Attempting an SA100 spectrum on 2Lyr and 4Lyr/5Lyr

[SteveBz]

Hi Folks,

After a lot of hassle I finally have a setup that I hope will work.  I have a small Newtonian (114 f/4) with a QHY5iii178mm mini, piggy-backing a large Newtonian (see photo).  The whole lot is guided with an OAG.  Last night I tried it out for the fist time with mixed results.

I really wanted to take a comparative spectrum of the binary stars 4Lyr/5Lyr known, together with 2Lyr, as the 'Talons of the Eagle'.  Here's 4- and 5-Lyr through the main scope - about 10 mins exposure:

And here through the piggy-backed mono (0.5 sec exposure):

Here is a single frame spectrum through the SA100 (0.5 sec exposure):

It just raises so many questions.  Is it even in focus?  Each spectrum shows 2 horizontal parallel lines and I guess the dark horizontal band down the middle is the secondary mirror.  I tried to get it in focus with a helical focuser so that the central obstruction didn't show, but it seemed to be impossible.  Is this as good as it gets?  I didn't expect the stars in the second image to be at right angles to the first image, but then I remembered I'd oriented parallel to the DEC axis it so that drift in RA would show up as increased breadth in the spectrum.

I took 20 each of these and also of 2Lyr (not shown) and for 2 Lyr, I tried very unsuccessfully to stack in BASS.  This image seems to be highly stretched, but BASS doesn't seem to stretch at all, and I couldn't even see the star or the spectrum in BASS.  I'm sure I can play around with the stretching and exposure parameters, but what about a), the focus and b), separating or stacking 2 spectra on the same image.  I could manually crop the 20 images, but BASS wants them all to be the same size and I'm not sure how to do that manually.  I was hoping to try to measure and compare the metallicity of the stars.  Each star in fact has a hidden companion, probably a red dwarf.  I don't imagine I can detect these, but it would be interesting to see what I can detect.

I'm a bit lost here and your input would be very welcome.

Kind regards

Steve

Edited August 29, 2021 by SteveBz

[robin_astro]

Hi Steve,

You have a good setup which will work well but I would suggest taking a step back and  start with a star with known clear features to focus on. You can then move to stars with less obvious features keeping the same focus. Main sequence A stars are often used as  they show clear Balmer absorption lines but an even better star at this time of year to cut your teeth on is P Cygni which shows very obvious Hydrogen and Helium emission lines.  The image you have posted is way too saturated to be able to see anything.  It is important that none of the pixels in the spectrum is saturated. To see the features clearly err on the under exposed side. 

Have you downloaded the Star Analyser manual ? I have tried to include some useful tips in there too

Cheers

Robin

 

[robin_astro]

There are some examples on my BAA page here which give you an idea what in focus spectra should look like 

https://britastro.org/observations/observation.php?id=20201216_234948_8cabda965bfe692f

https://britastro.org/observations/observation.php?id=20210406_144443_9e1c6a4cf219d14d

note though at that the low resolution of the Star Analyser as used in its standard configuration, many star spectra look almost featureless. It works best on stars with strong features

Cheers

Robin

Edited August 29, 2021 by robin_astro

[robin_astro]

Also have you tried imaging through the piggyback scope without the star Analyser? If you cannot get good well exposed sharp images without it, you will not get good spectra with it

Cheers

Robin

[SteveBz]

1 hour ago, robin_astro said:

Hi Steve,

You have a good setup which will work well but I would suggest taking a step back and  start with a star with known clear features to focus on. You can then move to stars with less obvious features keeping the same focus. Main sequence A stars are often used as  they show clear Balmer absorption lines but an even better star at this time of year to cut your teeth on is P Cygni which shows very obvious Hydrogen and Helium emission lines.  The image you have posted is way too saturated to be able to see anything.  It is important that none of the pixels in the spectrum is saturated. To see the features clearly err on the under exposed side. 

Have you downloaded the Star Analyser manual ? I have tried to include some useful tips in there too

Cheers

Robin

 

I'll try P Cygni on the next clear night.

[SteveBz]

1 hour ago, robin_astro said:

There are some examples on my BAA page here which give you an idea what in focus spectra should look like 

https://britastro.org/observations/observation.php?id=20201216_234948_8cabda965bfe692f

https://britastro.org/observations/observation.php?id=20210406_144443_9e1c6a4cf219d14d

note though at that the low resolution of the Star Analyser as used in its standard configuration, many star spectra look almost featureless. It works best on stars with strong features

Cheers

Robin

Beautiful spectra.  I hope I can emulate them.

[SteveBz]

59 minutes ago, robin_astro said:

Also have you tried imaging through the piggyback scope without the star Analyser? If you cannot get good well exposed sharp images without it, you will not get good spectra with it

Cheers

Robin

No I haven't. Scary thought. I'll have a think about how to test it.

