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

Takahashi FSQ106EDX4 Chromatic Aberration

  • Please log in to reply
26 replies to this topic

#1 thekubiaks

thekubiaks

    Vostok 1

  • *****
  • topic starter
  • Posts: 165
  • Joined: 16 Dec 2019

Posted 20 October 2020 - 10:05 AM

I need your guys opinion on something.   I bought a new Takahashi FSQ106EDX4 and started imaging M31.   First thing I noticed is Chromatic Aberation in all four corners.   I contacted Takahashi Houston and they forwarded my subs and master image to Japan for analysis.  Japan said that this was some kind of tangential something but within the specs of this scope.  

 

I figured a scope of this caliber would not have this problem.

 

He said I could fix it with some trick in Pixinsight.   I'm waiting to hear from Pixinsight to tell me how that is done.

 

Fortunately I am within the 90 day return window, I may do that.

 

What do you guys think of this image?  All four corners are the same with the aberration pointed towards the middle of the image.  The CA starts at about 30mm from center and gets worse as you get to the outer edges.

 

Keep or return or fix in post process??

 

+++ There is another thread discussing this too and here is an additional image

 

https://www.cloudyni...ars/?p=10591376

Attached Thumbnails

  • FSQ CA.jpg

Edited by thekubiaks, 20 October 2020 - 10:13 AM.


#2 dhaval

dhaval

    Vendor

  • *****
  • Vendors
  • Posts: 1,964
  • Joined: 21 Jul 2008
  • Loc: Round Rock, TX

Posted 20 October 2020 - 10:42 AM

I haven't used a TAK FSQ106EDX4, but this is not looking right unless there is something significantly wrong with the rest of the imaging train. 

 

I would return this for sure.

 

CS!



#3 dcornelis

dcornelis

    Ranger 4

  • -----
  • Posts: 301
  • Joined: 09 Oct 2012

Posted 20 October 2020 - 10:55 AM

A not well collimated tak FSQ probably, happens more than it should, a Belgian astrophotographer his name escapes me, has sent back 3 or 4 before he got a good one from his dealer in France.



#4 thekubiaks

thekubiaks

    Vostok 1

  • *****
  • topic starter
  • Posts: 165
  • Joined: 16 Dec 2019

Posted 20 October 2020 - 10:59 AM

I haven't used a TAK FSQ106EDX4, but this is not looking right unless there is something significantly wrong with the rest of the imaging train. 

 

I would return this for sure.

 

CS!

The imaging train from the back of the FSQ is the CAA, an M72 to M68 adapter, Gerg Neumann CTU, M68 30.8mm extension, ZWO OAG, ZWO 2" x 7 EFW with all Chroma 3nm LRGB and NB, and then ASI6200MM-P
 


Edited by thekubiaks, 20 October 2020 - 10:59 AM.


#5 rgsalinger

rgsalinger

    Fly Me to the Moon

  • *****
  • Moderators
  • Posts: 7,363
  • Joined: 19 Feb 2007
  • Loc: Carlsbad Ca

Posted 20 October 2020 - 11:19 AM

I don't think that there is any refractor made that will cover a full frame camera without a flattener. I know that my AP155 won't do it. 



#6 Madratter

Madratter

    Voyager 1

  • *****
  • Posts: 12,875
  • Joined: 14 Jan 2013

Posted 20 October 2020 - 11:52 AM

I don't think that there is any refractor made that will cover a full frame camera without a flattener. I know that my AP155 won't do it. 

It is a 4 element system. Flattening built in.


  • rgsalinger, bobzeq25 and thekubiaks like this

#7 Stelios

Stelios

    Voyager 1

  • *****
  • Moderators
  • Posts: 10,449
  • Joined: 04 Oct 2003
  • Loc: West Hills, CA

Posted 20 October 2020 - 02:58 PM

.  The CA starts at about 30mm from center and gets worse as you get to the outer edges.

30mm from center, so you have a good image circle of 60mm? If so, I would keep the scope, as your sensor will likely be much smaller. OTOH, how would you have got those images then?

 

If you meant stars are good in a 30mm image circle, then I would return it. You don't want to be fixing stars in post-processing, or having to crop.

 

Your images are very reminiscent of what happens when a flattener has been spaced too far. I bet the corrector elements inside the FSQ need a fraction of a mm adjustment. But that should be Takahashi's problem, not yours. 



