First go with Star Analyser

[NadrejN]

Hello All

Bought Star Analyser 100 some month ago, but had not started using it until now - should not wait so long with this interesting field.

In the past week I recorded around 30 different spectra, mostly named stars and as a quick start, just put them all in the same image to get a feeling for the differences in different stars. BASS project was used for processing, but only stretch was applied to get them all at a similar brightness.

Spectra are arranged from O on top to M at the bottom - used Simbad database for clasification.

Still need to invest much more time in each spectra and try to get the data out.

Very interesting topic and many new opportunities to learn.

Comments much appreciated.

Best regards

Andrej

[Merlin66]

Andrej,

Looks like you're off to a great start!

Well done.

Onwards and Upwards.

[Ruud]

Hi Andrej,

interesting results indeed. I wonder what fireworks look like through a spectrometer or LED lamps or ... It reawakens the young scientist in me.

Good luck with the Star Analyser and thank you for sharing your data.

[Gasman]

Excellent start Andrej!

Look forward to seeing the plots from those

Happy new year to all on the forum too!

Steve

[jnp]

Excellent start Andrej. You can already get a feel of how the stars get hotter and see the major lines, especially Hydrogen Balmer.

 

Still need to invest much more time in each spectra

Not sure what process you followed, but, so far so good.  The common slitless spectra processing steps are:

• stack images
• correct for rotation/tilt
• set binning and sky subtraction regions.
• calibration
• instrument response correction

I wonder what fireworks look like through a spectrometer......... It reawakens the young scientist in me......

Would expect to see emission lines specific to the metals/salts mixed with the gunpowder (red=strontium, blue=copper, yellow=sodium etc). I agree spectro opens the door to plenty of science.  Much more than taking pretty pictures.

Happy New Year!

John
 

[Thom]

... I agree spectro opens the door to plenty of science. Much more than taking pretty pictures...

But pretty pictures have their place too!

Sent from my GT-I9505 using Tapatalk

[NadrejN]

Hello All

Thanks for all comments and a happy new year.

@John:

For first post image no processing was done. Was too thrilled with what I was seeing to go into processing .

Image acquisitions were done with Grasshopper 3, 1200 mm focal length and exposure length was adjusted by eye to get just below saturation on the spectra. For each star 20% out of 100 frames were stacked - maybe in some cases hand-picking frames would give better results, but here did not see much of a difference. Additional errors could be caused by lack of flat and dark frame correction.

I followed processing steps same as you wrote. Instrument response was done by dividing acquired star spectra with reference spectra. Interesting here is, that with different stars give different response - need to do some more reading.

Below is data for Castor, a binary star - as image shows clearly. But, incredibly, each component is in fact a binary star itself. ...wish I had a bigger scope.

Stacked image

Stacked image, rotated and stars moved closer to spectra for easier viewing.

Castor A - unmodified

Castor A - processed

Castor B - unmodified

Castor B - processed

Thanks,

Andrej

[Dave In Vermont]

Wow!

A very nice job. It always amazes me what these things can do. I have the Rainbow Optics models, which are about the same as the SA 100, and I'm always finding new things to try with it.

Clear & Colourful Skies,

Dave

[Merlin66]

Dave,

The Rainbow Optics version is actually a 200 l/mm grating, so double the dispersion of the SA100 and similar to the new SA200.

[robin_astro]

Hi Andrej,

Interesting here is, that with different stars give different response - need to do some more reading.

Very nice results there.  

The instrument response should not vary depending on the star as it is only a function of the instrument. Here are some measurements I made to demonstate this.  

http://www.threehillsobservatory.co.uk/astro/spectroscopy_21.htm

Measuring the instrument response accurately is not easy though and there are many possible reasons why your instrument response can vary from star to star (or even from observation to observation of the same star.)

1. The instrument response you measure includes the extinction effect of the atmosphere which varies with altitude of the target in  the sky and from night to night. For this reason it is best to measure the instrument response on the night using a reference star (typically using a main sequence A star) at similar elevation to the target you are taking a spectrum of. You can then use this to correct your target star spectrum. In practise though, provided you do not go too near the horizon (eg above 30 deg say your instrument response should not vary by much except at the extreme blue end.

2.  The library spectrum you are usign to compare with may not represent the true spectrum of the star observed. For various reasons, not all spectra are identical even if the published spectral type is  the same. Again if you want best accuracy you need to use a spectrum of the actual star, taken by a professional observatory. The MILES database is typically used for low resolution spectra.

