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

Reflection Artifacts in the ASI6200

  • Please log in to reply
457 replies to this topic

#276 sharkmelley

sharkmelley

    Soyuz

  • *****
  • Posts: 3,730
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 21 September 2020 - 06:24 AM

Inevitably any explanation of the "((o)) rings" effect will be quite technical but I'll do my best to make it understandable.  The formula giving the intensity of Airy rings for centrally obstructed optics can be found in the "Obscured Airy Pattern" section of the Wiki Airy Disk article.  It uses a first order Bessel function of the first kind, which is the function BesselJ in Microsoft Excel.  Anyone, including me, can plot it!  So here is a plot of a slice through the circular Airy pattern of a star with no central obstruction (CO) and a star with a 55% CO (e.g. the RH200 telescope):

 

AiryRings1.jpg

 

The effect of the CO is to narrow the width of the central peak and to throw more energy into the 1st and 2nd rings.  Everyone is already familiar with this.  But the amplitude of successive rings dies down quite quickly which means that few people are familiar with what happens in those outer rings. So here's a plot where I have applied a scaling function to the rings, to make the effect visible:

 

AiryRings2.jpg

 

There is an obvious periodic modulation to the amplitude of the CO rings, with a maximum approximately every 4.5 rings.

 

We are now in a position to relate this pattern to the example image from AtmosFearIC (a few posts ago), which I've copied here for convenience:

 

RH200_QHY183_HaFilter.jpg

 

This example used the QHY183 camera on the RH200 with f-ratio f/3. Using the well known formula for unobstructed optics, the radius of the first zero in the Airy pattern is:

  • radius = 1.22 x wavelength x focal_ratio

This gives 2.4 microns for the H-alpha wavelength (656nm) and f/3 optics. The spacing of successive rings is slightly less than this. But the pixel pitch of the QHY183 camera is also approx 2.4 microns so the rings of unobstructed f/3 optics are too closely spaced to be resolved by the camera. 

 

However my plot above shows that there is an amplitude modulation that occurs in the case of a 55% CO (like the RH200) with a cycle of approx 4.5 rings.  This is easily sampled by the camera and my conjecture is that this is exactly what we are seeing in the above image.

 

Mark

 


  • R Botero, jdupton, vehnae and 9 others like this

#277 Der_Pit

Der_Pit

    Surveyor 1

  • *****
  • Posts: 1,601
  • Joined: 07 Jul 2018
  • Loc: La Palma

Posted 21 September 2020 - 06:53 AM

Chapeau Mark!  Sheer brilliance bow.gif


  • sharkmelley, rockstarbill and tkottary like this

#278 sharkmelley

sharkmelley

    Soyuz

  • *****
  • Posts: 3,730
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 21 September 2020 - 07:00 AM

Now let's return to the Gamma Cygni diffration ring example in the OP:

 

asi6200_ha_1.jpg

 

The scope was AG Optical 10" iDK (f/6.7 with central obstruction 56%) and an ASI6200 camera with an effective pixel pitch of 7.52 microns since the image was 2x2 binned.

 

Using the formula 

radius = 1.22 x wavelength x focal_ratio

gives an unobstructed ring spacing of slightly less than 5.36 microns for H-alpha.

The amplitude modulation for a central obstruction of 56% again gives a cycle of approx 4.5 rings i.e 4.5*5.36 = 24.1 microns.

With an effective pixel pitch of 7.52 microns this is a spacing of slightly less than 3.2 pixels which pretty much agrees with what we see in the image.

 

Mark


  • Dean J., rockstarbill, Lead_Weight and 2 others like this

#279 xthestreams

xthestreams

    Explorer 1

  • -----
  • Posts: 66
  • Joined: 18 Feb 2020
  • Loc: Melbourne, Australia

Posted 21 September 2020 - 08:06 AM

Is it too soon to say my brain hurts? 



