#121  
Old 02-10-2017, 09:36 AM
peter_4059's Avatar
peter_4059 (Peter)
Big Scopes are Cool

peter_4059 is offline
 
Join Date: Jun 2007
Location: SE Tasmania
Posts: 4,532
Quote:
Originally Posted by Slawomir View Post
Thank you for sharing your experience Lee - your images certainly are great and inspire me and many others

Galaxy imaging is something I would like to do in the future, so this discussion is of great interest to me.

I put together a simple excel spreadsheet in an attempt to roughly estimate which system would be optimal for galaxy imaging with the seeing we have in coastal Australia.

It looks like my camera's small pixels are always limited by any smallish optics slower than f/6 until we get to 140 mm aperture and above, when seeing will become limiting factor. So for my current camera, either a 250 mm f/4ish Newton or 140 mm f/6ish refractor would do for galaxy imaging, with Newton being significantly faster but would require collimating.

Those two might be my options since I feel that in the future we will see more new sensors with smaller pixels that will most likely will be more affordable than sensors with large pixels.
Suavi,

Can you explain the rational for your spreadsheet? Are the numbers arc seconds per pixel? What are your criteria for optimal sampling and seeing limited?

I get 0.97 aspp with my QSI683 (5.4 micron pixels)/Paracorr (1.15 focal length multiplier)/SN10 (10" f4) combination.
Reply With Quote
  #122  
Old 02-10-2017, 10:28 AM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
Quote:
Originally Posted by Slawomir View Post
Shallower wells but also usually less read noise. However, in spite of having lower read noise, dynamic range is generally lower with smaller pixels, so more shorter exposures are needed if we want to control saturation of stars. The biggest advantage of small pixels on a small chip for me is substantially lower cost of the entire imaging apparatus.

As I understand, 2.4 micron pixels are in most cases best matched with fast telescopes (f/5 and faster).
I’d agree with a seeing-limited approach to pixel/scope choice, but with newer technology the smaller pixels aren’t always at a substantial advantage.

For example, the well depth of the IMX178 is about 15000 whereas the ICX694 is around 19000. Not only does the 178 have lower read noise, but with the use of BSI it has a reportedly higher QE too. With a 14-bit ADC the 178 isn’t at any significant disadvantage either.
Reply With Quote
  #123  
Old 02-10-2017, 10:43 AM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
Quote:
Originally Posted by peter_4059 View Post
That's what I suspected. I guess the trade off then is in image processing time?

Is it possible to get the same depth in an image with many short exposures vs fewer longer ones?
1) processing time - depends on the capabilities of your computer. Registration and Integration is multithreaded in PixInsight now, so a modest quad core and a solid state drive works wonders

2) depth - so long as your images are sky limited, stacking many short images should still amount to a similar numerical result. If a photon arrives from a faint part of an object every 10 minutes and you take 10 minute subs, over the course of an hour you collect 6 photons. With 1 minute subs you still collect 6 photons in the same time...
Reply With Quote
  #124  
Old 02-10-2017, 10:54 AM
Peter Ward's Avatar
Peter Ward
Galaxy hitchhiking guide

Peter Ward is offline
 
Join Date: Dec 2007
Location: The Shire
Posts: 8,090
Quote:
Originally Posted by Camelopardalis View Post
2) depth - so long as your images are sky limited, stacking many short images should still amount to a similar numerical result. If a photon arrives from a faint part of an object every 10 minutes and you take 10 minute subs, over the course of an hour you collect 6 photons. With 1 minute subs you still collect 6 photons in the same time...
Wouldn't many of the one minute exposures have no signal at all? Hence you collecting noise rather than signal.

Also, if the camera is downloading, that lone photon might be missed entirely....as opposed to the camera with its shutter being open most of the time.
Reply With Quote
  #125  
Old 02-10-2017, 11:15 AM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
Quote:
Originally Posted by Peter Ward View Post
Wouldn't many of the one minute exposures have no signal at all? Hence you collecting noise rather than signal.