#8 thekubiaks

thekubiaks

    Vostok 1

  • *****
  • topic starter
  • Posts: 165
  • Joined: 16 Dec 2019

Posted 20 October 2020 - 03:12 PM

30mm from center, so you have a good image circle of 60mm? If so, I would keep the scope, as your sensor will likely be much smaller. OTOH, how would you have got those images then?

 

If you meant stars are good in a 30mm image circle, then I would return it. You don't want to be fixing stars in post-processing, or having to crop.

 

Your images are very reminiscent of what happens when a flattener has been spaced too far. I bet the corrector elements inside the FSQ need a fraction of a mm adjustment. But that should be Takahashi's problem, not yours. 

The stars are good in a 30mm circle, fixing stars post processing should not be required.  That is Takahashi's problem but they say the scope is in specs.  They say just a tangential something, maybe astigmatism.  Either way, for 6 grand, it should be flawless and they should be tripping all over themselves to offer to repair it.



#9 sn2006gy

sn2006gy

    Apollo

  • ****-
  • Posts: 1,084
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 20 October 2020 - 03:24 PM

Do you have a sub to share to poke around?



#10 Madratter

Madratter

    Voyager 1

  • *****
  • Posts: 12,875
  • Joined: 14 Jan 2013

Posted 20 October 2020 - 03:46 PM

My opinion, is that scopes like Taks are worth the money BECAUSE of the well corrected images with large image circles. If that image circle is really only good to 60mm instead of 80mm, then that is what they should be advertising. They can call it what they want, but IMO if the scope has significant color problems at 60mm and they are advertising 88mm, that should be their problem.

 

Now in fairness, the actual advertising is here:

 

https://www.takahash...or-telesco.html

 

And they clearly show spot diagrams with some color aberrations visible at 22mm from center. With the prevalence of cameras with small pixel pitches, color affects will indeed be visible when air is still enough to not smear everything.

 

In the actual advertising, they advertise that it is flat to 88mm without a reducer. But the color correction is advertised with the spot diagrams. It would have been very helpful if those spot diagrams went out to 44mm, but they stop at 22mm.



#11 thekubiaks

thekubiaks

    Vostok 1

  • *****
  • topic starter
  • Posts: 165
  • Joined: 16 Dec 2019

Posted 20 October 2020 - 04:03 PM

My opinion, is that scopes like Taks are worth the money BECAUSE of the well corrected images with large image circles. If that image circle is really only good to 60mm instead of 80mm, then that is what they should be advertising. They can call it what they want, but IMO if the scope has significant color problems at 60mm and they are advertising 88mm, that should be their problem.

 

Now in fairness, the actual advertising is here:

 

https://www.takahash...or-telesco.html

 

And they clearly show spot diagrams with some color aberrations visible at 22mm from center. With the prevalence of cameras with small pixel pitches, color affects will indeed be visible when air is still enough to not smear everything.

 

In the actual advertising, they advertise that it is flat to 88mm without a reducer. But the color correction is advertised with the spot diagrams. It would have been very helpful if those spot diagrams went out to 44mm, but they stop at 22mm.

You are right on the money, they stopped at 22mm for a reason.  My stars are indeed flat out to 44mm though.


Edited by thekubiaks, 20 October 2020 - 04:05 PM.


#12 thekubiaks

thekubiaks

    Vostok 1

  • *****
  • topic starter
  • Posts: 165
  • Joined: 16 Dec 2019

Posted 20 October 2020 - 04:07 PM

Do you have a sub to share to poke around?

https://www.dropbox....iSKeDallaa?dl=0
 



#13 Madratter

Madratter

    Voyager 1

  • *****
  • Posts: 12,875
  • Joined: 14 Jan 2013

Posted 20 October 2020 - 04:16 PM

As for a possible strategy to fix this in PixInsight:

 

How fast color information varies is far less important to the eye than lumenance. You could create a mask using  PixelMath that allows more and more correction towards the edges of the frame. Then take the RGB color image and in CIE L*a*b extract the channels using the ChannelExtraction process. Then using your mask, do a convolution (blur) on the color channels a* and b*. You can be fairly aggressive, especially since you are using a mask and the center portion won't be affected much if at all. Then do a channel combination back to RGB using ChannelCombination in CIE L*a*b* mode again. L should have been untouched. a* and b* are your convolved channels.

 

I do a variation of this without a mask on most of my images with a fairly small convolution (around 2) to help knock out some of the color noise.


Edited by Madratter, 20 October 2020 - 04:16 PM.


