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

What HFR star size are you getting with long FL (2800mm)?

  • Please log in to reply
21 replies to this topic

#1 GeneralT001

GeneralT001

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1625
  • Joined: 06 Feb 2012
  • Loc: Nanaimo, BC

Posted 21 April 2017 - 02:13 AM

Hi,

 

First time trying to setup/use the C11 Edge and the best HFR I could achieve was 3.75. Is this to be expected at long FL or are my skies just craptacular? It seemed I got somewhat better HFR with the 100mm scope, so is it just a matter of the FL?



#2 happylimpet

happylimpet

    Surveyor 1

  • *****
  • Posts: 1864
  • Joined: 29 Sep 2013
  • Loc: Southampton, UK

Posted 21 April 2017 - 03:53 AM

Is that in arcseconds? Not amazing if so.



#3 spokeshave

spokeshave

    Viking 1

  • -----
  • Posts: 973
  • Joined: 08 Apr 2015

Posted 21 April 2017 - 04:25 AM

If you're referring to the HFR metric in SGP, it is measured in pixels. As such, any comparison to another system with a different image scale is meaningless. It is only useful as a relative metric - meaning it can tell you how good your focus is or if focus is drifting, or even how much seeing changes from night to night, but it is useless as a comparison to other scope/camera combinations.

Tim

#4 leemr

leemr

    Mariner 2

  • -----
  • Posts: 246
  • Joined: 21 Mar 2014

Posted 21 April 2017 - 04:46 AM

There's also variations in how these things are calculated, so to have any meaningful discussion it would help if everyone stated what application they measured it in, or better yet, if everyone used the same application for measurement. I'd recommend using PixInsight's FWHMEccentricity script with a consistent model function (I think I saw a paper once that recommends Moffat 4), then normalized to arc-seconds for discussion purposes. Sorry I cannot help with the exact question as I'm only imaging at 840mm / 0.9"px



#5 kd4pbs

kd4pbs

    Apollo

  • *****
  • Posts: 1139
  • Joined: 13 Feb 2013

Posted 21 April 2017 - 09:34 AM

Using SGP, and the ASI1600 camera, at 2000mm I can obtain between 2.25 and 3.25 HFR numbers with autofocus depending on which filter I'm on and duration of exposure.  So while it might not be a "cross platform" number, if you're using your 1600, it is probably close enough to gauge.  I'd say adding the extra focal length and adjusting for variations in exposure, that might be about right.  Then again, my scope probably is not as good optically as yours, since mine is not the coma-free type SCT, and I believe I could refine my collimation a bit.
As long as your stars look round, and you're focused well, I think you're going to find it's as good as it gets.  Focal length, as well as seeing conditions plays a huge part on how small your stars appear.



#6 GeneralT001

GeneralT001

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1625
  • Joined: 06 Feb 2012
  • Loc: Nanaimo, BC

Posted 21 April 2017 - 10:02 AM

Thanks all for the replies.

 

I should have specified that yes, I am using SGP so that number (3.75) is pixels? I am using the ASI1600mm. I was not able to get the Auto Focus routine in SGP to show a "V" or a "U" shape.

 

If my initial HFR is 3.75, does that mean I have to go out of focus to 13 pixels (3.5 x HFR) to get my numbers for a step size calculation. It seems you would really have to move the focuser a fair bit for that?



#7 bobzeq25

bobzeq25

    Cosmos

  • *****
  • Posts: 9368
  • Joined: 27 Oct 2014

Posted 21 April 2017 - 10:07 AM

Assuming you're using the full 2800mm, that's an image scale of 0.55 arc sec per pixel.   So, if 3.75 is pixels, that's 2 arc sec.  I'm jealous, my skies won't let me get near there.  <smile>  I wouldn't dream of going for that image scale, it would be a waste.  0.9 is my usual number for "high resolution", widefield my usual choice.


Edited by bobzeq25, 21 April 2017 - 10:08 AM.