Also, if the camera is downloading, that lone photon might be missed entirely....as opposed to the camera with its shutter being open most of the time.
1) Many of the exposures would indeed record no photon, only background sky. However, since the SNR increases with the number of subs, the statistics quickly discerns between noise and faint signal

2) Given the random nature of photon arrival, over the course of the 10 minutes the window of opportunity to miss the photon is longer with many shorter subs but over the course of the hour it is improbable (although not impossible) that the photon arrives while the camera is downloading on each occasion, and thus the stack still records a statistically significant photon
Reply With Quote
  #126  
Old 02-10-2017, 11:26 AM
Peter Ward's Avatar
Peter Ward
Galaxy hitchhiking guide

Peter Ward is offline
 
Join Date: Dec 2007
Location: The Shire
Posts: 8,090
Quote:
Originally Posted by Camelopardalis View Post
1) Many of the exposures would indeed record no photon, only background sky. However, since the SNR increases with the number of subs, the statistics quickly discerns between noise and faint signal

2) Given the random nature of photon arrival, over the course of the 10 minutes the window of opportunity to miss the photon is longer with many shorter subs but over the course of the hour it is improbable (although not impossible) that the photon arrives while the camera is downloading on each occasion, and thus the stack still records a statistically significant photon
So I'll play devil's advocate here. Draw this to an extreme...say millisecond exposures. Still think we'll be collecting signal..if any...as quickly as the open shutter case?
Reply With Quote
  #127  
Old 02-10-2017, 11:36 AM
Merlin66's Avatar
Merlin66 (Ken)
Registered User

Merlin66 is offline
 
Join Date: Oct 2005
Location: Junortoun Vic
Posts: 8,904
Peter,
The statistics of the "Shot noise" probability says yes... some of those very short exposures will record a photon.
Over an extended period (ie leaving the shutter open) the probability of collecting the same number of photons is the same.
Reply With Quote
  #128  
Old 02-10-2017, 11:45 AM
peter_4059's Avatar
peter_4059 (Peter)
Big Scopes are Cool

peter_4059 is offline
 
Join Date: Jun 2007
Location: SE Tasmania
Posts: 4,532
If the shutter is open for 1 min and the download time is 8 seconds then the percent of time capturing is 88%. If the shutter is open for 10 mins and the download time is still 8 seconds then the percent of time capturing increases to 99%. I'd have thought the probability of collecting the elusive photon must be higher if the shutter is open a larger percentage of the time?
Reply With Quote
  #129  
Old 02-10-2017, 11:50 AM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
Quote:
Originally Posted by Peter Ward View Post
So I'll play devil's advocate here. Draw this to an extreme...say millisecond exposures. Still think we'll be collecting signal..if any...as quickly as the open shutter case?
Actually this is where it gets a little more interesting...the “traditional” long exposure paradigm = software gets in the way here, as it doesn’t make any sense to take millisecond exposures with conventional software, working through an ASCOM driver which takes an exposure, downloads, then gives the OK signal for the software to initiate the next sub.

These devilish CMOS sensors come equipped with video capabilities, since many of them were designed for use in security cameras and driver aids in cars requiring a constant video stream.

So using the appropriate tool such as SharpCap you capture the entire video stream, no shutter involved...
Reply With Quote
  #130  
Old 02-10-2017, 11:56 AM
Peter Ward's Avatar
Peter Ward
Galaxy hitchhiking guide

Peter Ward is offline
 
Join Date: Dec 2007
Location: The Shire
Posts: 8,090
Quote:
Originally Posted by Merlin66 View Post
Peter,
The statistics of the "Shot noise" probability says yes... some of those very short exposures will record a photon.
Over an extended period (ie leaving the shutter open) the probability of collecting the same number of photons is the same.
Yep, I get that bit...it's all the other camera noise sources going to zero (thermal, read, bulk layer , etc.etc. ) and not swamping the signal I am having trouble getting my head around.
Reply With Quote
  #131  
Old 02-10-2017, 11:56 AM
Camelopardalis's Avatar
Camelopardalis (Dunk)
Drifting from the pole

Camelopardalis is offline
 
Join Date: Feb 2013
Location: Brisbane
Posts: 5,425
Quote:
Originally Posted by peter_4059 View Post
If the shutter is open for 1 min and the download time is 8 seconds then the percent of time capturing is 88%. If the shutter is open for 10 mins and the download time is still 8 seconds then the percent of time capturing increases to 99%. I'd have thought the probability of collecting the elusive photon must be higher if the shutter is open a larger percentage of the time?
Correct, but as the nature of the photon arrival is random, on average you’re only disadvantaged 1/8 of the time...

But also see shutterless software use above. More CMOS chips are rolling shutter than global shutter, so unless the photon strikes the pixel at the instant it is being read, then there is no loss taking place.
Reply With Quote
  #132  
Old 02-10-2017, 12:03 PM
Atmos's Avatar
Atmos (Colin)
Ultimate Noob

Atmos is offline
 
Join Date: Aug 2011
Location: Melbourne
Posts: 6,980
I’m afraid everyone seems to be missing the point, we’re discussing integration time, not time under the stars. One hour worth of subs is going to capture one hours worth of photons. Whether that hour is taken in single 10 minute subs over five nights or a single hour in one night makes no statistical difference as photon capture is a random event.
Reply With Quote
  #133  
Old 02-10-2017, 12:10 PM
Slawomir's Avatar
Slawomir (Suavi)
Registered User

Slawomir is offline
 
Join Date: Sep 2014
Location: North Queensland
Posts: 3,240
Quote:
Originally Posted by peter_4059 View Post
Suavi,

Can you explain the rational for your spreadsheet? Are the numbers arc seconds per pixel? What are your criteria for optimal sampling and seeing limited?

I get 0.97 aspp with my QSI683 (5.4 micron pixels)/Paracorr (1.15 focal length multiplier)/SN10 (10" f4) combination.
Yes, arcseconds per pixel.

Criteria? Pretty much got them out of thin air , but yes, I believe, perhaps naively, that sampling for galaxy imaging between 1" pp and above 0.6" pp should in a costal location yield comparable and optimally sampled data, as long as aperture will not limit the resolution and guiding will be spot on. Going below 0.6-0.7"pp for deep sky imaging in a typical coastal location in AU is a bit pointless, I presume
Reply With Quote
  #134  
Old 02-10-2017, 12:14 PM
peter_4059's Avatar
peter_4059 (Peter)
Big Scopes are Cool

peter_4059 is offline
 
Join Date: Jun 2007
Location: SE Tasmania
Posts: 4,532
Quote:
Originally Posted by Atmos View Post
I’m afraid everyone seems to be missing the point, we’re discussing integration time, not time under the stars. One hour worth of subs is going to capture one hours worth of photons. Whether that hour is taken in single 10 minute subs over five nights or a single hour in one night makes no statistical difference as photon capture is a random event.
I'm discussing the pros and cons of small pixels and long vs short subs...not sure what you are discussing?
Reply With Quote
  #135  
Old 02-10-2017, 12:32 PM
Atmos's Avatar
Atmos (Colin)
Ultimate Noob

Atmos is offline
 
Join Date: Aug 2011
Location: Melbourne
Posts: 6,980
Quote:
Originally Posted by peter_4059 View Post
I'm discussing the pros and cons of small pixels and long vs short subs...not sure what you are discussing?
Below

Quote:
Originally Posted by peter_4059 View Post
If the shutter is open for 1 min and the download time is 8 seconds then the percent of time capturing is 88%. If the shutter is open for 10 mins and the download time is still 8 seconds then the percent of time capturing increases to 99%. I'd have thought the probability of collecting the elusive photon must be higher if the shutter is open a larger percentage of the time?
Reply With Quote
  #136  
Old 04-10-2017, 06:22 PM
Shiraz's Avatar
Shiraz (Ray)
Registered User

Shiraz is offline
 
Join Date: Apr 2010
Location: ardrossan south australia
Posts: 4,918
Hi again Andy.

apart from resolution, sensitivity also varies with aperture. The attached graphs show how resolution and relative sensitivity vary with aperture for practical systems around 0.75arcsec sampling.

resolution flattens out pretty well by 200mm (within a few % of best possible), but the sensitivity increases with aperture. ie, set your focal length to 1.5m to match your pixels and then get the fastest scope that your mount will carry.

Although small scopes with small pixels will get similarly resolved results to larger scopes, they will take a lot longer to get there. Chips with small pixels do not make large scopes obsolete - they do however bring high resolution imaging within the reach of small scope/mount owners.

cheers Ray
Attached Thumbnails
Click for full-size image (size.jpg)
108.6 KB50 views
Reply With Quote
  #137  
Old 04-10-2017, 07:43 PM
Andy01's Avatar
Andy01 (Andy)
My God it's full of stars

Andy01 is offline
 
Join Date: Mar 2012
Location: Melbourne
Posts: 3,253
Quote:
Originally Posted by codemonkey View Post
IMO you can get some decent results using a moderate sized refractor with small pixels... it's certainly been my approach and, not to blow my own trumpet, but I think my images hold up well to much larger, much harder to manage and potentially much more expensive scopes
Yes your images are inspirational Lee, amazing results

Quote:
Originally Posted by strongmanmike View Post
I agree, your images are great and I certainly don't yawn when looking at them....
+1^

Quote:
Originally Posted by DJT View Post
I have been chasing ARP14, a tiny little galaxy, from Sydney with an RC8 (1628mm) and an older ST2000XM with a 7.2 pixel size. Its doable but needs patience.
But this would be a struggle if the plan is to do single nights at a dark site... it might be a struggle, unless you have your setup process nailed completely....these things can hold focus and collimation all night so that's a plus.
Yes, good arguments for the CF RC8

Quote:
Originally Posted by Shiraz View Post
... set your focal length to 1.5m to match your pixels and then get the fastest scope that your mount will carry.
Thanks again Ray - as mentioned above I'm leaning towards a carbon fibre RC8 to suit my EQ6/5.4m pixels combo.
1625mmFL, 7Kgs, short OTA (less wind issues than a newt at my exposed location), good bang for the buck but will need collimation & upgraded focusser.

Quote:
Originally Posted by Shiraz View Post
Given that you might want something with >100 mm clear aperture (~200mm+ if obstructed) and 1.5m fl, that will work on an EQ6 - you could consider:
- an 8 inch f8 RC with field flattener (will take a long time, but give best resolution)
Agreed, a bit slow at f8 but the short OTA, longer FL & best resolution are attractive.
Maybe more so than a shorter FL, longer OTA 8" f4 CF Newt, and anything 10" is too heavy, as is the MN190 MakNewt.

I'd like to one day get decent A4 size printable images to enter in print competitions, so this is looking like a good entry point to start working towards that goal.

Thanks again to all for your input, glad several other IIS'ers have got something out of this thread as well.

Last edited by Andy01; 04-10-2017 at 07:47 PM. Reason: typo
Reply With Quote
  #138  
Old 05-10-2017, 09:36 AM
rally
Registered User

rally is offline
 
Join Date: Sep 2007
Location: Australia
Posts: 896
The OP's question was about the minimum focal length required and the discussion has wandered around this and its been interesting, but I cant help but think that the FL he needs is more likely constrained by the FL that the rest of his system can support.

Anyone who has imaged at long FL has discovered how much more exponentially worse all the problems become.

Things that werent even considerations become significant problems and obstacles.

Talking about 1 arc sec seeing or sub arc sec image scale becomes a moot point if the rest of the noise in your system is being measured in many 10's of arc secs or even arc minutes !

eg your mount has more Periodic error than a few arc secs and there is no PEC, if mount backlash exists and is a problem, if mount capacity is at its limit, mount's ability to hold stability with light breeze etc, if polar alignment is not perfect - eg under say 10 arc secs, if tracking is limited or if there is no tracking and pointing correction for example, if your guiding errors are more than an arc sec, if you can't hold focus across an imaging session, if guiding flexure or mirror flop is occurring . . . its a long list.

You simply cannot capture the theoretical image detail in your subs if the mechanical noise in your system exceeds the telecope spot size, image scale or the seeing.
If that noise greatly exceeds the theoretical limits then its a failed exercise.

So IMO the answer needs to at least consider as much about what the OP's existing system is capable of yielding, since the budget is only $2000 it doesnt allow for any other upgrades other than OTA.

So for example
What is the existing PE and backlash ?
What is a typical PA for this system and the methods used
What is the tracking and guiding providing now at low FL and how will that be affected by increasing FL etc.
The choice of camera and sub length is relevant too - if the minimum sky noise and read noise limited sub length is sufficiently long but the mounts ability to track accurately or remain focussed for that length of time is compromised then its going to be difficult to obtain good results.

Imaging at longer FL is a whole new ball game, the status quo is not the same as imaging at a forgiving 770mm on a small light weight scope at an image scale of around 1.5 as/p

Its all doable and encouragement given, but knowing what you are up against is an important consideration.
I've seen EQ6's (and I see this one is belt modded) that had well over 1 arc minute of mechanical error - backlash and PE
Even after hypertuning, relubing, swarf removal and bearing upgrades the errors were still in the 10's of arc secs.

My 2c worth

Cheers
Rally
Reply With Quote
  #139  
Old 05-10-2017, 09:45 AM
Merlin66's Avatar
Merlin66 (Ken)
Registered User

Merlin66 is offline
 
Join Date: Oct 2005
Location: Junortoun Vic
Posts: 8,904
Rally,
Back, as you say to the original question....
I'm glad to see that seeing conditions are mentioned.
I semi-regularly (weather limited) use the C11 on the NEQ6pro at f10 (2800mm fl) with a reflective slit plate guiding (similar to an on axis OAG) and Lodestar.
My target is to keep the FWHM star image central on a 20 micron slit for at least subs of 10 mins. At "good" seeing of 2 arc sec the FWHM is 27 micron.
Using PHD2 (or the AA6 guiding) this seems doable.
Just my 2c
Reply With Quote
  #140  
Old 05-10-2017, 08:19 PM
gregbradley's Avatar
gregbradley
Registered User

gregbradley is offline
 
Join Date: Feb 2006
Location: Sydney
Posts: 17,877
Quote:
Originally Posted by Atmos View Post
The size of the CFZ is determined by the focal radio as it is the steepness of the light curve. The size of the pixels also has to be taken into account as this is the area in which the CFZ occupies.

When you have a larger slower telescope you typically have larger pixels and at higher resolution you are never diffraction limited. Take what Mike and Trish have, a 20” CDK with a 16803. When the seeing is less than perfect the CFZ aid basically made larger via atmospheric blurring.

4-5” F/5 refractors really struggle because their CFZ is the determining factor of perfect focus where as a CDK can largely be determined by the seeing conditions.

A large CDK scope can be very difficult to focus in poor seeing. So much so that its really a waste of time trying to image with that sort of long focal length scope in poor seeing.

I have had it where it seems in focus you take another focus shot and the seeing pushes it out of focus, you adjust and all you are doing is chasing the seeing and its all over the place and unpredictable.

Temperature compensation is a good tool but its somewhat imperfect because it assumes a straight line response between temperature and focus and I don't know that it is. Its not a bad approximation but I imagine its not as accurate as continuous autofocus or refocusing at some regular interval which matches your scopes focus shift with temperature.

For example I have noticed my Honders will shift focus noticeably after an hour of thermal equalizing and then the differences become much less after its settled.

The carbon fibre scopes so far have been very forgiving of temperature focus shift.

Some scopes have built in adjusters to match the focus shift. Roland Christen has said aluminium tubes focus shift matches reasonably closely the shift in the lens focal point so it tends to cancel out.

Greg.
Reply With Quote
Reply

Bookmarks

Thread Tools
Rate This Thread
Rate This Thread:

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +10. The time is now 08:24 PM.

Powered by vBulletin Version 3.8.7 | Copyright ©2000 - 2024, Jelsoft Enterprises Ltd.
Advertisement
Testar
Advertisement
Bintel
Advertisement