Jump to content

  •  

CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.

Photo

Quasar redshift or garbage?

  • Please log in to reply
26 replies to this topic

#1 descott12

descott12

    Soyuz

  • *****
  • topic starter
  • Posts: 3,933
  • Joined: 28 Sep 2018
  • Loc: Charlotte, NC

Posted 04 March 2024 - 04:46 PM

Hello all,

I recorded 33 minutes of B3 0754+394 last night with my new setup.

It is a C8 + Starizona reducer + ZWO OAG with a 220MM mini for guiding and an SA-100 and a 178MM for the science camera.

I still have some issues getting a good image in PHD2 but the guiding went pretty well and I got 100 20 second exposures.

But this thing is really dim. Supposedly mag 14.36 and I think all I recorded was noise.

In the attached screenshot, you can see a peak that would be near where Ha should be to calculate of redshift of 0.096 but I can't match the other peaks to any other balmer lines. I am pretty sure this is just garbage. What do you think?

Attached Thumbnails

  • B30754+394.png

  • mborland likes this

#2 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 04 March 2024 - 05:57 PM

You are following in the footsteps of a great pioneer of amateur spectroscopy  in the early electronic sensor era with this one. At the end of the last millennium, Maurice Gavin was pushing the boundaries, measuring Quasars using a forerunner of the Star Analyser. It was what he was doing that inspired me to get into spectroscopy. This article from the BAA Journal in 2000 show his spectrum of 0754+394 where he has identified H beta, gamma, delta but unfortunately this copy is poor and does not show the spectrum clearly

https://articles.ads...JBAA..110..127G

I will see if I can locate a better copy

 

Cheers

Robin


Edited by robin_astro, 04 March 2024 - 05:58 PM.


#3 descott12

descott12

    Soyuz

  • *****
  • topic starter
  • Posts: 3,933
  • Joined: 28 Sep 2018
  • Loc: Charlotte, NC

Posted 04 March 2024 - 06:03 PM

You are following in the footsteps of a great pioneer of amateur spectroscopy  in the early electronic sensor era with this one. At the end of the last millennium, Maurice Gavin was pushing the boundaries, measuring Quasars using a forerunner of the Star Analyser. It was what he was doing that inspired me to get into spectroscopy. This article from the BAA Journal in 2000 show his spectrum of 0754+394 where he has identified H beta, gamma, delta but unfortunately this copy is poor and does not show the spectrum clearly

https://articles.ads...JBAA..110..127G

I will see if I can locate a better copy

 

Cheers

Robin

Hi Robin, Thanks for the article. That will be very helpful. How much integration time do you think I will need to see something? I am guessing over an hour minimum


Edited by descott12, 04 March 2024 - 06:04 PM.


#4 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 04 March 2024 - 06:14 PM

 

https://articles.ads...JBAA..110..127G

I will see if I can locate a better copy

 

 

The copy on the BAA website is a little better, though it is only available to members. Here is a screen grab of the relevant image

 

Cheers

Robin

 

Maurice_Gavin_QSO_spectra.png


  • lambermo and descott12 like this

#5 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 04 March 2024 - 06:26 PM

Hi Robin, Thanks for the article. That will be very helpful. How much integration time do you think I will need to see something? I am guessing over an hour minimum

Maurice's spectrum was 30 min but he was using a rather low dispersion which is the key to going deep with the Star Analyser as it improves  the brightness of  the spectrum against the sky background which is normally the limiting factor. Here is one of mine  (QSO B0050+124) at similar magnitude, from the early days of the Star Analyser with a modified webcam 

http://www.threehill.../spectra_21.htm

 I used  32 Angstrom/7.4um pixel dispersion, though I did not note the total exposure. Only H alpha is obvious in the spectrum there.

 

Cheers

Robin


Edited by robin_astro, 04 March 2024 - 06:40 PM.


#6 descott12

descott12

    Soyuz

  • *****
  • topic starter
  • Posts: 3,933
  • Joined: 28 Sep 2018
  • Loc: Charlotte, NC

Posted 04 March 2024 - 07:14 PM

Maurice's spectrum was 30 min but he was using a rather low dispersion which is the key to going deep with the Star Analyser as it improves  the brightness of  the spectrum against the sky background which is normally the limiting factor. Here is one of mine  (QSO B0050+124) at similar magnitude, from the early days of the Star Analyser with a modified webcam 

http://www.threehill.../spectra_21.htm

 I used  32 Angstrom/7.4um pixel dispersion, though I did not note the total exposure. Only H alpha is obvious in the spectrum there.

 

Cheers

Robin

Good to know. I will adjust the dispersion and try again



#7 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 05 March 2024 - 07:16 AM

 using a rather low dispersion which is the key to going deep with the Star Analyser as it improves  the brightness of  the spectrum against the sky background which is normally the limiting factor. 

I should add a caveat to that. 

 

.....provided the features eg emission lines are wide eg as in QSO or supernovae for example.

 

If the feature is narrower than the resolution it tends to get "washed out" 

 

Robin


  • descott12 likes this

#8 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 05 March 2024 - 06:20 PM

Here is a quick 5min raw image tonight with the ALPY200. H alpha is clear. Also H beta and what is probably [OIII]

 

QSO_0754_394_alpy200_300s_20240305.png

 

Robin



#9 descott12

descott12

    Soyuz

  • *****
  • topic starter
  • Posts: 3,933
  • Joined: 28 Sep 2018
  • Loc: Charlotte, NC

Posted 05 March 2024 - 07:14 PM

Here is a quick 5min raw image tonight with the ALPY200. H alpha is clear. Also H beta and what is probably [OIII]

 

attachicon.gif QSO_0754_394_alpy200_300s_20240305.png

 

Robin

That is beautiful. Is that B3 0754+394?  Only 5 minutes??  There looks like some UV shifted into the visible too??

My understanding is that the Alpy 200 is mostly linear...what was the dispersion? I'd like to play with your image a bit.


Edited by descott12, 05 March 2024 - 07:16 PM.


#10 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 05 March 2024 - 07:57 PM

Yep your QSO

https://simbad.cds.u...nt=1E 0754+39.3

I got couple more exposures before the clouds rolled back in so I will reduce it tomorrow and can let you have the preprocessed image sum to play with. 

 

I use a 2nd order fit to 4 Balmer lines from the reference star for wavelength calibration of the ALPY200 but yes, the commercial version of the ALPY200 (which I used here) is close to linear compared with my original version as it uses a flat grating rather than an angled grism. 

 

Cheers

Robin



#11 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 05 March 2024 - 08:11 PM

Dispersion is ~20A/9um pixel (4.54um 2bin)



#12 descott12

descott12

    Soyuz

  • *****
  • topic starter
  • Posts: 3,933
  • Joined: 28 Sep 2018
  • Loc: Charlotte, NC

Posted 05 March 2024 - 08:21 PM

Yep your QSO

 

Thanks. My  spectrum didn't look ANYTHING like yours! I must try again. I was playing with yours a bit and the peaks match well for a z =0.096 and a shift of 459 A.

HA and HB were spot on. The 2 OIII peaks were close but H gamma was not a good match.

But I guesstimated a linear dispersion of 20.6 and I am assuming the redshift is constant across the spectrum (I think I recently read that that is not the case). But overall, this is pretty cool stuff.

Thanks for the data.

Attached Thumbnails

  • RobinsQuasar.png

Edited by descott12, 05 March 2024 - 08:22 PM.


#13 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 05 March 2024 - 08:22 PM

Here is a pro spectrum of it in NED (Lick 3m telescope)

https://ned.ipac.cal...:S:w2000_sp.png



#14 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 06 March 2024 - 02:00 PM

Here is a pro spectrum of it in NED (Lick 3m telescope)

https://ned.ipac.cal...:S:w2000_sp.png

Here is my reduced spectrum from last night with the ALPY200 (blue) overlaid on the spectrum from NED (orange). My total exposure was 35min but only 5 min was completely clear of cloud (looking at the counts, probably equivalent to ~15min cloud free)

 

QSO.png

 

QSO_0754_394_alpy200_300s_20240305_THO_raw_preproc.png


Edited by robin_astro, 06 March 2024 - 02:07 PM.

  • PrestonE likes this

#15 descott12

descott12

    Soyuz

  • *****
  • topic starter
  • Posts: 3,933
  • Joined: 28 Sep 2018
  • Loc: Charlotte, NC

Posted 06 March 2024 - 04:20 PM

Here is my reduced spectrum from last night with the ALPY200 (blue) overlaid on the spectrum from NED (orange). My total exposure was 35min but only 5 min was completely clear of cloud (looking at the counts, probably equivalent to ~15min cloud free)

 

attachicon.gif QSO.png

 

attachicon.gif QSO_0754_394_alpy200_300s_20240305_THO_raw_preproc.png

Very nice. I am looking forward to trying this one again once the sky clears.



#16 mborland

mborland

    Viking 1

  • *****
  • Posts: 617
  • Joined: 17 Mar 2020
  • Loc: Tucson

Posted 07 March 2024 - 09:28 AM

This is a very interesting discussion.

 

Perhaps one of you can tell me if there is a list of such objects that are within range of amateur equipment. I usually just click around in Stellarium, but seldom find anything with z>0.01.

 

I'm thinking if I can sample enough objects, perhaps I can establish a correlation between red shift and magnitude. Using magnitude as a proxy for distance, I could thus verify in a qualitative way the correlation between red shift and distance.

 

--Michael



#17 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 07 March 2024 - 12:09 PM

This is a very interesting discussion.

 

Perhaps one of you can tell me if there is a list of such objects that are within range of amateur equipment. I usually just click around in Stellarium, but seldom find anything with z>0.01.

 

I'm thinking if I can sample enough objects, perhaps I can establish a correlation between red shift and magnitude. Using magnitude as a proxy for distance, I could thus verify in a qualitative way the correlation between red shift and distance.

 

--Michael

Here is a million of them.

https://heasarc.gsfc.../milliquas.html

If you click browse this table you can build a list based on whatever parameters you set. Unless you have a large statistical sample, I dont think luminosity will correlate that well with distance though as active galaxies are not standard candles like type 1a supernovae for example. You could perhaps use galaxy size though as I did here as a very crude estimate in the local universe

 

https://britastro.or...c30f1b2e71339b3

 

Cheers

Robin


  • mborland likes this

#18 mborland

mborland

    Viking 1

  • *****
  • Posts: 617
  • Joined: 17 Mar 2020
  • Loc: Tucson

Posted 10 March 2024 - 12:21 AM

Here's my attempt from last night, using 6x720s exposure at gain 400 with an ALPY200 and ASI533MM on a 120-mm APO refractor.

 

QSO-B0754+394.png

 

My peaks don't agree as well with the professional spectrum as Robin's, a sign that I'm having some trouble with the wavelength calibration using A0V stars. The H-alpha line isn't clear, for one thing. The other Balmer lines are riding on a continuum that varies pretty strongly, making it doubtful that the locations are being found accurately.

 

Perhaps I need to iterate the wavelength calibration and response correction.  Or maybe I need to try to calibrate to something with strong Balmer emission lines.

 

--Michael

 


  • robin_astro likes this

#19 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 10 March 2024 - 07:13 AM

 The H-alpha line isn't clear, for one thing. The other Balmer lines are riding on a continuum that varies pretty strongly, making it doubtful that the locations are being found accurately.

 

 

Nice focus ! Better than mine. You have managed to split the [OIII] lines.

 

Tricks I use for wavelength calibration of the ALPY200 are to first process the reference star without using a flat which gives a "less peaky" shaped spectrum then rectify this spectrum,  just cropping out the 4 Balmer lines and using minimum smoothing before measuring the wavelengths. I use the dispersion tool in ISIS (remembering to tell it that I am looking at absorption rather than emission lines !)  This gives a linear fit with a typical RMS of 2-3 Angstrom. Not wonderful but good enough. 


Edited by robin_astro, 10 March 2024 - 09:56 AM.

  • mborland likes this

#20 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 10 March 2024 - 09:55 AM

 

Tricks I use for wavelength calibration of the ALPY200 are to first process the reference star without using a flat which gives a "less peaky" shaped spectrum then rectify this spectrum,  just cropping out the 4 Balmer lines and using minimum smoothing before measuring the wavelengths. I use the dispersion tool in ISIS (remembering to tell it that I am looking at absorption rather than emission lines !)  This gives a linear fit with a typical RMS of 2-3 Angstrom. Not wonderful but good enough. 

Raw spectrum

 

1.png

 

Remove Balmer lines to produce "continuum" and divide

 

2.png

 

Measure Balmer line positions in "rectified" spectrum and do linear fit

 

3.png


Edited by robin_astro, 10 March 2024 - 09:56 AM.