#8 GeneralT001

GeneralT001

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1625
  • Joined: 06 Feb 2012
  • Loc: Nanaimo, BC

Posted 21 April 2017 - 10:28 AM

Assuming you're using the full 2800mm, that's an image scale of 0.55 arc sec per pixel.   So, if 3.75 is pixels, that's 2 arc sec.  I'm jealous, my skies won't let me get near there.  <smile>  I wouldn't dream of going for that image scale, it would be a waste.  0.9 is my usual number for "high resolution", widefield my usual choice.

Hi Bob,

 

Can you confirm that the image scale is .55 arcsec? I came out with .28 arcsec? Whats the formula to calculate the seeing in arc sec for that, that you used to get 2 arcsec?

 

 

OK, did some research and you got that .55 arcsec based on 2x2 binning? So, I probably need to bin 2x2 with the ASI1600 with this scope?


Edited by GeneralT001, 21 April 2017 - 10:54 AM.


#9 kd4pbs

kd4pbs

    Apollo

  • *****
  • Posts: 1139
  • Joined: 13 Feb 2013

Posted 21 April 2017 - 11:34 AM

Autofocus with SGP takes some tweaking to get right.  It's a matter of getting it perfectly in focus with, say a focus mask first.  From there, you'll know where your center of focus is.  Starting with the filter focus position set to the focuser position you found using a mask, and using at least with 9 auto focus data points, play with the step size until the focus routine starts far enough out that you can see the stars are little doughnuts instead of stars.  Make sure the "disable smart focus" is checked when you're using an SCT - this is very important!

Oh, and I get .28 arcseconds per pixel as well - and assume that Bob's talking about binning 2x2.  And yes, you should be binning 2x2 or even 3x3 when running an autofocus routine - there's no real benefit to going 1x1.


Edited by kd4pbs, 21 April 2017 - 11:36 AM.


#10 bobzeq25

bobzeq25

    Cosmos

  • *****
  • Posts: 9368
  • Joined: 27 Oct 2014

Posted 21 April 2017 - 01:57 PM

 

Assuming you're using the full 2800mm, that's an image scale of 0.55 arc sec per pixel.   So, if 3.75 is pixels, that's 2 arc sec.  I'm jealous, my skies won't let me get near there.  <smile>  I wouldn't dream of going for that image scale, it would be a waste.  0.9 is my usual number for "high resolution", widefield my usual choice.

Hi Bob,

 

Can you confirm that the image scale is .55 arcsec? I came out with .28 arcsec? Whats the formula to calculate the seeing in arc sec for that, that you used to get 2 arcsec?

 

 

OK, did some research and you got that .55 arcsec based on 2x2 binning? So, I probably need to bin 2x2 with the ASI1600 with this scope?

 

Let's see if I messed up.  Take the focal length and divide it by 200 (actually 206, note the sig).  2800/200 = 14.  Take that number divide it into the pixel size.  3.8/14 = 0.27.

 

Yep, you're right.  Brain fade <grin>.

 

That makes the HFR 1 arc sec.  Almost too good to be true.

 

And the image scale pretty oversampled, by any standard.  I believe with a CMOS camera it's equally effective to bin in processing.  StarTools actually does variable binning, don't know about PI, much less PS.

 

Minor point.  This follows the well known rule.  Small pixels for short focal lengths, big ones for long.  A focal length reducer also deals somewhat with the situation, has plusses and minuses.


Edited by bobzeq25, 21 April 2017 - 02:00 PM.


#11 spokeshave

spokeshave

    Viking 1

  • -----
  • Posts: 973
  • Joined: 08 Apr 2015

Posted 21 April 2017 - 05:03 PM

Keep in mind that HFR is a radius, FWHM is a diameter. So you would need to double HFR to compare it to FWHM. Even then, though, you can't compare even twice the HFR to FWHM. They are calculated completely differently. HFR is the radius at which half of the light is concentrated within the radius and the other half is outside of the radius. For a Gaussian distribution, twice the HFR encompasses approximately 1.2 standard deviations whereas FWHM encompasses approximately 2.4 standard deviations. So HFR (actually twice the HFR) is quite a bit narrower. or the same Gaussian distribution, an HFR of 1" corresponds to a FWHM of 4.8".