#14 thekubiaks

thekubiaks

    Vostok 1

  • *****
  • topic starter
  • Posts: 165
  • Joined: 16 Dec 2019

Posted 20 October 2020 - 09:25 PM

The images produced by the FSQ106 look good until you really start going deep.  I think with the new small micron full frame cameras, the imperfections that were below the surface are now appearing.  Here is this weeks M31, the skies and M31 cooperated, still more processing needed. You will definitely see the CA once you get to a viewing size of 200% +.    I couldn't get the huge image down below 500KB for this forum preview.

 

JPG

 

https://www.dropbox....M31 V3.jpg?dl=0

 

XISF

 

https://www.dropbox....31 V3.xisf?dl=0


Edited by thekubiaks, 20 October 2020 - 09:26 PM.


#15 AtmosFearIC

AtmosFearIC

    Apollo

  • -----
  • Posts: 1,236
  • Joined: 10 Dec 2015
  • Loc: Melbourne

Posted 20 October 2020 - 10:21 PM

Given that you’re looking at 20mm off axis the fact that it doesn’t have 8 micron spot sizes isn’t surprising at all. This lateral CA is something I’ve seen I a LOT of FSQ106 images and it is where there is a slightly different focal length for each colour. Each channel has a low spot size (well corrected) but a different focal length so blue on the inside of the stars and red on the outside.

In PixInsight you use StarRegistration on the RGB masters (use G as the reference) and do a dynamic distortion correction on the R&B.

It’s a part of my workflow with a Mewlon 250 as it shows the same kind of lateral CA.

#16 alan.dang

alan.dang

    Ranger 4

  • *****
  • Posts: 338
  • Joined: 15 Dec 2011

Posted 20 October 2020 - 11:04 PM

The FSQ-130ED and VSD100 are probably the only designs that can cover a medium format sensor with small pixels and clean stars. Both are out of production and more expensive than the FSQ106. The VSD100 has limited back focus and limited ability to support auto focusing which makes it a tougher scope to work with than the FSQ106.

Other good options are the TEC110FL with flattener or a AP130GTX with QuadTCC. Again, both more expensive than the FSQ-106ED.

Dollar for dollar, it’s hard to beat the proven performance of the FSQ-106ED without spending more. That said, it is from the era of 9 micron pixels and won’t hold up to today’s ultra resolution sensors.

#17 gregbradley

gregbradley

    Mariner 2

  • -----
  • Posts: 208
  • Joined: 23 Jan 2008

Posted 21 October 2020 - 03:39 AM

You are right on the money, they stopped at 22mm for a reason.  My stars are indeed flat out to 44mm though.

That is 22mm from the centre, ie. 44mm across the whole circle.

 

Do you have any other camera that you can test it with? It doesn't look like chromatic aberration but rather some odd reflection or perhaps something to do with the filters.

 

Did you refocus for each filter? Have you used these filters for a long time? Its looks to me more like you got bad seeing when you did the blue subexposures or not quite in focus. Blue subs often have the most bloated stars so I often will use decon on the worst coloured sub to reduce star sizes to match the other colours better. Here it looks like blue stars are larger than the other colours. Your filters may not be parfocal, seeing could have deteriorated, tracking may have gotten worse or your focus went off as FSQ is quite sensitive to temperature shifts. Every 1C change you should refocus. You should also check the focus point of each LRGB filter and work out what the offsets need to be for the different filters.

 

Unless you do this assuming its the scope may be a dead end as you've got the wrong target.

 

I have had 2 FSQs and they never looked like that with 16803 sensors and full frame 11001 sensor.

 

Greg.


Edited by gregbradley, 21 October 2020 - 03:42 AM.


#18 gregbradley

gregbradley

    Mariner 2

  • -----
  • Posts: 208
  • Joined: 23 Jan 2008

Posted 21 October 2020 - 03:45 AM

The FSQ-130ED and VSD100 are probably the only designs that can cover a medium format sensor with small pixels and clean stars. Both are out of production and more expensive than the FSQ106. The VSD100 has limited back focus and limited ability to support auto focusing which makes it a tougher scope to work with than the FSQ106.

Other good options are the TEC110FL with flattener or a AP130GTX with QuadTCC. Again, both more expensive than the FSQ-106ED.

Dollar for dollar, it’s hard to beat the proven performance of the FSQ-106ED without spending more. That said, it is from the era of 9 micron pixels and won’t hold up to today’s ultra resolution sensors.

I find that hard to believe. FSQ is one of the most perfect optical scopes you can get.

There are plenty of images using KAF8300 which is only 4.54microns so not that different to 3.76 microns.

I'd be checking the filters for focus position and try another camera before jumping to conclusions that it is the scope. 

