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

Guidestars: color, brightness, and number

  • Please log in to reply
127 replies to this topic

#1 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • topic starter
  • Posts: 8739
  • Joined: 12 Oct 2007

Posted 26 October 2015 - 07:09 AM

There was recently a discussion about the color of a typical guidestar - and I think it's an interesting topic that is worth looking into. This is somewhat technical - but it pertains to an issue that is important to anyone who has trouble finding guidestars, or maximizing their signals for good guiding.

Nowadays star catalogs are available to answer these questions directly with some code. There are many studies of star distributions - but they tend to focus on intrinsic physical properties of stars - rather than apparent properties as seen from the Earth. But the apparent properties are what matter for finding guidestars - since a close faint star is just as bright as a distant bright one.

So here is a basic view of the issue - a simple view of the distribution of guidestars in the entire sky. Since the stars tend to be denser near the milky way - this is shown in galactic coordinates, centered on the center of the galaxy.

This shows all stars in the UCAC4 catalog brighter than VMag=10. Many people guide on fainter stars - but 10 seems like a reasonable cutoff for desirably brightish guidestars.

It should be no surprise that stars are most likely found near the center - but there is a trade off because there is also a lot of dust in the plane of the milky way - so you see big patches of relatively few brightish guidestars - even though they are 'there.' They are just darkened by dust.

But as you move away from the milky way, they definitely reduce in numbers.

Frank

Attached Thumbnails

  • StarDensityMag10GalacticLoRes.png

  • korborh likes this

#2 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • topic starter
  • Posts: 8739
  • Joined: 12 Oct 2007

Posted 26 October 2015 - 07:19 AM

Now the question is - what color are they? I took all the UCAC4 stars with good B and V magnitude values, and plotted histograms for those brighter than mag 8, mag 9, and mag 10 - vs. their B-V color index. The B-V corresponds to the color temperature of the star - with blue stars slightly negative, red stars positive, and the sun in the middle around 0.6.

The result shows clearly that more than half of the mag 10 stars are actually bluer than the sun - and only a fraction are particularly cool and red. This is because such red stars are also fainter - and you only see them if they are close by.

In addition, it shows that as you increase in depth by one magnitude, the number of guidestars increases by a factor of 4. This is consistent with simple volumetric geometry, because if a star is 4 times fainter, it is 2 times farther away. And if it is 2x farther away, it is part of a volume that is 8x bigger. So if you see stars fainter by a factor of K, then you are seeing a volume of stars bigger by K^1.5. Thus a step of one magnitude is a factor of 2.512 in intensity, or 2.512^1.5=3.98 in volume, or number of stars. And that is shown directly in the histograms. You get about a factor of 4x more stars with each added magnitude that you can reach.

The net conclusion is that on average guidestars are about the color of the sun - and if you want to find them easily - you should go as faint as possible.

Frank

Attached Thumbnails

  • StarColorHistogram.png

  • korborh likes this

#3 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • topic starter
  • Posts: 8739
  • Joined: 12 Oct 2007

Posted 26 October 2015 - 07:24 AM

Now when the starlight enters the telescope, there will be losses across the spectrum due to mirrors and lenses - and then the detector itself will have varying sensitivity across the spectrum. But both the optics and the sensor sensitivity tends to be peaked near the visible - and it drops off toward the UV and IR.

This is shown in the following plot - for three different sensor types, and for Celestron XLT system transmission. Not many telescopes show the full transmission of the system - but this curve is published. Other OTA's will have slightly different values - but again it would be peaked near the visible and drop off either side.

I also show typical RGB filters - in this case Astrodon E-type, which have a gap between green and red. Note that this leaves a big chunk of the spectrum completely blocked off.

Attached Thumbnails

  • FiltersStarsSensors.png

  • korborh likes this

#4 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • topic starter
  • Posts: 8739
  • Joined: 12 Oct 2007

Posted 26 October 2015 - 07:30 AM

Finally, here is a plot of the net flux, in electrons, for a fairly sensitive QHY5l-ii guide camera. It is cmos and slightly less sensitive than the Sony 694 - but the both drop off toward the IR. If you multiply the losses due to reflection and QE for a typical guidestar with temperature 5500K, you get a big reduction as you move to the IR, while the flux is very strong over the visible where the star emission is strong.

If you only used the signal from wavelengths above 750nm, you would reduce the electron signal by a factor of about 9 from purely unfiltered. And this would have a reduction in visible guidestars by a factor of nearly 30.

Attached Thumbnails

  • TotalFluxQHY.png

  • korborh likes this

#5 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • topic starter
  • Posts: 8739
  • Joined: 12 Oct 2007

Posted 26 October 2015 - 07:35 AM

This should show the importance of signal from visible light in maximizing the odds of seeing a guidestar in a given view. Not only will filters reduce the signal, they will strongly reduce the number of guidestars that can be seen.