Tim
  • happylimpet and Jon Rista like this

#12 GeneralT001

GeneralT001

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1625
  • Joined: 06 Feb 2012
  • Loc: Nanaimo, BC

Posted 21 April 2017 - 05:13 PM

Autofocus with SGP takes some tweaking to get right.  It's a matter of getting it perfectly in focus with, say a focus mask first.  From there, you'll know where your center of focus is.  Starting with the filter focus position set to the focuser position you found using a mask, and using at least with 9 auto focus data points, play with the step size until the focus routine starts far enough out that you can see the stars are little doughnuts instead of stars.  Make sure the "disable smart focus" is checked when you're using an SCT - this is very important!

Oh, and I get .28 arcseconds per pixel as well - and assume that Bob's talking about binning 2x2.  And yes, you should be binning 2x2 or even 3x3 when running an autofocus routine - there's no real benefit to going 1x1.

Yeah, I don't have a focus mask for the C11, so I was using the SGP frame/focus to get it as sharp as I could with my eyeball.....wouldn't that be close enough as a starting point?

 

I didn't set any filter focus position. I'm using the Lum filter - do I have to enter that position in the filter before running the focus routine?

 

No, I didn't disable smart focus, will do that tonight.



#13 GeneralT001

GeneralT001

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1625
  • Joined: 06 Feb 2012
  • Loc: Nanaimo, BC

Posted 21 April 2017 - 05:23 PM

Keep in mind that HFR is a radius, FWHM is a diameter. So you would need to double HFR to compare it to FWHM. Even then, though, you can't compare even twice the HFR to FWHM. They are calculated completely differently. HFR is the radius at which half of the light is concentrated within the radius and the other half is outside of the radius. For a Gaussian distribution, twice the HFR encompasses approximately 1.2 standard deviations whereas FWHM encompasses approximately 2.4 standard deviations. So HFR (actually twice the HFR) is quite a bit narrower. or the same Gaussian distribution, an HFR of 1" corresponds to a FWHM of 4.8".

Tim

 

I'm willing to believe that my FWHM is closer to the 4 or 5 range!

 

Do I have this totally wrong in my head:

 

- with a longer FL (2800mm) the FWHM of a star has to be larger than with a short FL (200mm) because the star is magnified more and therefore covers more pixels. Is that correct?



#14 bobzeq25

bobzeq25

    Cosmos

  • *****
  • Posts: 9368
  • Joined: 27 Oct 2014

Posted 21 April 2017 - 05:41 PM

 

Keep in mind that HFR is a radius, FWHM is a diameter. So you would need to double HFR to compare it to FWHM. Even then, though, you can't compare even twice the HFR to FWHM. They are calculated completely differently. HFR is the radius at which half of the light is concentrated within the radius and the other half is outside of the radius. For a Gaussian distribution, twice the HFR encompasses approximately 1.2 standard deviations whereas FWHM encompasses approximately 2.4 standard deviations. So HFR (actually twice the HFR) is quite a bit narrower. or the same Gaussian distribution, an HFR of 1" corresponds to a FWHM of 4.8".

Tim

 

I'm willing to believe that my FWHM is closer to the 4 or 5 range!

 

Do I have this totally wrong in my head:

 

- with a longer FL (2800mm) the FWHM of a star has to be larger than with a short FL (200mm) because the star is magnified more and therefore covers more pixels. Is that correct?

 

Generally not.  It's complex, but two things that figure in are the higher resolution of a large aperture scope, and the quality of the optics.  The "magnification of the star" doesn't really figure in to it, stars are (without optics) extremely tiny points of light that cannot be "magnified" even by a large telescope.  For example, Betelgeuse is 0.05 arc sec, about one-tenth of a pixel at your image scale. 

