarized. Spectropolarimetry is also a powerful tool to measure magnetifc field, even if those are very weak in massive stars such as Be stars.
-Interferometry had been developped in recent years and both stars and disk have been observed with this technics, only available for large and/or close stars.
-Spectroastrometry
-Photometry, specially space based high frequency precise photometry (BRITE, SMEI, CoRoT, Kelper...), leading to very fine study of the pulsations modes of the Be stars. Ground based photometry are also important as geometry changes in how we see the Be star system (in case of binarity) lead to variation of the absolute flux received to us.

Anyway, check out those two articles, they are very interesting to read:
https://arxiv.org/abs/1310.3962
https://arxiv.org/abs/1708.07360



Reading in the pool... :-)