These to effects are generally relatively small though and if you are seeing large differences then it is probably due to processing problems.  The most common ones are.

a) Having a false zero by not accurately subtracting the background (for example it is important to make sure that there is no contamination from the spectrum in the subtracted background)

Not correctly removing the lines from the continuum and doing either too much or too little smoothing.  (It is for this reason that A or B stars normally used to measure the instrument response as they contain few lines and are more easily identified and removed. )

I hope these tips help. You are certainly off to a very good start based on your results so far 

Goog Luck !

Robin

[robin_astro]

Another potential problem when trying to measure and correct spectra for instrument response is non linearity of the camera.  Webcams and video cameras for example my have a gamma response built in. Unless the gamme correction can by switched off (set to 1) Spectra taken using these types of camera cannot be accurately corrected for instrument response.   For many applications though there is no need to know the exact shape of the continuum. the features in the spectrum are usually the most important information and these are visible even without an accurate instrument response correction.  For example one option is to  remove the broad continuum shape completely, producing a level spectrum with the features superimposed on it. (known as normalising the spectrum relative to the continuum.) You can see an example of this at the top of this page here.

http://www.threehillsobservatory.co.uk/astro/spectroscopy_11a.htm

Robin

[NadrejN]

Robin, thanks very much for the detailed answer. I have actually already read some of the links you provided, very good source of information. 

Do not know how I managed to miss the Miles database reference, looks very interesting. 

Did a quick extraction of instrument responses yesterday and when displaying them together form looks more or less the same. They are shifted around a lot and all responses from O, A, B stars have a hump starting at around 725nm. This are stars that don't have much signal in my data at those wavelengths and with K an M stars the hump disappears. Maybe caused by atmosphere.

[robin_astro]

Hi Andrej,

The far red end might be affected by atmospheric telluric bands beyond ~6800A which will appear in your spectrum but not in the library version. You can see them identified here for example.

http://www.astrosurf.com/buil/us/vatlas/vatlas.htm

The fact that it appears in the hot stars rather than the cool ones suggests though that it might more likely  be contamination from the second order spectrum which will appear as a ghost spectrum at twice the wavelength. Although it is quite weak with the Star Analyser (perhaps only 1/10 as strong as the main spectrum), because hot stars have so much higher intensity in the blue compared with in the red, the contamination beyond ~3800A (which will appear in the second order at 7600A) can be significant.  If you are specifically interested in looking at the region beyond ~7600A  you can add an order filter (a yellow/red short wavelength cut filter) to block the contaminating blue end of the second order spectrum. You can see an example of this technique with the Star Analyser here 

http://www.threehillsobservatory.co.uk/astro/Classifying_red_stars_using_a_Star_Analyser_VdS_poster.pdf

Robin

[NadrejN]

Last weeks were quite busy and with Lovejoy in a good position, focus changed.

Nevertheless, here are spectra of Aldebaran and Mirach (Mirak), which are K5III and M0III spectral type stars and quite different from Castor.

It is interesting, that the recorded spectra are very similar with very little difference between them.

Aldebaran - original recorded spectra.

Aldebaran - instrument response corrected spectra.

Mirach (Mirak) - original recorded spectra.

Mirach (Mirak) - instrument response corrected spectra.

Best regards

Andrej

[Ruud]

Hi Andrej,

what kind of telescope are you using? I ask because in the corrected spectra there is a strong drop in intensity around 925 nm. Is this caused by the atmosphere, absorption in glass (refractor) or a drop in reflectivity of the mirror (reflector) at this wavelength?

Anyway, the spectral lines are are plentiful and well defined.  

[NadrejN]

Hi Ruud

Equipment is:

OrionOptics 12" reflector (dont know that it is some particular model), Grasshopper GS3-U3-28S4M.

According to documentation of camera - http://www.ptgrey.com/support/downloads/10304 - there is already a very low respose at those wavelengths. I think the cutoff is combination of everything including processing, but definitely the signal is very weak there.

I really lack knowledge about this, hopefully someone can shed a bit more light on it.

[Ruud]

Thanks. I'll keep track of what you're doing. I think it's very interesting.

[robin_astro]

Yes there is definitely something wrong with the spectra at the far red end. They should not drop off like that. Cool stars continue to be as bright or even brighter into the infra red.  It is difficult to get good results at that end however without taking the precautions I mentioned to avoid 2nd order contamination. I would recommend trimming off the part of the  spectra above ~8000A. which is unreliable.  Below there the corrected spectra look excellent.

Robin