  • mborland likes this

#21 mborland

mborland

    Viking 1

  • *****
  • Posts: 617
  • Joined: 17 Mar 2020
  • Loc: Tucson

Posted 10 March 2024 - 08:16 PM

Robin,

 

That really worked well to get reliable linear fits. I got 1.3 A rms error from one star and 2.1 from another.

 

However, when I use those on my QSO+B0754+394 spectrum, I got a surprise.

QSO-B0754+394-V2.png

There's a systematic offset of about 20 A between my spectrum and the professional one. That's hard to explain.

 

I did this several times, with two different stars (one A0V and one A2V) and got essentially the same result. I also tried using spectrum-continuum instead of spectrum/continuum in the dispersion determination, thinking that the division might introduce a systematic error due to the slope of the continuum. In the end, the results were the same.

 

--Michael

 



#22 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 11 March 2024 - 07:28 AM

 

 

There's a systematic offset of about 20 A between my spectrum and the professional one. That's hard to explain.

 

 

Hi Michael,

 

The will be a small correction to heliocentric reference frame but 20 A is too large for that. What software are you using ?

 

Some possible things to check.

  

Check the position of some sky lines eg the O I lines at 5577.338   6300.304. Are they in the right place? (if using ISIS check the sky not removed box)

 

Has something in the spectrograph moved (was lose) between taking the reference and QSO spectra?  (Do you have matching calibration lamp spectra for both observations that you can check?

 

Has something has gone wrong with the geometric correction which has displaced the spectra in the X axis relative to each other?(eg If the slant/smile is not correct and the two spectra are at different positions along the slit)

 

Cheers

Robin


Edited by robin_astro, 11 March 2024 - 07:33 AM.


#23 robin_astro

robin_astro

    Mercury-Atlas

  • *****
  • Posts: 2,567
  • Joined: 18 Dec 2005
  • Loc: UK

Posted 11 March 2024 - 09:51 AM

 

Some possible things to check.

  

 

Also did your calibration work ok when applied to the reference star ie when fully processed including the response correction did it overlay precisely on the published spectrum ?



#24 mborland

mborland

    Viking 1

  • *****
  • Posts: 617
  • Joined: 17 Mar 2020
  • Loc: Tucson

Posted 11 March 2024 - 11:25 PM

Robin,

 

Thanks for the suggestions.

 

I'm using ISIS for most of my data processing. I perform the response normalization externally, so I can use a different program for smoothing the response prior to division.

 

I checked those two sky lines in my QSO-B0754+394 spectrum and find they are at 5589 A and 6300 A. They are pretty broad, which makes it hard to identify the exact wavelength, but they don't seem to be 20 A off. Unfortunately, I didn't take Neon/Ar calibration spectra since I found those not very useful with the ALPY 200.

 

 

I think I'm going to start from scratch with the processing.  Perhaps I made an error somewhere with my multiple calibration attempts etc.

 

--Michael


Edited by mborland, 12 March 2024 - 12:52 AM.


#25 mborland

mborland

    Viking 1

  • *****
  • Posts: 617
  • Joined: 17 Mar 2020
  • Loc: Tucson

Posted 12 March 2024 - 12:52 AM

Robin,

 

I reprocessed everything from scratch and the 20 A offset persists using the original reference/calibration star (HD43378).  The check of the reference star data to the MILES database after calibration and response correction looks good. I don't see any shifted lines.

HD43378Check.jpg

 

I took another star spectrum after taking the data for QSO+B0754+394, namely, HD97633. It isn't a good reference star choice for the QSO, but I have the data so I used it regardless. The results are compared here, zoomed in on the OIII lines. This new result is closer to the professional result by almost ~10A, not quite half the discrepancy I saw before.

ZoomedComparison.jpg

I'm not sure the source of the change. The vertical offset was the same to within ~2 pixels and the tilt was the same to within about 0.1 deg. I'll have to disassemble things and ensure there's nothing loose, I guess.

 

Also, I think I'll try finding a bright Z=0 object with strong Balmer emission lines to practice, then come back to high red-shift objects.

Dave, sorry I high-jacked your thread.

--Michael


Edited by mborland, 12 March 2024 - 12:53 AM.



CNers have asked about a donation box for Cloudy Nights over the years, so here you go. Donation is not required by any means, so please enjoy your stay.


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics