Revised Capella

[johnb]

So I have recalibrated and now have the H Alpha at 6563, am i on the right lines, what are the other dips as they dont seem to match anything ?

John B

[the rev]

Looks similar to my first Capella spectrum

[jnp]

Hi John

I normally just line up with a second image of a star having known spectral lines. Here is my Capella from the weekend, with An A class having lots of Balmer lines. Both images need to be taken with the same sprectral dispersion setup (gratting, camerra etc).

Hope the quick screenshot below helps?

cheers

John

[Jack Martin]

John,

Your doing well I also work under London skies and some nights can only see the brightest stars!

Keep practising.

Jack

[Merlin66]

As mentioned...

Richard Walker's Spectral Altas is the amateur's bible.

http://www.ursusmajor.ch/astrospektroskopie/richard-walkers-page/index.html

(Latest English version of the atlas..down to bottom, rhs)

[robin_astro]

Hi John

I normally just line up with a second image of a star having known spectral lines. Here is my Capella from the weekend, with An A class having lots of Balmer lines. Both images need to be taken with the same sprectral dispersion setup (gratting, camerra etc).

Hope the quick screenshot below helps?

cheers

John

Yes that's the way to do it. In general you are taking a spectrum of an object because you dont know what it is. You cannot therefore rely on identifying any features. (These are difficult to see in Capella at low resolution in any case as there are so many lines and they blend together)

You first calibrate your spectrograph using a known standard star (A type stars are good as they show clear Balmer lines) You can then apply this calibration to any other star you measure with the same setup, using the zero order and the dispersion measured from the calibration star. You then can measure the wavelengths of the lines in any other star.

Cheers

Robin

www.threehillsobservatory.co.uk

[johnb]

Yes that's the way to do it. In general you are taking a spectrum of an object because you dont know what it is. You cannot therefore rely on identifying any features. (These are difficult to see in Capella at low resolution in any case as there are so many lines and they blend together)

You first calibrate your spectrograph using a known standard star (A type stars are good as they show clear Balmer lines) You can then apply this calibration to any other star you measure with the same setup, using the zero order and the dispersion measured from the calibration star. You then can measure the wavelengths of the lines in any other star.

So I assume you calibrate against a standard star you mean of the same type ? ie ABO etc. Is that what the reference stars are for in Rspec, that being the case Capella is a G8III so in Rspec I would use the g8iii.dat ?

John B

[robin_astro]

Hi John,

You dont need to use the same type star for calibration as you are just using it to calibrate the spectrograph (both for wavelength and to compensate for the instrument spectral response)

In general, say you are measuring some new unknown star. You have no idea what type of star it is and want to take a spectrum to find out. First you must calibrate your spectrograph, (This is a one off exercise provided you do not change anything in your setup.) Then you can measure the spectrum of any star and apply the calibration to it.

For the Star Analyser, the calibration is done using a known star. An A type star is usually used as it has very clear easily identified lines. (A G typer star is a bad choice for this as the lines are indistinct and blended together at the Star Analyser resolution) You use this standard star to work out the wavelength scale (how many A/pixel) and also to measure how the sensitivity of the spectrograph varies with wavelength.

You then take your "unknown" star , in this case Capella. You first wavelength calibrate the spectrum using the zero order as zero Angstrom and the A/pixel calculated from your calibration star. You then have a correctly wavelength calibrated spectrum of your unknown star with all the lines in their correct position.

You then need to compensate for the fact that the spectrograph is not equally sensitive at all wavelengths. This variation in sensitivity (The instrument response) is also calculated using the standard A star (you divide your A star spectrum by the library version of the spectrum, the result after smoothing is the instrument response). You can then use this instrument response to correct the spectrum of any "unknown" star you take.

If you are using Rspec there are useful tutorials on the website. Similarly there are tutorials on the Visual spec website which take you through these steps

HTH

Robin

[robin_astro]

You then take your "unknown" star , in this case Capella. You first wavelength calibrate the spectrum using the zero order as zero Angstrom and the A/pixel calculated from your calibration star. You then have a correctly wavelength calibrated spectrum of your unknown star with all the lines in their correct position.

You then need to compensate for the fact that the spectrograph is not equally sensitive at all wavelengths. This variation in sensitivity (The instrument response) is also calculated using the standard A star (you divide your A star spectrum by the library version of the spectrum, the result after smoothing is the instrument response). You can then use this instrument response to correct the spectrum of any "unknown" star you take.

If you calibrate your setup for wavelength and instrument response using Vega for example and then apply these calibrations to your Capella spectrum, you can then overlay each of the spectra in the Rspec library (actually from the Pickles database of typical spectra) to discover what spectral type Capella is. You should find that the G8III spectrum matches up well.

Note that at this resolution most of the broad features are not produced by a single line due to one particular element but are a combination of many lines.

Here is what the spectrum of Capella actually looks like at high resolution. (Just a small part at 0.1A resolution, 1000x that of the Star Analyser, from the ELODIE spectrograph used to detect exoplanets at the Observatoire Haute Provence)

Cheers

Robin

[johnb]

Hi John,

You dont need to use the same type star for calibration as you are just using it to calibrate the spectrograph (both for wavelength and to compensate for the instrument spectral response)

In general, say you are measuring some new unknown star. You have no idea what type of star it is and want to take a spectrum to find out. First you must calibrate your spectrograph, (This is a one off exercise provided you do not change anything in your setup.) Then you can measure the spectrum of any star and apply the calibration to it.

For the Star Analyser, the calibration is done using a known star. An A type star is usually used as it has very clear easily identified lines. (A G typer star is a bad choice for this as the lines are indistinct and blended together at the Star Analyser resolution) You use this standard star to work out the wavelength scale (how many A/pixel) and also to measure how the sensitivity of the spectrograph varies with wavelength.

You then take your "unknown" star , in this case Capella. You first wavelength calibrate the spectrum using the zero order as zero Angstrom and the A/pixel calculated from your calibration star. You then have a correctly wavelength calibrated spectrum of your unknown star with all the lines in their correct position.

You then need to compensate for the fact that the spectrograph is not equally sensitive at all wavelengths. This variation in sensitivity (The instrument response) is also calculated using the standard A star (you divide your A star spectrum by the library version of the spectrum, the result after smoothing is the instrument response). You can then use this instrument response to correct the spectrum of any "unknown" star you take.

If you are using Rspec there are useful tutorials on the website. Similarly there are tutorials on the Visual spec website which take you through these steps

HTH

Robin

So this is the Calibrating your spectrum video on the RSpec site ?

Thanks for putting up with all the questions

John B

[the rev]

Yes, it's worth working through the rspec video, I used them as a step by step guide for my first few stars

[johnb]

Yes, it's worth working through the rspec video, I used them as a step by step guide for my first few stars

thanks going to give it a try, probably more questions to follow.....

[johnb]

Well I thought I was reasonably ok with this but the more I look the more confused I get. The video talks about calibration which is fine it then goes on to use a reference file from the library which is not calibrated for the instrument etc. If i reset the program to its defaults and then say load vega i can calibrate against as the lines are well known and documented, so if I carry out a 2 step calibration and then exit the program when I open the program and load a new image is the software still calibrated or do I start over ?

Sorry if im not getting this.

John B

[the rev]

Each time I start I begin with an A class star which I have taken on the same night. This is because an A class has obvious lines, so I use the star as the 0 and the the Hb as the second. Once calibrated I then just do a one point calibration for the rest using the 0 point

[johnb]

Each time I start I begin with an A class star which I have taken on the same night. This is because an A class has obvious lines, so I use the star as the 0 and the the Hb as the second. Once calibrated I then just do a one point calibration for the rest using the 0 point

Ahh penny dropping (at last) so you take a fresh A star each session to provide the reference and then as you say a one point calibration. Thanks that really help - think I will worry about instrument response etc further down the line

John B

[the rev]

Nice thing about instrument response is you only have to do it once until you start doing really serious stuff. And it is simple a case of dividing it from the calibrated image and you have it done.

But initially there is no need. I spent my first few months finding lines and learning what different stars looked like. Lots of fun and really interesting.

Kate