Focus quality and aberrations will also spread the light of the star and reduce its SNR - having the same effect of reducing the limiting magnitude, and reducing the odds of finding one.

None of this will matter too much if you are using a guidescope - or very long guide exposures. But if you do want to find guidestars easily and quickly - and guide on them with strong signal - remember that they are most likely similar to our sun in color - and would benefit from signal in the visible.

Frank
  • korborh likes this

#6 Hilmi

Hilmi

    Vendor-Astronomical Solutions Co.

  • *****
  • Vendors
  • Posts: 5628
  • Joined: 07 Mar 2010
  • Loc: Muscat, Sultanate of Oman

Posted 26 October 2015 - 09:04 AM

I understand that you had brought up this statistic in order to methodically prove your point about the value of an ONAG. I applaud your methodical and scientific approach to the issue, but would like you to add one more factor. From my understanding, the ONAG is targeted at cameras with larger sensors. Can you now take into account the area of sky visible with a large sensor camera vs a tiny chip through pick-off prism?

 

I would assume that you really don't get the benefit of an ONAG unless you have a guide camera with a large sensor such a Atik 314L+ or that new Kodak APS C sensor that operates without a shutter. Can't remember it's name.



#7 jhayes_tucson

jhayes_tucson

    Fly Me to the Moon

  • *****
  • Posts: 7090
  • Joined: 26 Aug 2012
  • Loc: Bend, OR

Posted 26 October 2015 - 10:06 AM

Frank,

I don't really understand your agenda here.  You've set out to prove something that is fairly obvious.  Anytime you introduce a spectral filter, the total signal goes down.  If you filter out blue light, you'll detect fewer blue stars and if you filter out red light, you'll see fewer red stars.  Introducing the stellar population into the issue is an interesting way to try to hand waive your way into some kind of conclusion about how many stars can be found in the NIR; though, I agree completely with Hilmi's point that you are missing an important parameter if you don't consider sky coverage.  

 

There is absolutely nothing wrong with guiding in the visible or in the NIR as long as you have a suitable signal.  The question isn't how many more stars can you detect in the visible compared to the NIR?  The question is can you find a sufficient number of guide stars in the NIR to achieve a good guide signal?  The reality is that there is no problem finding guide stars in the NIR--even with a fairly small sensor.  For over a year, I have used a Lodestar X2 in the NIR and I have never once had any problem finding a suitable star to guide on.  Honestly, I have a lot more trouble finding guide stars in the visible with my guide-scope but that's only because I use a NexGuide guider that has a pretty insensitive camera and a really crummy display.  

 

The take away is that if you want to get a good signal, make sure that you use a good camera that has good responsivity over the spectral region of interest.  That's true for imaging cameras and guide cameras alike.

 

John



#8 Alex McConahay

Alex McConahay

    Cosmos

  • *****
  • Posts: 8072
  • Joined: 11 Aug 2008
  • Loc: Moreno Valley, CA

Posted 26 October 2015 - 11:20 AM

You know, Freestar8n,  if your point is that ONAG cannot work well, there is an easier way to prove it than putting out a bunch of graphs. Just put the ONAG in an imaging train and try it out. 

 

Now, I'm not particularly excited about spending $1400 plus maybe on that experiment. But others have tried it, and even more will be trying it in the near future. Let's let those with the money decide if this is a good product or not. 

 

So far, we have had testimony from John that he has "never once had any problem finding a suitable star to guide on." And he is using a Lodestar 2, a rather common guide camera (albeit a good one).

 

In the thread that got locked, I heard that several other people had tried it, and most had found success. One had a buddy who had not. (If I recall correctly.) My only question in that thread was why it was not a more common product if it were in fact so powerful. My issue of the number of users appears to be related to cost and backfocus rather than performance (as far as I can tell).  As I said, several people reported good success. This answers my questions about why not many are using it. It appears I am wrong. People are using it. Successfully.  

 

Why don't we just let the results from the field roll in a little before we get too far off into the theoretical about what can and cannot work. It's not my money or yours, but that of a number of people who seem to be willing to try it out. 

 

Why not call a truce on this for a bit while we get a few more of these in the field, and see how they work.

 

Alex



#9 dawziecat

dawziecat

    Gemini

  • *****
  • Posts: 3020
  • Joined: 20 Oct 2010
  • Loc: Rural Nova Scotia

Posted 26 October 2015 - 12:34 PM

Actually, I am surprised a bit by Frank's conclusions. I expected that, by far and away, MOST guidestars would be O, B, A and F stars. The idea that cooler stars are far more numerous in an absolute sense is indeed a red herring of a sort because they are so intrinsically faint.

 

I have made an act of faith. I have already ordered the adapters I will need for the ONAG-XM and will order the guider itself as soon as a payment settles into my account later this week.

 

As for Hilmi's observation that the ONAG seems targeted at those with larger format guide cams, that excludes me. That is, as I see it, a concern for remote imagers. The movable X-Y stage will allow we "non-remote" imagers access to the entire FOV of the imaging OTA to find a guide star. And without an OAG prism shadow to "flat out." I will be using the ONAG with my Lodestar X2. No interest in buying another "real CCD cam" just to use as a guider.



#10 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • topic starter
  • Posts: 8739
  • Joined: 12 Oct 2007

Posted 26 October 2015 - 03:18 PM

Actually, I am surprised a bit by Frank's conclusions. I expected that, by far and away, MOST guidestars would be O, B, A and F stars. The idea that cooler stars are far more numerous in an absolute sense is indeed a red herring of a sort because they are so intrinsically faint.


Hi-

Yes - in a way it is surprising to me that there is such a long tail into the red. It isn't surprising that most guidestars are bluer than the sun - but it is somewhat surprising that so many are out to the red. But when you take those and multiply by transmission and sensor response - there isn't much signal.

Part of the issue is that many of the stars are toward the milky way and reddened by dust - so in fact they are redder than they should be based on assumptions of color, distance, and luminosity. In other words - they are redder than their true color temperature. But it's not enough to make a big difference in terms of actual signal.

There is obvious structure in the histogram - with three humps - and my guess is that it is related to local structure in the spiral arms. The bump far on the right is way redder than it should be - which suggests there is a disproportionate population of stars close to us. I checked the approximate location of the three humps - and they are generally distributed in the milky way and not, for example, in the Magellanic clouds. I'm sure someone has looked at this and published it - but I haven't found it. The stuff I am presenting here is something I worked on some time ago - and I just wrote it up since there seemed to be general interest - and guidestar signal is very important to autoguiding.

Even years ago somehow there was a general notion that sony sensors had extended IR response - and therefore you would get better signal from a guidestar if it was red - and in fact that is true - if you are using normal OAG unfiltered. The main effect is that two guidestars with the same VMag may have different signal due to their color and how their spectrum overlaps with the system response. So it is something I was aware of - but these histograms show that there should be no expectation of guidestars being particularly red.

I hear people talking about challenges of finding stars with OAG guiding - and a key point here isn't just the color issue - but the strong dependence of guidestar density on limiting magnitude. A filter is just one way of limiting magnitude - but others are defocus, aberration, and any blockage of the pupil as seen from the sensor. These factors may be obvious - but looking at OAG setups and hearing people say that aberrations don't affect centroiding - combined with complaints that OAG is hard to use - make me suspect that people don't realize just how important these factors are in finding guidestars and guiding on them well.

Frank

#11 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • topic starter
  • Posts: 8739
  • Joined: 12 Oct 2007

Posted 26 October 2015 - 03:36 PM

As for people using different guiding methodologies with no problem - keep in mind that some people have no problem finding guidestars with 3 second exposures - while others have no problem finding them with 0.1s exposures. If you allow exposure time to be a free parameter - then any system will find something to guide on eventually. And the difficulty for all methods will get worse as you stray from the milky way.

Star clusters and nebulosity - including planetaries - will tend to be near the milky way and have a higher density of guidestars - which will benefit any guiding method. But distant galaxies have no association with guidestar density, and galaxies far from the milky way will have a lower probability of finding a guidestar by chance on the guider. There will always be one in there of some magnitude - but there may be only one in the general area that is particularly bright - and it can be found and guided on systematically if you know where it is and can center it on the guider.

A real killer for any method that guides on stars away from the optic axis is coma - which used to be common in sct's. But now there are many imaging options that are well corrected for coma and even astigmatism - so that someone revisiting off-axis methods will find many more guidestars than before - because of the huge dependence of aberrations on limiting magnitude, and the dependence of limiting magnitude on guidestar density - along with amazing and very recent improvements in camera sensitivity across the visible spectrum.

Frank

#12 korborh

korborh

    Apollo

  • *****
  • Posts: 1125
  • Joined: 29 Jan 2011
  • Loc: Oregon

Posted 26 October 2015 - 05:31 PM

Actually, I am surprised a bit by Frank's conclusions. I expected that, by far and away, MOST guidestars would be O, B, A and F stars. The idea that cooler stars are far more numerous in an absolute sense is indeed a red herring of a sort because they are so intrinsically faint.

 

Yes, this was a really good point by Frank challenging misinformation in the ONAG advertising on the CN thread.


Edited by korborh, 26 October 2015 - 05:32 PM.


#13 vpcirc

vpcirc

    Fly Me to the Moon

  • ****-
  • Posts: 5169
  • Joined: 09 Dec 2009

Posted 26 October 2015 - 06:03 PM

As soon as we have everything up and running I'll be giving a live demo to the astro imaging channel and will also post real time video. I can save fits files as well so people can measure for themselves. I'll make sure I plate solve so the exact star data is known. People can draw their own conclusions. I have no clue how well this is going to work, and I'm keeping an open mind. If it fails there's a 30 day return policy. If it works as advertised, it's worth every cent to me for the nightmares it saves from long focal length rotator guiding. We see no need to worry about StarLock as the optics in an RCOS allow the focus remain virtually unchanged throughout the night. 

 

Look for my post of results, in the meantime, I'm not wasting any more of my time on this topic!



#14 jhayes_tucson

jhayes_tucson

    Fly Me to the Moon

  • *****
  • Posts: 7090
  • Joined: 26 Aug 2012
  • Loc: Bend, OR

Posted 26 October 2015 - 07:39 PM

 

 

Yes, this was a really good point by Frank challenging misinformation in the ONAG advertising on the CN thread.

 

 

 

What's up with this statement?   I saw a technical discussion but I didn't see an ad containing misinformation.  Maybe I missed something.  What was the "misinformation" and where was the ad?

 

John



#15 Corsica

Corsica

    Vendor (Innovationsforesight)

  • *****
  • Vendors
  • Posts: 633
  • Joined: 01 Mar 2010
  • Loc: PA - USA

Posted 26 October 2015 - 10:46 PM

I do not remember having written anywhere anything wrong about the ONAG and NIR, but if I did please let me know.

 

The last thing I would do is to misinform anybody (yet I understand that no being a native English speaker may have some side effects :-)) the ONAG exists today because I designed and made it to provide another option than guide scopes and OAGs, solving some of their own limitations.
Since I and other amateurs were please with the ONAG performance I eventually decided to make it available to more amateurs like me, and for me the only realistic way to do it was to sell it.

I do not have any other agenda and if I wrote a mistake, we all make some, I'll be more than happy to correct it (them).
I think our eduction page on guiding in NIR discloses the drop in NIR with star surface temperature too:

 

http://www.innovatio...r-infrared-nir/

 

My point is, and has been, that there is many stars with a lot of NIR, and I would argue that even hotter stars still have enough amount of NIR for guiding with today's sensor technology (like the lodestar).
I do not dispute Frank's analysis and I always said and written that you see a drop in signal level when using NIR too, every systems have limits, OAG, and ONAG as well.

I work with lot of optical sensors for other applications and I knew that unfiltered silicon sensors are quite sensitive in NIR (up to 1000nm) and that the black body radiation has a heavy trail in NIR (https://en.wikipedia...wiki/Black_body).

Now as mentioned before using NIR for guiding decrease quite a bit the actual star profile for long term exposures (seconds) otherwise inflated by the seeing, this is also true for the HFD.
Although the guide star shape experiences some astigmatism with an ONAG, which is used for auto-focus (http://www.innovatio...ocus-focuslock/), it remains close to the actual HFD (not FWHM), which I would consider as a important parameter for guiding using centroid, and also in terms of final SNR.

At the end of the day SNR is what does matter, the signal is only one piece of the ratio, function of exposure time, f/#, star profile, star temperature, size of the guider FOV, ...

 

For completeness I should say that we have ONAG applications where customers use or custom/optional astigmatism corrector for the guider port if needed, this provide near diffraction limited image now, but at the expense of not be able to use FocusLock. As a matter of act our early ONAG XT used to feature this corrector in standard, but we eventually dropped it since it was not necessary for just guiding, ONAG owners were successful without it.

 

By the way with large imaging chips you could face challenging guide stars using on OAG with some optics when way off axis too, not to mention diffraction.

If the goal here is demonstrated that there is a drop in signal using the NIR I fully agree, please take note of this, if the goal on the other hand is to assume that the ONAG cannot work well, or at all, in the real applications and therefore I used misinformation to sell the product I would have to strongly disagree.

Having say that I do understand this is complex topic and some arguments can be lost, challenge, or taken out of context here, or even misunderstood but I just cannot accept the idea that I promote (with passion) I product I made years to bring to this level using trickery, this is just not true nor fair.

 

I am sure you can find people who did not have a good experience with an ONAG, the same is for sure true with an OAG and I am one of them (and I know how to use an OAG too).
We had few returns (by the way Mike we have a 60 days return policy :-)) but as far as I can remember once only a customer argued it could see any guide star. When this happens I usually suggest to support the customer, in a latter case it was a defective lodestar, but in this case he did not want any help.

You do not know if this is the real problem or just a good reason for returning a product, you just do not ask, and  you take the product back.

 

However there are a lot of happy users who do not see the NIR as big issue using an ONAG, very often they would tell you that advantages offsets this, the very early review from S&T reports the same thing, and that was long before the new lodestar X2 which is twice more sensitivity in NIR (the best is yet to come on this matter):

 

http://www.skyandtel...ploads/ONAG.pdf

 

At the end of the game is you ask whether the ONAG and related real time auto-focus technologies solve all problems, my answer would obviously be no (backfocus is certainly a limitation, but NIR not so much), but it does provide new opportunities. As any new technologies it is more expensive early (so good ONAGs), this is just the usual path. I stand behind my point that the ONAG does work, and in my option quite well (a bad thing to say from a vendor -:)).

 

What I have difficulty to understand is arguments from some people who have never seen nor used an ONAG, and I wonder why they would not believe any (good) user feedback or even try and see the actual advantages and limitations, this just fits my vision of science.

 

As you known I am registered as a vendor in here, and for those who read my previous posts in different threads they know that I always try to stay fair in my comment/reply, to comply with CN's policy hopefully readers and the moderator will agree that reply was motivated.



#16 Corsica

Corsica

    Vendor (Innovationsforesight)

  • *****
  • Vendors
  • Posts: 633
  • Joined: 01 Mar 2010
  • Loc: PA - USA

Posted 26 October 2015 - 11:42 PM

One point I'll like to bring in the discussion, which is not specific to the ONAG, is guider pixel size and binning (yet another important parameter to play with).

 

Binning improves SNR in various way depending whether you are read noise limited (short exposures, few seconds) or photon noise limited (longer exposures, few minutes).
When autoguiding using short exposures (few seconds) we are usually read noise limited.
In this configuration the SNR gain with binning is more dramatic than for longer exposures (minutes) where the photon noise is dominant.
For short exposures (in this context of binning not seeing) a bin 2x2 can increase the SNR by a factor up to 3x, a bit more usually than the expected 2x, this is because in this context we are mainly read noise limited.
Now longer exposures eventually settle, for a 2x2 binning, to a SNR improvement around 2x since we become photon noise limited, this is a bit less than a star magnitude.
With focal lengths above one meter you can easily bin 2x2, or more, without any significant drop in guide star centroid calculation accuracy with most guiders (I should credit Alan Holmes from SBIG for championing low spatial resolution for autoguiding achieving sub pixel accuracy with just a 100mm focal length lens and a ST-i).
The SNR improvement largely offsets the spatial resolution difference for dim stars with long focal length scopes most of the time. This provides the opportunity to find more suitable guide stars (the limiting magnitude is increased by near one).
Now longer guiding exposures is an other way to improve SNR, a 2.5 seconds versus a one second exposure increases the guide star magnitude by one too. As I mentioned before guiding is a complex problem with various aspects when you consider it from a global point of view.

 

I just thought this could be informative in the general guiding framework, regardless with device you may use.



#17 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • topic starter
  • Posts: 8739
  • Joined: 12 Oct 2007

Posted 27 October 2015 - 03:33 AM

This thread is about the statistics of guidestar properties - and how they relate to their detection with various sensors and filters. At no point have I ever said that you cannot guide on typical stars with IR. In fact - from the plots above - you could also guide on them with a UV filter - but in both cases you would have greatly reduced signal, and fewer possible guidestars.

A key point is that the statistics of guidestar colors are based on apparent stellar properties rather than physical ones. Histograms are commonly used in astronomy - but for stars they usually show physical properties - and I don't know of a study similar to what I show above.

I hear people talking as if this stuff is obvious - and that's great if everyone is in agreement. But it often happens in CN and amateur astro in general that strange misconceptions take hold and spread like wildfire. I am reminded of the Jonathan Swift quotation, "Falsehood flies, and truth comes limping after it..."

So if anyone said you couldn't guide with an IR filter - they would be crazy. Of course you can. You can guide with pretty much any filter except an actual lens cap - as long as you expose long enough. You can also do well without guiding at all - with certain mounts.

If anyone does think apparent properties of stars are directly linked to their physical properties - that would certainly be an error and cause confusion - both about astronomy - and the ability to find guidestars. It doesn't matter what fraction of stars in the galaxy are deep red if they are so faint you can't even see them.

The point of my study is that guidestars are fairly normal stars, color-wise - but overall slightly bluer than the sun - as a direct consequence of the color-luminosity relation captured in the H-R diagram. With typical sensors and optics, a rough estimate for the loss of signal using an IR filter is about a factor of 9, which reduces the number of guidestars in a patch of sky by about a factor of 29. As long as no one sees a problem with these numbers - we can agree they capture a real disadvantage of guiding with an IR filter - and the potential advantages of guiding in that manner are a separate topic.

Frank

#18 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • topic starter
  • Posts: 8739
  • Joined: 12 Oct 2007

Posted 27 October 2015 - 03:42 AM

When autoguiding using short exposures (few seconds) we are usually read noise limited.


Indeed you will be - if you block almost all the signal before it arrives on your sensor. But that is not "we."

Frank

#19 Corsica

Corsica

    Vendor (Innovationsforesight)

  • *****
  • Vendors
  • Posts: 633
  • Joined: 01 Mar 2010
  • Loc: PA - USA

Posted 27 October 2015 - 04:40 AM

This thread is about the statistics of guidestar properties - and how they relate to their detection with various sensors and filters. At no point have I ever said that you cannot guide on typical stars with IR. In fact - from the plots above - you could also guide on them with a UV filter - but in both cases you would have greatly reduced signal, and fewer possible guidestars.

Frank

 

 

When autoguiding using short exposures (few seconds) we are usually read noise limited.


Indeed you will be - if you block almost all the signal before it arrives on your sensor. But that is not "we."

Frank

 

 

 

It seems for me that we basically agree that we can guide in NIR.
 

Although this thread is about guide star properties it is fundamentally related to another one concerning the ONAG (from MiIke, now locked at his request) and uses its cut-off wavelength in the analysis.

From what I could tell the ONAG NIR has been the key element and reason for this discussion so far, therefore I feel concern by the topic as soon as it challenges the practical use of NIR for guiding, or the validity of the ONAG arguments/results.

 

I understand that the underlying question becomes: Is it practical and possible to do ONAG NIR guiding with reasonable guider FOV and exposure time with today sensor technologies?

 

My answer to this, as well as many hundreds of ONAG user across 4 years or so, is yes it is possible to do this with guiding exposures in the range of 2 seconds for most setup, we actually do it all the time.
Many ONAG users use a lodestar guider (with the Sony Exview chips) and as John as mentioned  they seldom use the integrated X/Y stage even with a small sensor area.
 

Others, like me use a larger chip like the KAF8300 (APS-C size) and they never have had to hunt for the guide star with the stage. From my personal experience using a Moravian G2-4000 monochrome (KAI4022 chip) there is always at least one, very often many guide stars, available with 2 second exposure at f/8 (RCT 10").

 

I want to make clear that the ONAG does not required a larger guiding chip (a lodestar will do, with the use of the X/Y stage if needed, similar of rotating with an OAG), but it offers the opportunity to do so, which is useful for automation and remote operations. Also this is the only option to do multi-star guiding using the same scope than imaging, (not to mention real time auto-focus) this is maybe a new option for amateurs (professional use constellation guiding since a while) in my opinion it will be a game changer eventually, at least for most setups as long as you can use a larger guider.

 

With short exposures you may still well be read noise limited (> photon noise) even with bright guide stars for most sensors. Now if you are photon noise limited indeed a binning 2x2 still increases your SNR by 2, a 3x3 by 3, and for 4x4 by 4, with whatever you use for guiding (guide scope, OAG, ONAG) it is good thing.



#20 freestar8n

freestar8n

    Vendor - MetaGuide

  • *****
  • Vendors
  • topic starter
  • Posts: 8739
  • Joined: 12 Oct 2007

Posted 27 October 2015 - 05:38 AM

Well - if you are talking about using 2 second exposures to see a guidestar - that is equivalent to a well set up OAG seeing the same typical guidestar with the same SNR in 0.2 seconds - with no filter. The read noise and signal in both exposures will be the same. And if you are happy finding a guidestar with 2 second exposures with a filter - you could see about 29 with no filter. This isn't at all about what is possible or easy - it is about the loss of guidestars with a given camera when used with a filter. The easiest of all in terms of imaging is high end mount with no guiding at all - but that also is not the topic of this thread.

You apparently are in agreement with my analysis and the numbers I have provided - which is great - and my main point. My other point is that the percentage of red stars in our galaxy has no bearing on guidestar color.

You are also in agreement with me that guiding with a filter of any kind should work - given a long enough exposure and/or a large enough sensor.

Your other comments on optimal imaging and so forth are completely opposite my views - and I have no interest in discussing them here. I judge guiding methodology by results obtained with mid-range equipment in long exposures - based on realized fwhm - and I am not aware of results to back your claims. They are commonly repeated anecdote based on simplified and incorrect views of how guiding works in the presence of turbulence and mount errors. They completely ignore the role of latency, and the role of centroid errors with non-Gaussian star spots. A key to guiding a mid-range mount is low latency - and that means you do not have time to waste on a long guide exposure to see a single star that happens to fall on your sensor.

But with a high end mount - where everything is more casual - you can use a big detector and not move it to find a guidestar - and use NIR guiding and long guide exposures - and it should work well - and have the advantage of not needing special flats. I have already said this before - and I hope you agree with it. But - again - with a high end mount that is well set up - many people are going unguided - and getting great results.

Frank

#21 Hilmi

Hilmi

    Vendor-Astronomical Solutions Co.

  • *****
  • Vendors
  • Posts: 5628
  • Joined: 07 Mar 2010
  • Loc: Muscat, Sultanate of Oman

Posted 27 October 2015 - 07:16 AM

Frank may I propose a truce based on agreeing on the following points?

 

  • Guiding with your software is effective, but will not work with ONAG
  • Guiding with ONAG is effective but will not work with Frank's guiding software
  • Optimal guiding with an ONAG is either with a large guide chip or with a guide chip specifically sensitive to IR
  • Guiding with 2 second exposures is one way to do things and works very well with many people (I use 3 second exposure and get perfectly round and tight stars, tested with 20 minute exposure)
  • Guiding with many fast exposures and then calculating average centroid values is another way of doing things, but it is not the only way to guide
  • We all agree to disagree amiably


#22 vpcirc

vpcirc

    Fly Me to the Moon

  • ****-
  • Posts: 5169
  • Joined: 09 Dec 2009

Posted 27 October 2015 - 07:34 AM

Hilmi

ONAG will work with his software

A large Guide chip is not necessary' I'm only going to use one because my FOV is very restricted to begin with at 3742 mm and I want to get rid of rotating

Most Guide cameras will work fine with this

This has come down to a long running disagreement that has apparently spilled over from many different instances.

Most of us really don't care and will wait to see for ourselves or from the examples provided by others and make our decisions then.

I guess it's because I'm from the show me state, I'm too ignorant to give a rat's (&^)(^% about the physics, it either works or it doesn't

Since I'm from the show me state, I'm going to show you which ever way the results turn out to be from short exposures to long. I'm not going to give you a two page explanation, just data and trust you can decide from that.



#23 Mark Sibole

Mark Sibole

    Mercury-Atlas

  • *****
  • Posts: 2853
  • Joined: 30 Nov 2007
  • Loc: Fife Lake Mi.

Posted 27 October 2015 - 08:12 AM

Call me old school but I never went by all of these wild calculations .I grabbed an imager threw it on a scope and made it work.I never did the pixle scale things to match up a camera to a scope.As for guiding ive guided several different ways from using the main tube to an OAG to a mini orion guide scope and never had a problem finding a guide star and never had an issue getting 20 minute to 30 minute sub frames error free all night.I know everyone has their own methoid to their madness but sometimes I think people get to carried away spending hours on math pixel scale arc seconds and so forth.Like I said its to each his own but ill match any of my images with the ones who spend hours on end doing the math to make sute on paper its all right.


  • Faris likes this

#24 Corsica

Corsica

    Vendor (Innovationsforesight)

  • *****
  • Vendors
  • Posts: 633
  • Joined: 01 Mar 2010
  • Loc: PA - USA

Posted 27 October 2015 - 08:22 AM

Frank,

 

I do agree with the fundamental not with the all conclusion, the drop in SNR when using NIR only versus the full spectrum is more around a factor 3 to 5 in average depending of the chip, not 9 or 10, but going deep on this would not bring much more in the discussion and we would not agree probably, this is close enough.

 

If you target a 0.2 seconds for guiding this is a totally different context than a few seconds, obviously our definition of short and long exposures is off by at least a factor 10. Now you enter the world of AO type guiding (using a mount I presume). I do know you will probably not agree with me on that matter, however the turbulence theory is not mine, it as been extensively study since many decades, I think I mentioned few article/websites in my last posts here more, here a more comprehensive list.

 

Fried - Statistics of a Geometric Representation of Wavefront distortion – Nov 1965 JOSA, Vol 55 Number 11
Noll R.J., 1976, "Zernike polynomials and atmospheric turbulence", J. Opt. Soc. Am. A 66, 207-211

http://citeseerx.ist...p=rep1&type=pdf
http://www.ctio.noao...part1/turb.html
From Differential Image Motion to Seeing, A. Tokovinin Publications of the Astronomical Society of the Pacific, 114:1156–1166, 2002 October
http://ao.osa.org/ab...t.cfm?&id=43466

http://www.innovatio...eeing-tutorial/

 

I would argue that those are not anecdotal works and they are used every day across the world, I have some experiences with basic AO (home made for some) which just behave the same way, as it should. I would also dispute  that I am just repeaingt blinding what was written without understating the basic.

 

The fundamental points are:

 

- Short term images (of a star) eventually freeze the tilt/tip component of the wave front (the main aspect for our scope since we are usually below one meter in aperture, yet other wave front error may still exist). The diffracted PSF at best is a Bessel function (circular aperture) not a Gaussian.

 

The FWHM for short exposures (10ms to 100ms depending of the coherence time t0) is given by:

 

FWHM_short = 202.1*lambda/(3.4*r0)

 

Lambda => wavelength, r0 => Fried's parameter (see http://www.innovatio...eing-tutorial/)

 

- Long term exposures (seconds) is the result of averaging many of the above short terms exposures and lead to a Gaussian like star profile, the related FWHM is now:

 

FWHM_long = 202.1*lambda/(r0)

 

Which translates in a 3.4 larger FWHM

 

I am not sure how much improvement one may get using lucking imaging techniques, by drooping the "bad" sub-frames.

The potability to get a good frame (RMS wave front error ~< lambda/6) is given by:

 

P=5.6*exp(-0.16*(D/r0)^2)       (D/r0 > 3.5)

 

Where D is the scope diameter, if we assume and average r0 of 5cm here D/r0~=5, ND P is about 10%, you may have to drop 90% of the guide star frames missing the opportunity to average the noise (supposed white here). My guess is you will end up with the same over all FWHM, or worse, but I do not have much experience in this field, there may be a value that I do not get doing so, that would be a very interesting and fascinating  to topic to discuss, but in another thread.

 

- The angular separation for which the turbulence are still correlated across the FOV is in the range of arc-seconds, it is known as the isoplanatic patch.

 

- r0 and t0 scales as the 6/5th power of the wavelength therefore the FWHM_short and FWHM_long scles down as the 5/6th power of the wavelength. This  is about a 70% reduction in FWHM from 550nm to 850nm.

 

So going fast provides tighter guide stars (if you go fast enough you eventually end with the above FWHM_short value),I fully agree with that statement however the isoplanatic patch is only few arc seconds in size, which means guide star corrections applied few arc-second away from the guide star will not work as well (from a seeing stand point), and it may make things eventually worse since the local seeing at that location in the FOV (let's say one arc-minute away) is not correlated with the guide star anymore (please see the Palomar's example here http://www.innovatio...eeing-tutorial/). This may end with a nice looking guide star on the imager, but we do not usually image just the guide star, although this is useful when doing double star work for instance.

 

Now from a practical stand point it is quite possible and common that a mount (with PEC) short term errors (gear box, ..) is much more than the seeing limit, requiring short term corrections, and if we need to go fast we may eventually even need an AO (tilt/tip device such the SBIG AO-X) which may be a better options to minimize mount latency and inertia (I used to do that with my CGE mount and an AO8, than AOL).

 

If this is the situation we are talking about one should go for it, but keep in mind that when using only one guide star you will have some difficulties to separate the star wander due to the seeing and due to the mount, of course it does not matter much if your mount is much worse left unguided (noise wise) than the seeing.

One way to improve the situation and extract some guidance information related to the mount and not the seeing is the multi star guiding approach, as long that those stars are distant from each others more than the isoplantatic angle, an usually easy requirement for our typical FOV.

 

There are also conditions where the seeing is manly from the boundary layer (near ground) and has a quite large isoplanatic angle, let say if you have a driveway near by, or a roof, any dominant turbulence close to the scope, in this case you could do better with fast guiding.

So I am not sure we are totally on disagreement here, this is more a matter to precising the scope of the problem and the related assumptions. There is not one solution fits all (ONAG or not).

 

P.S: If you want to see nice results using ONAG NIR guidance I would suggest to visit the websites of those users who left a feedback (some are quite know imagers) in our website, my observation site is seldom a good place those days for imaging anything, and those people do a better job than I could anyway.



#25 Corsica

Corsica

    Vendor (Innovationsforesight)

  • *****
  • Vendors
  • Posts: 633
  • Joined: 01 Mar 2010
  • Loc: PA - USA

Posted 27 October 2015 - 08:31 AM

 

Frank may I propose a truce based on agreeing on the following points?

 

  • Guiding with your software is effective, but will not work with ONAG
  • Guiding with ONAG is effective but will not work with Frank's guiding software
  • Optimal guiding with an ONAG is either with a large guide chip or with a guide chip specifically sensitive to IR
  • Guiding with 2 second exposures is one way to do things and works very well with many people (I use 3 second exposure and get perfectly round and tight stars, tested with 20 minute exposure)
  • Guiding with many fast exposures and then calculating average centroid values is another way of doing things, but it is not the only way to guide
  • We all agree to disagree amiably

 

Thanks for the suggestion, I am fine with your suggestion,I doubt we ever settle all the differences here anyway and I have also a business to run and customers to support, not enough time for everything, the life is too short :-)




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