[SteveBz]

So 2 test objectives for the next night. P Cygni and plain scope.

(I attached the manual for other readers).

Tx.

StarAnalyser100manual.pdf

Edited August 29, 2021 by SteveBz

[SteveBz]

So it was quite clear tonight until about 11pm.  In fact just as I was about to try P Cygni, as you suggested.  Then it clouded over. But I did get another go at 2Lyr, which came up 2 hours earlier.  Here's the image, cropped at source and then stacked with Bass (about 20 images of 30 secs each):

And here's the spectrum, calibrated.

As you can see I've calibrated it to the the H-Beta line and the star (@ 0 nm).  Given that messing with the camera is sort of outside my comfort zone at this stage, I'm going to look for a pair of binaries separated E-W, not N-S, so I can use the same technique to take a spectrum of both stars.

In spite of all my earlier experiments, this feels like the first real one.

Any suggestions welcome, as always.

Kind regards

Steve.

 

[robin_astro]

Excellent!  You are on your way. Lots of nice Balmer absorption lines in this A star.  The shape of the continuum perhaps looks a bit flat topped so double check that  there are no saturated pixels in that region. The small peaks will likely be the zero orders of faint background stars.  

Cheers

Robin

[SteveBz]

4 hours ago, robin_astro said:

The small peaks will likely be the zero orders of faint background stars.  

OK, thanks for that, I wondered what they were.

[SteveBz]

9 hours ago, robin_astro said:

The small peaks will likely be the zero orders of faint background stars.  

So you were right.  When I cropped it more tightly some of the peaks disappeared.  (That one at about 250 is a little stubborn).  The H-alpha line seems very faint, as do the telluric lines.  When I look at the unstacked lines, they do indeed look a little intense at the bright points..  Next clear night I'll play some more with the exposure and the focus, and also more tightly cropping the area of interest frame on the sensor to avoid background stars.  The image is about 1 degree away from horizontal.  Should I fiddle or leave it? 

Edited September 4, 2021 by SteveBz

[SteveBz]

I'm trying to produce a spectrum corrected for instrument response.  I've followed the instructions and this is what I get.  Seems a bit noisy.

Do I need to produce lots of correction profiles and stack them?

Is it possible to find standard profiles for individual cameras, eg QHY5iii178mm?

Kind regards

Steve.

  

 

[robin_astro]

Something's gone wrong there I am afraid. The absorption lines have change to emission lines and the spectrum continuum should rise from red to blue, not be flat if you have corrected for the response. Take a look at my page here for how to correct for response of the instrument (and atmosphere)

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

also my presentation " Low Resolution Slitless Spectroscopy - Robin Leadbeater on the BAA website here

https://www.britastro.org/downloads/15701

Cheers

Robin

[SteveBz]

Hi People,

So here's P Cygni.  15 sec x 40 stacked, tilted and croped.

And what I hope is the sensor corrected spectrum:

Do I have it the right way round?

I feel the little Bresser 114mm and the 178mm mini are working well together. I put a helical focuser on the Bresser for fine focal adjustments really because it was very light.

Tx

Steve.

[robin_astro]

A very nice spectrum with clear Hydrogen and Helium emission lines in the powerful stellar wind of this blue supergiant.  It looks like you have rectified the spectrum though (flattened it by removing the shape of the continuum)  rather than correcting for the response. That is fine if that is what you intended but a response corrected spectrum of this hot star should slope upwards from red to blue as in this example in the BAA database by Kevin Gurney

https://britastro.org/specdb/data_graph.php?obs_id=1658

The full details of how to do this are in this document on my website here

http://www.threehillsobservatory.co.uk/astro/Relative_flux_calibration_20201007.pdf

but  the  steps in outline are

1. Take a spectrum of a star with a known published spectrum (eg  you could use your 2 Lyr  which is type A2v a version of which from the Pickles library can be found in your spectrum processing program eg Visual spec, ISIS, RSpec, BASS)

2. Produce a raw digitised spectrum of it, doing dark subtraction and subtracting the sky background and then wavelength calibrate it

3. Divide this spectrum by the published spectrum of the star to produce the response

4. Produce a raw spectrum of the target star ie P Cygni

5. Divide  this raw spectrum by the response produced in 3 to produce the response corrected spectrum of the target star

 

Cheers

Robin

[robin_astro]

56 minutes ago, SteveBz said:

what I hope is the sensor corrected spectrum

Note than when you are correcting for the response (also known as flux calibration) you are not only correcting for the sensor response but also everything else between the sensor and the starlight arriving at the top of our atmosphere. This includes the effect of the atmosphere, the telescope optics and the transmission of the grating which is wavelength dependent and as significant as  the camera sensor. By measuring the response by taking a known star we include  all these effects in one go, all explained in the document above

Cheers

Robin

Edited September 4, 2021 by robin_astro

[SteveBz]

Hi People,

That was quite hard and probably a bit wobbly.  Here is what I hope is a response curve produced from 2 Lyr applied to the 6 Lyr A (if that's the proper nomenclature - ie the main star of 6 Lyr):

I've had to save the response file as a dat file and then loaded it using 'Add/Image', then divided.  I'm sure there is an easier way!

