Presentation of Franck Boubault
Posted: Thu Sep 27, 2012 8:16 pm
Hello,
My name is Franck BOUBAULT, and I live in the south of France, near Bordeaux.
I use a LISA spectrometer with an ATIL 314L camera for acquisitions, and an ATIK TITAN camera for guiding. The spectrometer is coupled with a classical f/10 Schmidt-Cassegrain telescope (200 mm diameter), with a x0.63 focal reducer.
I present my first fully-processed spectra for three stars (Chi Bootes, Vega and Alpha Herculis) that have been processed with a home-made software, that have been improved with including many suggestions from Christian Buil at the last Pro-Am Spectro Meeting at La Rochelle (again, many thanks Christian !). For each star, the measured spectra is in black, and the reference one is in red.
The spectral calibration deserve some comments. I’ve tried to perform an accurate spectral calibration for the whole LISA spectral range, which was a tricky task for me. At the beginning, I only used the brightest Neon lines available with the LISA calibration lamp. In this case, the spectral calibration was satisfactory for wavelenght longer than 585 nm but worse for wavelengths below 585 nm, particulary on the blue side of the spectra. This point has already been discussed on the group. To obtain a better spectral calibration for the LISA spectral range, I now use the following procedure :
- Preliminary spectral calibration with using the brightest Neon lines. At this stage, I consider that the spectral calibration is only correct for wavelenght longer than 585 nm
- Measurement of the position (in pixels) of the Balmer lines visible in the spectra of a hot star (Chi Boo or Vega in my case), with using a lorentzien fitting of the lines profiles.
- Computation of the Doppler-shift of the the observed Balmer lines, with calculating the observed wavelength of the Halpha line (using the Neon lines calibration) and using the theoretical wavelength of the Halpha line. I then deduce the heliocentric velocity of the star at the time of the observation.
- Computation of the theoretical wavelength of the others Balmer lines of the star at the time of the observation due to their Doppler –shift (with using the previously computed heliocentric velocity).
- Definitive spectral calibration with using the brightest Neon lines and the set of Balmers lines (except H alpha) corrected from their Doppler-shift due to their heliocentric velocity
As an example of this procedure, consider the case of Vega, I use 22 Neon lines and three Balmer lines (Hbeta, Hgamma and Hdelta). I use a 3-degree spectral calibration polynom. To assess the quality of the spectral calibration, I compute the statistical values of the residuals, defined by the absolute differences between the computed wavelenghts of the 25 lines used for calibration and their theoretical wavelengths. I obtain a standard deviation of the residual equal to 0.024 nm, with a maximal value equal to 0.069 nm. Since the standard deviation is about one-tenth of the average pixel spectral range of the LISA spectrometer, I consider that the spectral calibration is satisfactory.
Note also the spike around 525 nm for the Chi Boote spectra. I suppose that it is due to a cosmic ray, however, despite many efforts, I have not been able to definitly identify the pixel responsible of this spike.
For Chi Bootes and Vega, the measured and reference spectra are very similar. However, the differences between the measured and reference spectra are greater for Alpha Herculis. Maybe it could be due to the difference between the spectral type of Alpha Herculis (M5IB-II) and the closest reference spectra I have found (m5iii). Or may be AlphaHerculis is a peculiar star or a double star ? Could someone explain the source of these differences ?
In order to assess the relative validity of the Chi Bootes and Vega spectra, I’ve compared the relative instrumental response computed for both spectra. As expected, the responses clearly show differences. They may be explained by the angular height difference of the stars at the time of the observations, but I would be very interested in some additionnal explanations. Are the amplitudes of the differences realistics ?
Many thanks for your comments and answers
My name is Franck BOUBAULT, and I live in the south of France, near Bordeaux.
I use a LISA spectrometer with an ATIL 314L camera for acquisitions, and an ATIK TITAN camera for guiding. The spectrometer is coupled with a classical f/10 Schmidt-Cassegrain telescope (200 mm diameter), with a x0.63 focal reducer.
I present my first fully-processed spectra for three stars (Chi Bootes, Vega and Alpha Herculis) that have been processed with a home-made software, that have been improved with including many suggestions from Christian Buil at the last Pro-Am Spectro Meeting at La Rochelle (again, many thanks Christian !). For each star, the measured spectra is in black, and the reference one is in red.
The spectral calibration deserve some comments. I’ve tried to perform an accurate spectral calibration for the whole LISA spectral range, which was a tricky task for me. At the beginning, I only used the brightest Neon lines available with the LISA calibration lamp. In this case, the spectral calibration was satisfactory for wavelenght longer than 585 nm but worse for wavelengths below 585 nm, particulary on the blue side of the spectra. This point has already been discussed on the group. To obtain a better spectral calibration for the LISA spectral range, I now use the following procedure :
- Preliminary spectral calibration with using the brightest Neon lines. At this stage, I consider that the spectral calibration is only correct for wavelenght longer than 585 nm
- Measurement of the position (in pixels) of the Balmer lines visible in the spectra of a hot star (Chi Boo or Vega in my case), with using a lorentzien fitting of the lines profiles.
- Computation of the Doppler-shift of the the observed Balmer lines, with calculating the observed wavelength of the Halpha line (using the Neon lines calibration) and using the theoretical wavelength of the Halpha line. I then deduce the heliocentric velocity of the star at the time of the observation.
- Computation of the theoretical wavelength of the others Balmer lines of the star at the time of the observation due to their Doppler –shift (with using the previously computed heliocentric velocity).
- Definitive spectral calibration with using the brightest Neon lines and the set of Balmers lines (except H alpha) corrected from their Doppler-shift due to their heliocentric velocity
As an example of this procedure, consider the case of Vega, I use 22 Neon lines and three Balmer lines (Hbeta, Hgamma and Hdelta). I use a 3-degree spectral calibration polynom. To assess the quality of the spectral calibration, I compute the statistical values of the residuals, defined by the absolute differences between the computed wavelenghts of the 25 lines used for calibration and their theoretical wavelengths. I obtain a standard deviation of the residual equal to 0.024 nm, with a maximal value equal to 0.069 nm. Since the standard deviation is about one-tenth of the average pixel spectral range of the LISA spectrometer, I consider that the spectral calibration is satisfactory.
Note also the spike around 525 nm for the Chi Boote spectra. I suppose that it is due to a cosmic ray, however, despite many efforts, I have not been able to definitly identify the pixel responsible of this spike.
For Chi Bootes and Vega, the measured and reference spectra are very similar. However, the differences between the measured and reference spectra are greater for Alpha Herculis. Maybe it could be due to the difference between the spectral type of Alpha Herculis (M5IB-II) and the closest reference spectra I have found (m5iii). Or may be AlphaHerculis is a peculiar star or a double star ? Could someone explain the source of these differences ?
In order to assess the relative validity of the Chi Bootes and Vega spectra, I’ve compared the relative instrumental response computed for both spectra. As expected, the responses clearly show differences. They may be explained by the angular height difference of the stars at the time of the observations, but I would be very interested in some additionnal explanations. Are the amplitudes of the differences realistics ?
Many thanks for your comments and answers