#280 RogeZ

RogeZ

    Apollo

  • *****
  • Posts: 1,497
  • Joined: 21 Nov 2004
  • Loc: Palm Beach Gardens, FL

Posted 21 September 2020 - 02:37 PM

And of course, the bigger aperture throws more energy into the rings regardless of focal ratio which we also see from the data here.

#281 Dean J.

Dean J.

    Viking 1

  • -----
  • Posts: 652
  • Joined: 13 Aug 2011
  • Loc: Vista, CA

Posted 21 September 2020 - 09:23 PM

So, judging from Mark's great analysis and the lack of further discussion, it seems that the theory/argument that there is something wrong with the camera or the filters has been abandoned. 


  • rockstarbill likes this

#282 rockstarbill

rockstarbill

    Cosmos

  • *****
  • topic starter
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 21 September 2020 - 09:29 PM

So, judging from Mark's great analysis and the lack of further discussion, it seems that the theory/argument that there is something wrong with the camera or the filters has been abandoned. 

At this point the camera seems to be absolved of any wrongdoing. The filters are still in question, and a test will be done with a 5nm HA Astrodon filter to see if there is any difference. This pertains more to the grid pattern than any of the other effects.

 

I did find that the filter wheel carousel in my EFW 2" seems to have a problem. It is not seating correctly, which suggests it is warped or something else. It will turn fine but looks to be making an very minor oblong movement. This could explain why my tilt was so bizarrely bad (although was very minor at first light), as well as some of the strange optical behavior I was seeing. One of the screws will not seat properly to allow it to thread into the center holes, although if I pull the same screw out, it will thread into the same hole just fine without the carousel in place.

 

I passed this feedback to Sam, hopefully I can get this replaced so I can test some more and hopefully be done with this mess so I can get back to what I enjoy most -- imaging. wink.gif


Edited by rockstarbill, 21 September 2020 - 09:54 PM.


#283 arjanb

arjanb

    Sputnik

  • *****
  • Posts: 44
  • Joined: 11 Dec 2018
  • Loc: Raleigh, NC

Posted 21 September 2020 - 11:37 PM

This is very useful. I see the ((o)) pattern still exists on the 6200 image, but far less so. 

 

If you get a chance and can point the scope at Sadr (Gamma Cyg) in Cygnus and fire off a 5 min HA exposure, I would love to see that sub.  

this is a 2 minute exposure at Gain 100, Offset 10 (for some reason it dropped to 10 from 50).

 

Scope = RASA11". Astronomik MaxFR 2" 12nm Ha filter. Can't do 5 minutes because i think it would overexpose everything at F/2.

 

Sadr.jpg

 

i would be interested to hear what you think makes the "stipple" pattern of the diffraction spikes shooting off the star. I feel it's some internal reflection in probably some Baader UFC adapter or even Corrector. I heard someone say it's just black coated aluminum which isnt the best anti-reflective esp in IR.


  • rockstarbill likes this

#284 rockstarbill

rockstarbill

    Cosmos

  • *****
  • topic starter
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 21 September 2020 - 11:39 PM

https://www.cloudyni...200/?p=10521229

 

A few posts up. Well done by Mark to show that CO based telescopes can produce this effect and it is completely normal. As odd as that may be to think through. 



#285 sharkmelley

sharkmelley

    Soyuz

  • *****
  • Posts: 3,730
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 22 September 2020 - 02:15 AM

this is a 2 minute exposure at Gain 100, Offset 10 (for some reason it dropped to 10 from 50).

 

Scope = RASA11". Astronomik MaxFR 2" 12nm Ha filter. Can't do 5 minutes because i think it would overexpose everything at F/2.

 

attachicon.gifSadr.jpg

 

i would be interested to hear what you think makes the "stipple" pattern of the diffraction spikes shooting off the star. I feel it's some internal reflection in probably some Baader UFC adapter or even Corrector. I heard someone say it's just black coated aluminum which isnt the best anti-reflective esp in IR.