Kind regards

Steve.

PS  Maybe this is it.  There is a chart edit option and you can enter the instrument response file.  Looks a bit better, but obviously the same graph.

Edited September 8, 2021 by SteveBz

[robin_astro]

Nice result.  The pair are called 6 Lyr, 7 Lyr or Zet01 Lyr, Zet02 Lyr or HD173648,HD173649

https://simbad.u-strasbg.fr/simbad/sim-id?Ident=*zet01 Lyr

https://simbad.u-strasbg.fr/simbad/sim-id?Ident=*zet02 Lyr

How does the spectrum of 6 Lyr compare with that of 2 Lyr ? Can you see  a difference ?

Cheers

Robin

 

Edited September 8, 2021 by robin_astro
typo

[Merlin66]

Steve,

Well done mate!

A couple of points:

We still tend to use Angstrom rather than nm for our spectra.

Use the "Chart/Crop X axis range" to hide the ends of the profile say 3700A to 7500A

Onwards and Upwards

 

 

 

[SteveBz]

2 hours ago, Merlin66 said:

We still tend to use Angstrom rather than nm for our spectra.

Yes, I did see that and there have even been (quite intemperate) arguments on SGL regarding it.  Why do we do that since I haven't even got the A-circle-thingy on my keyboard?

It's not as much hassle as the Light-year vs parsec argument, since it's still metric and it just needs an extra x10 or /10, which we can all probably live with 😄

2 hours ago, Merlin66 said:

Use the "Chart/Crop X axis range" to hide the ends of the profile say 3700A to 7500A

Good point.  I tried to do that on the second version.  Should I have cropped it further?

Finally, do you build lot's of instrument response files over time and then 'stack' them to get a more accurate response curve? 

[robin_astro]

4 hours ago, SteveBz said:

Yes, I did see that and there have even been (quite intemperate) arguments on SGL regarding it. 

Both are used for optical spectra in the professional community, Angstrom traditionally, particularly when identifying a particular line eg H alpha 6563 Angstrom. In publications though some Journals now insist on the SI unit nm

Cheers

Robin

[robin_astro]

4 hours ago, SteveBz said:

Finally, do you build lot's of instrument response files over time and then 'stack' them to get a more accurate response curve? 

Steve,

As I said earlier what you are measuring is not just the instrument response. (This is a common misunderstanding which is why I emphasise this in the document on my website). It also includes the effect of the atmosphere which is significant and changes from night to night and with height above the horizon (air mass). Although you can get a rough result by using a response from a previous night, provided the difference in air mass is not too great, for the most accurate results it is a good idea to take a response on the night at similar air mass, particularly for targets low in the sky where the air mass is high. See this example by Christian Buil of  the effect of air mass on the spectrum

http://www.astrosurf.com/buil/atmosphere/transmission1.png

From his page explaining the effect of the atmosphere

http://www.astrosurf.com/buil/atmosphere/transmission.htm

Francois Teyssier has made a nice  spreadsheet to find suitable bright stars with known spectra near your target. There is a link to it in my document on flux calibration

Cheers

Robin 

[vlaiv]

3 minutes ago, robin_astro said:

As I said earlier what you are measuring is not just the instrument response.

How about obtaining instrument response with Tungsten-Halogen lamp acting as artificial star (pinhole and enough distance)?

Are those precise enough? I don't mean special ones that are classified as calibration source (too expensive) - but regular, of the shelf ones?

[robin_astro]

13 minutes ago, vlaiv said:

How about obtaining instrument response with Tungsten-Halogen lamp acting as artificial star (pinhole and enough distance)?

Are those precise enough? I don't mean special ones that are classified as calibration source (too expensive) - but regular, of the shelf ones?

You can get a rough response by this method. In fact this is essentially what happens with a slit spectrograph when you take a flat, all explained in my document. At the end of the day though it is not particularly useful as you still have to include the effect of the atmosphere which is done by taking a standard star. Professionals do tend to separate the effect of the instrument and of the atmosphere using a standard instrument response and then correct for the atmosphere separately by taking telluric standards on the night but they rely on much more stable atmospheric conditions. In general most amateurs take a nearby reference star on the night which includes all effects in one go.

Cheers

Robin

Edited September 9, 2021 by robin_astro