Struggling with my Pollux

[Hawksmoor]

I have been using my homemade spectrometer attached to my127mm Refractor to obtain spectral profiles for Castor, Pollux, Regulus and Arcturus. I have used  BASS software to process the data. I have calibrated Castor using the clear hydrogen calibration lines and obtained a response curve for my set up on the night. I have calibrated the profiles for Pollux, Regulus and Arcturus  using data points from Castor and then applied  response curve to them. 

Whilst, Regulus and Arcturus look at least to be the right sort of shape Pollux looks to be more like an A class star  than a K. The original wavelength calibrated uncropped and response modified chart looks better than the profiles after further processing.  Anybody got any ideas as to what has gone wrong? I am losing the will to live, which at my age is dangerous!  I would be very grateful for help.

I have attached  a series of my charts in png format

George

 

[rl]

I have to confess I'm struggling a bit to identify features for certain.  Are these single exposures; would stacking help improve the signal-to-noise?

What are the y-axis units? ADC counts?

The Castor trace on your last plot (blue line) shows most clearly what are probably the H-beta and H-gamma lines. The Planck curves do sort of follow the expected pattern based on temperature.

[Hawksmoor]

14 hours ago, rl said:

I have to confess I'm struggling a bit to identify features for certain.  Are these single exposures; would stacking help improve the signal-to-noise?

What are the y-axis units? ADC counts?

The Castor trace on your last plot (blue line) shows most clearly what are probably the H-beta and H-gamma lines. The Planck curves do sort of follow the expected pattern based on temperature.

Hi rl

You must be getting fed up with my spectral ramblings but thank you for having a look and commenting.

The captured images (used SharpCap 4.1) were stacked 50 x png frames using BASS software (unaligned as recommended in the BASS software manual). I used Dark frames stacked and applied to the lights to reduce noise. The spectra were flipped horizontally and rotated approx 1 deg for horizonal. The spectra were then binned and areas each side chosen for subtraction.

What appears to be noise may be noise but also an effect created by my transmission grating, which has 500 lines per mm rather than  say a star Analyser at 100 lines per mm.

The Y Axis units on the cropped profiles were normalised using the visual magnitude for each star - applied using the  button in the normalise image drop down menu in BASS software. The uncropped profile was as  was recorded by my QHY5-11mono camera, so presumably ADC  counts.

The Castor plot was used to calibrate the other stars using 3 points (the 0 point, H-beta and H-gamma lines lines). The linear match used was, according to the BASS software, very close to spot on.

The last uncropped profile I posted seems more or less as I might have expected, I was however perplexed as to why on the cropped versions, where in particular I applied a response curve created from the Castor profile and the Miles data base  entry for Castor, the curve for Pollux seemed to change from convex to concave. Being both K stars, I thought it should look more like Arcturus and less like Regulus. I have checked the procedure I used  with the BASS manual , the BAA site and Robin's 3 Hill's site and for the life of me I cannot work out what I have done wrong. 

If no one can explain it - In desperation, I shall go back to my original captures  and use a different set to check that I haven't used  the wrong star capture  for Pollux from the many stars  taken on the night. Yes I am that disorganised and stupid - I put it down to old age and to spending too much time in the cold at the dead of night. I have to say that I am not convinced that this is the case, as the Pollux profile on the uncropped version looks ok. I am pretty sure it is some button or other in BASS that I am pushing or haven't pushed in the correct order that is causing the problem. Monkeys and typewriters!

Thanks once again for your interest , all help is appreciated .

George

 

 

 

[robin_astro]

Why do the zero orders appear at different positions in the different spectra ? If the equipment was unchanged the dispersion equation is the same and the zero orders should line up 

Cheers

Robin

[robin_astro]

There is definitely something wrong with the calibration I am afraid. The H beta, gamma are in the right place in Castor but are incorrect in Regulus (shifted well to the blue and still show some signal below 3500A where the camera will not be sensitive.  I suspect this is connected with the non alignment of the zero order

Is this a slitless system ?  If so you should include the zero order as one of the points and the fit (linear or non linear) should force the zero order to be at zero wavelength so you have a reference point to align all the spectra to. Also unless you have  very good tracking or guiding you should definitely align your sub-exposures (single point alignment on the zero order) otherwise any tracking errors will smear the spectrum.

Is the grating in the converging beam? If so the aberrations (chromatic coma and field curvature) will be severe. I deliberately  chose low dispersion gratings for the Star Analysers to minimise these effects.

Cheers

Robin

Edited June 2 by robin_astro
clarification

[robin_astro]

12 hours ago, Hawksmoor said:

The linear match used was, according to the BASS software, very close to spot on.

with a 500l/mm grating in the converging beam you should expect some non linearity because of the large dispersion angles involved. The low dispersion Star Analysers are very close to linear because the angles involved are small.

Cheers

Robin

[Hawksmoor]

13 hours ago, robin_astro said:

Why do the zero orders appear at different positions in the different spectra ? If the equipment was unchanged the dispersion equation is the same and the zero orders should line up 

Cheers

Robin

Hi Robin

Thank you so much for your very helpful comments which I shall try to answer.

I too was perplexed as to why the zero orders didn't line up. As the BASS software is a bit of a black box  to me and I had never tried to calibrate more than two stars in one go, I assumed it was an affect of applying calibration.  In retrospect, I haven't noticed this before, when calibrating one star using another. So as you said, it is therefore  down to capture. On the night, once inserted and aligned horizontal, I did not change or move the spectrometer relative to the scope but I did fiddle with the focus for each star and I guess this might change dispersion and result in the mismatch of zero orders - If this is so, it is 'very much an old man's silly mistake on a cold night in the backyard'. The only other changes I made were exposure and gain in SharpCap4.1, which I didn't think would make a difference?

My spectrometer is a slitless system working on a converging beam basis (no collimator or camera lens). I used  a blazed 500l/mm transmission grating because I found a supplier of an affordable technical film in Israel on line. I did however design a housing for the grating and QHY5ii-mono camera, which both held the camera at the appropriate angle (I  used the on-line spreadsheet calculator) and close enough to the sensor so as to keep the first order spectrum on it.

For Calibration of Castor I used the the zero order, H-gamma and H Beta. I then finally added H-alpha to test linearity.

My scope wasn't guided on the night but is on a permanent pier and EQ6Pro - so polar aligned (ish as I usually guide for long sub exposure imaging). I stacked the captured spectra using the facility in BASS software. The help guidance says not to use the align button. I didn't understand why but did follow the guidance.

Thank you so much for your assistance and  helpful and interesting posts on your web-site.

George

[Hawksmoor]

On 02/06/2024 at 01:23, robin_astro said:

There is definitely something wrong with the calibration I am afraid. The H beta, gamma are in the right place in Castor but are incorrect in Regulus (shifted well to the blue and still show some signal below 3500A where the camera will not be sensitive.  I suspect this is connected with the non alignment of the zero order

Is this a slitless system ?  If so you should include the zero order as one of the points and the fit (linear or non linear) should force the zero order to be at zero wavelength so you have a reference point to align all the spectra to. Also unless you have  very good tracking or guiding you should definitely align your sub-exposures (single point alignment on the zero order) otherwise any tracking errors will smear the spectrum.

Is the grating in the converging beam? If so the aberrations (chromatic coma and field curvature) will be severe. I deliberately  chose low dispersion gratings for the Star Analysers to minimise these effects.

Cheers

Robin

Great suggestion regarding the single point alignment on the zero order. tried this and saw an immediate improvement in the line profile. 

Thank you

George