It looks remarkably similar to the RASA8/ASI2600MC combination in post #264 by @buckeyestargazer.  It showed clearest in the green channel of that image:

 

RASA8_ASI2600MC.jpg

 

I'll need to think about the cause but one obvious question is why there are so many diffraction spikes.

 

Mark


  • arjanb likes this

#286 rockstarbill

rockstarbill

    Cosmos

  • *****
  • topic starter
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 22 September 2020 - 02:17 AM

It showed clearest in the green channel of that image:

 

The hue of the AR coating is green. I don't mention this to be a simpleton. Roland called this out as something that could influence imaging with optical systems with green coating. 



#287 sharkmelley

sharkmelley

    Soyuz

  • *****
  • Posts: 3,730
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 22 September 2020 - 02:33 AM

The hue of the AR coating is green. I don't mention this to be a simpleton. Roland called this out as something that could influence imaging with optical systems with green coating. 

Which AR coating?

 

Mark



#288 rockstarbill

rockstarbill

    Cosmos

  • *****
  • topic starter
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 22 September 2020 - 02:34 AM

Which AR coating?

 

Mark

Window.

 

https://ap-gto.group...n/message/72487

 

The context missing there in that post, was that it was about reflections caused by windows.


Edited by rockstarbill, 22 September 2020 - 02:39 AM.


#289 sharkmelley

sharkmelley

    Soyuz

  • *****
  • Posts: 3,730
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 22 September 2020 - 04:41 AM

It looks remarkably similar to the RASA8/ASI2600MC combination in post #264 by @buckeyestargazer.  It showed clearest in the green channel of that image:

 

attachicon.gifRASA8_ASI2600MC.jpg

 

I'll need to think about the cause but one obvious question is why there are so many diffraction spikes.

I'm a bit slow - I didn't immediately recognise the significance of the RASA!

 

I now realise that the cause of the diffraction spikes is probably the camera cable(s) passing in front of the aperture.  If we take the spider vanes on Newtonians as an example, they cause diffraction spikes whose coloured fringes have a spacing dependent on the wavelength of the light and on the thickness of the spider vane. In the same way, the spacing of the coloured fringes on the diffraction spikes caused by the cable(s) will be dependent on the wavelength of the light and on the thickness of the cable.

 

In other words this is an issue totally unrelated to our earlier discussions. 

 

Mark


Edited by sharkmelley, 22 September 2020 - 04:46 AM.

  • R Botero and rockstarbill like this

#290 rockstarbill

rockstarbill

    Cosmos

  • *****
  • topic starter
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 22 September 2020 - 04:50 AM

I'm a bit slow - I didn't immediately recognise the significance of the RASA!

 

I now realise that the cause of the diffraction spikes is probably the camera cable(s) passing in front of the aperture.  If we take the spider vanes on Newtonians as an example, they cause diffraction spikes whose coloured fringes have a spacing dependent on the wavelength of the light and on the thickness of the spider vane. In the same way, the spacing of the coloured fringes on the diffraction spikes caused by the cable(s) will be dependent on the wavelength of the light and on the thickness of the cable.

 

In other words this is an issue totally unrelated to our earlier discussions. 

 

Mark

Agreed.

 

My filter wheel is faulty, so my ability to continue testing will be hampered by my replacement. 

 

Meanwhile I will produce FLI ML16200 data and share that.


Edited by rockstarbill, 22 September 2020 - 04:58 AM.


#291 rockstarbill

rockstarbill

    Cosmos

  • *****
  • topic starter
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 24 September 2020 - 11:08 PM

Quick update.

 

Monkeybird747 was able to obtain a 5nm 50mm Round Astrodon filter. Testing on his new to him AGO 12.5" IDK showed that the halo grid pattern was still present. He will have to share the data, as I dont have it. The manufacturer suggested getting a better narrowband filter, so we tested the best in the business. No change.

 

On another note, I will not be able to contribute to this any further as I sold the camera and bought a FLI PL16803 instead. 


Edited by rockstarbill, 24 September 2020 - 11:09 PM.


#292 Monkeybird747

Monkeybird747

    Viking 1

  • *****
  • Posts: 839
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 24 September 2020 - 11:09 PM

My new-to-me 12.5" AGO iDK arrived yesterday, and I had a very small window to do some testing. One of the tests was to put my Astrodon 5nm Ha in the same wheel as my Chroma 3nm Ha and take some frames. Deneb, mag 2, 60s, bin 2, gain 100. Those settings were not purposely chosen, they were just what was already set, and I was firing though small holes in the clouds before total overcast settled in. I'll post up some data later, but I can confirm that both the large scale reflection grid, as well as the small concentric rings around the star cores, were present in BOTH filters at 60s. So....

 

Not sure where that leaves us. Here are my remaining burning questions:

  1. Are the grid pattern and concentric rings related and share a common cause, or are we looking at two different effects?
  2. If it happens with two different filters from two different high-quality manufacturers, at two different bandpasses, what does that say about the camera itself? Could the sensor window be complicit in all this by allowing some of these filter reflections to reach the sensor?
  3. Most importantly, what can be done to mitigate? I'm not interested in any talk of "don't shoot with bright stars in the frame". That would be some ASI1600/Panasonic (not ASI's fault) work-around style b.s., and I didn't pay $5k for a camera to have to point it delicately about the cosmos.

This would appear to make this camera not good for narrowband with medium to fast mirror optics. Why can't we have nice things?!

 

So what needs to happen here? Replace sensor window with better one? Different filter coatings? Different bandpass?

 

I declare Shenanigans! ~rabble-rabble-rabble-rabble~


  • rockstarbill likes this

#293 Lead_Weight

Lead_Weight

    Surveyor 1

  • *****
  • Posts: 1,690
  • Joined: 21 Jul 2016
  • Loc: Houston

Posted 24 September 2020 - 11:31 PM

What way were your filters facing? Highly reflective side should face the camera.



#294 rockstarbill

rockstarbill

    Cosmos

  • *****
  • topic starter
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 24 September 2020 - 11:32 PM

What way were your filters facing? Highly reflective side should face the camera.

That's the rub. Unlike my test of the LRGB Astrodon filters, there was no difference in the narrowband 5nm. The filter is the same on both sides. We tested and validated this with the visual edge test and a pen test. 

 

Note: The edge was visible on both sides. 


Edited by rockstarbill, 24 September 2020 - 11:34 PM.


#295 Monkeybird747

Monkeybird747

    Viking 1

  • *****
  • Posts: 839
  • Joined: 31 Aug 2017
  • Loc: Louisville, KY

Posted 24 September 2020 - 11:37 PM

I have the reflective side of all my filters facing camera. With the Astrodon 5nm Ha it is not as clear which side is the highly reflective side. I did the pen-tip reflection test, and on both sides of the filter the pen tip touches its reflection. I can flip the Astrodon and test, but flipping has had no effect on the large grids reflections.
  • rockstarbill likes this

#296 rockstarbill

rockstarbill

    Cosmos

  • *****
  • topic starter
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 24 September 2020 - 11:38 PM

narrow band filter is the key point

it will become better if you use a better narrow band filter absorb some light instead of reflecting lights

We just used a 5nm Astrodon filter. Problem reproduced again. What now? There aren't any better filters for this problem to hide behind. 

 

You could do this yourself (as I asked you to do multiple times) and build a camera with a fused silica AR coated window, test it with Astrodon filters if you think filter quality plays into this. Try something else if needed. Clearly the filter change alone did not provide the relief you assumed (wrongly) would happen. Clearly this camera suffers from this problem, yet all you have offered is trite and useless alternatives.

 

Asking your customers to solve problems for you is not cool. There is a problem, and all you have told us is to get better filters, that didn't solve the problem. The onus here needs to be on you now to help. Customers don't pay $4000 a pop to be your beta testers. Especially not with a production product. 


Edited by rockstarbill, 24 September 2020 - 11:48 PM.


#297 sharkmelley

sharkmelley

    Soyuz

  • *****
  • Posts: 3,730
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 24 September 2020 - 11:48 PM

Not sure where that leaves us. Here are my remaining burning questions:

  1. Are the grid pattern and concentric rings related and share a common cause, or are we looking at two different effects?
  2. If it happens with two different filters from two different high-quality manufacturers, at two different bandpasses, what does that say about the camera itself? Could the sensor window be complicit in all this by allowing some of these filter reflections to reach the sensor?
  3. Most importantly, what can be done to mitigate? I'm not interested in any talk of "don't shoot with bright stars in the frame". That would be some ASI1600/Panasonic (not ASI's fault) work-around style b.s., and I didn't pay $5k for a camera to have to point it delicately about the cosmos.

Without seeing your examples, it is not possible to comment.  However, there are some generic things we do know.

  1. The large scale grid pattern and the concentric rings have different causes.  The grid pattern involves diffraction off the sensor microlens array combined with a reflection off the filter.  The concentric rings are caused by physical optics - i.e. a pattern in the Airy disk rings caused by a large central obstruction on diffraction limited optics.  
  2. There is no problem with the camera.  Also the sensor window is designed to let all wavelengths reach the sensor so the sensor window cannot be blamed.
  3. There's no obvious way to mitigate either problem.  However, maybe some sensors have more reflective microlens arrays than others and hence show stronger large scale grid patterns.

Mark



#298 rockstarbill

rockstarbill

    Cosmos

  • *****
  • topic starter
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 24 September 2020 - 11:51 PM

Without seeing your examples, it is not possible to comment.  However, there are some generic things we do know.

  1. The large scale grid pattern and the concentric rings have different causes.  The grid pattern involves diffraction off the sensor microlens array combined with a reflection off the filter.  The concentric rings are caused by physical optics - i.e. a pattern in the Airy disk rings caused by a large central obstruction on diffraction limited optics.  
  2. There is no problem with the camera.  Also the sensor window is designed to let all wavelengths reach the sensor so the sensor window cannot be blamed.
  3. There's no obvious way to mitigate either problem.  However, maybe some sensors have more reflective microlens arrays than others and hence show stronger large scale grid patterns.

Mark

Good points to mention.

 

The QHY showed less propensity to cause this problem, but it uses a different grade sensor. The same grade of sensor that ZWO called on their Facebook page as a "waste of money". Seems it may have some truth that it wasnt a waste at all. Perhaps that sensor is better equipped to handle the light grasp of a large reflector/cassegrain telescope design.  



#299 sharkmelley

sharkmelley

    Soyuz

  • *****
  • Posts: 3,730
  • Joined: 19 Feb 2013
  • Loc: UK

Posted 25 September 2020 - 12:05 AM

The QHY showed less propensity to cause this problem, but it uses a different grade sensor. 

You mean someone has already tested them side by side?  I would be surprised if QHY's more expensive industrial version has a less reflective microlens array.

 

Mark


Edited by sharkmelley, 25 September 2020 - 12:05 AM.


#300 rockstarbill

rockstarbill

    Cosmos

  • *****
  • topic starter
  • Posts: 7,564
  • Joined: 16 Jul 2013
  • Loc: Snohomish, WA

Posted 25 September 2020 - 12:07 AM

You mean someone has already tested them side by side?  I would be surprised if QHY's more expensive industrial version has a less reflective microlens array.

 

Mark

Side by side, no. 

 

This post has data from the QHY600

 

https://www.cloudyni...200/?p=10515579

 

Way less going on here, although the center star does show some minor artifacts. No grid though, and that is a very bright star at 20 minutes exposure. The ASI6200 will show the grid very easily if the same test was done. 




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