 

What you're seeing are things like imperfections (such as optical aberrations) and noise, not a magnified star.  This chart is fun, can't guarantee the accuracy, but the general picture is correct.

 

http://www.hyperiont...omparison16.php

 

And, of course, blur due to atmospheric turbulance, aka "seeing" trumps everything.  Rough idea (not intended to be precise).  If your FWHM is 4 to 5, it's likely that your seeing is 3 to 4.  Note the sig, please.


Edited by bobzeq25, 21 April 2017 - 05:50 PM.


#15 spokeshave

spokeshave

    Viking 1

  • -----
  • Posts: 973
  • Joined: 08 Apr 2015

Posted 21 April 2017 - 05:42 PM

Keep in mind that HFR is a radius, FWHM is a diameter. So you would need to double HFR to compare it to FWHM. Even then, though, you can't compare even twice the HFR to FWHM. They are calculated completely differently. HFR is the radius at which half of the light is concentrated within the radius and the other half is outside of the radius. For a Gaussian distribution, twice the HFR encompasses approximately 1.2 standard deviations whereas FWHM encompasses approximately 2.4 standard deviations. So HFR (actually twice the HFR) is quite a bit narrower. or the same Gaussian distribution, an HFR of 1" corresponds to a FWHM of 4.8".

Tim

 
I'm willing to believe that my FWHM is closer to the 4 or 5 range!
 
Do I have this totally wrong in my head:
 
- with a longer FL (2800mm) the FWHM of a star has to be larger than with a short FL (200mm) because the star is magnified more and therefore covers more pixels. Is that correct?


If the FWHM is measure in pixels, then yes. However, FWHM is typically measured in arc seconds, which is corrected for image scale. As such, it represents the actual angular diameter of the star, independent of focal length. In this case, three primary factors influence FWHM - aperture (more aperture means smaller Airy disk), seeing and focus.

Tim
  • Jon Rista likes this

#16 akulapanam

akulapanam

    Surveyor 1

  • *****
  • Posts: 1866
  • Joined: 26 Aug 2012

Posted 22 April 2017 - 10:36 PM

 

 

Keep in mind that HFR is a radius, FWHM is a diameter. So you would need to double HFR to compare it to FWHM. Even then, though, you can't compare even twice the HFR to FWHM. They are calculated completely differently. HFR is the radius at which half of the light is concentrated within the radius and the other half is outside of the radius. For a Gaussian distribution, twice the HFR encompasses approximately 1.2 standard deviations whereas FWHM encompasses approximately 2.4 standard deviations. So HFR (actually twice the HFR) is quite a bit narrower. or the same Gaussian distribution, an HFR of 1" corresponds to a FWHM of 4.8".

Tim

 

I'm willing to believe that my FWHM is closer to the 4 or 5 range!

 

Do I have this totally wrong in my head:

 

- with a longer FL (2800mm) the FWHM of a star has to be larger than with a short FL (200mm) because the star is magnified more and therefore covers more pixels. Is that correct?

 

Generally not.  It's complex, but two things that figure in are the higher resolution of a large aperture scope, and the quality of the optics.  The "magnification of the star" doesn't really figure in to it, stars are (without optics) extremely tiny points of light that cannot be "magnified" even by a large telescope.  For example, Betelgeuse is 0.05 arc sec, about one-tenth of a pixel at your image scale. 

 

What you're seeing are things like imperfections (such as optical aberrations) and noise, not a magnified star.  This chart is fun, can't guarantee the accuracy, but the general picture is correct.

 

http://www.hyperiont...omparison16.php

 

And, of course, blur due to atmospheric turbulance, aka "seeing" trumps everything.  Rough idea (not intended to be precise).  If your FWHM is 4 to 5, it's likely that your seeing is 3 to 4.  Note the sig, please.

 

The problem with the chart is that it doesn't compare to scopes with correctors.  If the you are comparing to an RC with a corrector or an EdgeHD instead of a ACF the results are closer to each other.  