 

Greg.



#19 gregbradley

gregbradley

    Mariner 2

  • -----
  • Posts: 208
  • Joined: 23 Jan 2008

Posted 21 October 2020 - 03:47 AM

I find that hard to believe. FSQ is one of the most perfect optical scopes you can get.

There are plenty of images using KAF8300 which is only 4.54microns so not that different to 3.76 microns.

I'd be checking the filters for focus position and try another camera before jumping to conclusions that it is the scope. 

 

Greg.

 

I find that hard to believe. FSQ is one of the most perfect optical scopes you can get.

There are plenty of images using KAF8300 which is only 4.54microns so not that different to 3.76 microns.

I'd be checking the filters for focus position and try another camera before jumping to conclusions that it is the scope. 

 

The "chromatic aberration" has 2 little bumps top and bottom that look like tracking errors when imaging the blue subexposures.

Bad tracking, wind gusts, out of focus for blue, temperature shifts seem a more likely explanation. Its also only one image. Are there several like this?

 

Greg.



#20 Janco

Janco

    Mariner 2

  • *****
  • Posts: 279
  • Joined: 21 Aug 2017
  • Loc: Johannesburg, South Africa

Posted 21 October 2020 - 06:08 AM

I am a bit confused... which camera has a 60 mm diagonal sensor :O



#21 AtmosFearIC

AtmosFearIC

    Apollo

  • -----
  • Posts: 1,236
  • Joined: 10 Dec 2015
  • Loc: Melbourne

Posted 21 October 2020 - 06:55 AM

I am a bit confused... which camera has a 60 mm diagonal sensor :O

 

https://www.onsemi.c...KAF-50100-D.PDF



#22 sn2006gy

sn2006gy

    Apollo

  • ****-
  • Posts: 1,084
  • Joined: 04 May 2020
  • Loc: Austin, TX

Posted 21 October 2020 - 09:53 AM

Ran your sub through ccdware and it looks super flat, so that's good. 

 

I know several people running 6200+FSQ 106 and they get beautiful stars. However, I've also seen people suffer with blue on other scopes with the 6200 and baader filters - I would expect that chroma would be fine.

 

looks like this was imaged with SGP - were the flats/darks/bias all done at same gain? have you tried flipping your filters? Chroma says their filters are directional

 

https://www.chroma.c...and-orientation



#23 ArkabPriorSol

ArkabPriorSol

    Mariner 2

  • *****
  • Posts: 210
  • Joined: 23 Dec 2006
  • Loc: Austin, TX

Posted 21 October 2020 - 12:03 PM

I have a FSQ-106N and the QHY600 and what I've noticed is this sensor on a fast optical system is incredibly sensitive to precise focus. If my focus is not perfect I start to see more pronounced elongation of stars near the edge of the frame.

 

I wonder if your focus was slightly worse for one color filter than it is for the other. Color combining slightly out of focus red with slightly better focus blue might give the illusion of chromatic aberration. Is this something you can check for? Try comparing your FWHM between the color masters.

 

I'm struggling with this configuration too since I'm using a fairly imprecise moto-focus on the Takahashi focuser. I'm considering upgrading to a Nightcrawler with hopes to improve focus precision and eliminate tilt and star elongation.



#24 thekubiaks

thekubiaks

    Vostok 1

  • *****
  • topic starter
  • Posts: 165
  • Joined: 16 Dec 2019

Posted 21 October 2020 - 12:59 PM

This is interesting

Attached Thumbnails

  • FSQ106 Full Field Spot.jpg

  • sharkmelley likes this

#25 gregbradley

gregbradley

    Mariner 2

  • -----
  • Posts: 208
  • Joined: 23 Jan 2008

Posted 25 October 2020 - 01:12 AM

This is interesting

The thing is this is only 1 sample image, the stars generally do not have a blue halo around them which would indicate better that there was chromatic aberration. But rather blue specks inside the area of brighter stars. It seems to be more of a reflection or perhaps a processing error.

 

Its easy to blame the scope. I have never seen chromatic aberration from an FSQ106EDX. I used one for a few years and a 16803 sensor which is 50% taller than the IMX455.

 

More likely some handling error of an unknown nature. Is it even a reflection artifact off the camera's glass cover?

 

Bottomline I think is no easy conclusion from one image as a sample to analyse.

 

Greg.


Edited by gregbradley, 25 October 2020 - 01:12 AM.



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


Recent Topics






Cloudy Nights LLC
Cloudy Nights Sponsor: Astronomics