 

This is a nice comparison of corrected systems https://www.google.c...nKtH9hDyeGr1Mtg



#17 anismo

anismo

    Fly Me to the Moon

  • *****
  • Posts: 5540
  • Joined: 30 Jan 2014
  • Loc: Austin, TX, USA

Posted 22 April 2017 - 10:43 PM

With my Edge11 and ASI1600, (no reducer) I am at 0.28''/px and the best I can get it about 2.5 HFR (px) in very good seeing.

 

This has translated to FWHM of roughly about 5pxin PI subframeselector


Edited by anismo, 22 April 2017 - 10:44 PM.


#18 akulapanam

akulapanam

    Surveyor 1

  • *****
  • Posts: 1866
  • Joined: 26 Aug 2012

Posted 22 April 2017 - 11:53 PM

With my Edge11 and ASI1600, (no reducer) I am at 0.28''/px and the best I can get it about 2.5 HFR (px) in very good seeing.

 

This has translated to FWHM of roughly about 5pxin PI subframeselector

I'm really interested in seeing what the color version of your image ends up looking like.  I was very impressed with the mono version.



#19 akulapanam

akulapanam

    Surveyor 1

  • *****
  • Posts: 1866
  • Joined: 26 Aug 2012

Posted 22 April 2017 - 11:58 PM

I guess its also worth pointing out that with my RC at .92" per pixel (694 binned 2x) the best I could get was like 5 FWHM but with the Orion AG10 at about 1" per pixel I could get 3.7-4 FWHM last night.   Unfortunately, the AG10 is unusable at the moment except in the position you collimate (using a star) it because it has some massive mirror flop but I was impressed with the improvement.


Edited by akulapanam, 22 April 2017 - 11:58 PM.


#20 Jon Rista

Jon Rista

    Hubble

  • *****
  • Posts: 18089
  • Joined: 10 Jan 2014
  • Loc: Colorado

Posted 23 April 2017 - 12:43 AM

Hi,

 

First time trying to setup/use the C11 Edge and the best HFR I could achieve was 3.75. Is this to be expected at long FL or are my skies just craptacular? It seemed I got somewhat better HFR with the 100mm scope, so is it just a matter of the FL?

Measure both as FWHM in PixInsight. Then convert that FWHM into arcseconds using the image scale of each camera. My guess is that 3.75 HFR with your SCT is going to be better than whatever HFR you got with the 100mm refractor, because the image scale is going to be larger with the SCT. 



#21 GeneralT001

GeneralT001

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1625
  • Joined: 06 Feb 2012
  • Loc: Nanaimo, BC

Posted 23 April 2017 - 01:29 AM

With my Edge11 and ASI1600, (no reducer) I am at 0.28''/px and the best I can get it about 2.5 HFR (px) in very good seeing.

 

This has translated to FWHM of roughly about 5pxin PI subframeselector

Hey thanks, it sounds like my numbers are not to far out. I'm sure if I ever get some nice skies I'll be able to get a bit better numbers.



#22 GeneralT001

GeneralT001

    Surveyor 1

  • *****
  • topic starter
  • Posts: 1625
  • Joined: 06 Feb 2012
  • Loc: Nanaimo, BC

Posted 23 April 2017 - 01:35 AM

 

Hi,

 

First time trying to setup/use the C11 Edge and the best HFR I could achieve was 3.75. Is this to be expected at long FL or are my skies just craptacular? It seemed I got somewhat better HFR with the 100mm scope, so is it just a matter of the FL?

Measure both as FWHM in PixInsight. Then convert that FWHM into arcseconds using the image scale of each camera. My guess is that 3.75 HFR with your SCT is going to be better than whatever HFR you got with the 100mm refractor, because the image scale is going to be larger with the SCT. 

 

 

Hi Jon,

 

With the C11/ASI1600 combo and using the FWHMEccentricity Script I get a meanFWHM pixel size of just under 10. So if I multiply 10 x the .28 image scale I get a FWHM of about 2.8 arcsec for that sub? Is that right?




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.







Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics