FSQ 106 Purple stars

[Ken82]

I've been dealing with a purple star problem with my new Takahashi fsq 106 and was hoping to get some feedback as i believe it to be chromatic aberration as do others. 

I had sent my scope back to takahashi Europe for analysis and adjustment as i believed the problem was a simple miss alignment but there latest response is that the scope is a fine example.

I have done some extensive tests with this scope and compared it to my FSQ 85. Image below shows a direct comparison using the same camera/filters etc etc. Image is a simple alignment/combination of RGB and autostretch in pixinsight (no saturation has been altered) the difference is very clear. On the left hand side (FSQ 85) the star colour is nice and uniform and the right hand side (FSQ 106) the star clearly shows purple fringing or chromatic aberration. 
So this eliminates the problem being camera/filter related.  

I also contacted the pixinsight developers who gave me the following response - "In order to assess this, you need to look at the position angles of the fringes around the field. There is a general "rotation" of color from what it looks like to me when I inspected your RGB image. So... I disagree with vendor...this appears like optical aberrations (chromatic) to me.

So i checked my images for the rotation of colour and they are right.

Top right with purple stars pointing inwards 

Bottom right with purple stars pointing inwards

Bottom left with purple pointing inwards

I also sent my images to a couple of  independent optics specialists who gave me the following response    "You do have chromatic aberration in your telescope hence the colour on the stars."

Interestingly a friend mentioned that if your using a mono camera then chromatic aberration would show as different focal lengths for RGB which is why you get the rotation of colour. So again i tested this in comparison to my FSQ 85 using the same camera/filters etc. As it turns out this was also correct so RGB isn't being focused at the same focal length. I also compared different data sets from different nights eliminating temperature change etc etc and the results are fairly striking.

Takahashi FSQ 106

Red focal length - 530.99mm

Blue focal length - 530.48mm

Takahashi FSQ 85

Red focal length - 450.58mm

Blue focal length - 450.51mm

So between Red and Blue wavelengths my FSQ 85 has only a change of 0.07mm in focal length, so its apochromatic or focusing RGB at the same point. But my FSQ 106 shows a much larger change of 0.51mm so RGB isn't being focused at the same point hence i get the rotation of purple stars. Again this is a test carried out on different nights seeing conditions and temperatures and the results are consistent night to night.

Im feeling a complete lack of confidence! Lets not even talk about my focuser that was sent back to me adjusted for a DSLR weight!

Ken

Just peed off right now and hope they resolve this!

Note- thanks to those who have also been sending me messages of support and helping me workout the problem, thank goodness for SGL, cloudy nights etc etc.

[kirkster501]

Ken I sometimes get this with my FSQ85 and with my TEC140.  Note, sometimes, (i.e. intermittently).     I am sure it is something to do with light refracting in the atmosphere before it reaches the lens and the light being split already by the atmosphere.

Fringe killer in Photoshop sorts it straight away when it occurs.

[Ken82]

3 hours ago, kirkster501 said:

Ken I sometimes get this with my FSQ85 and with my TEC140.  Note, sometimes, (i.e. intermittently).     I am sure it is something to do with light refracting in the atmosphere before it reaches the lens and the light being split already by the atmosphere.

Fringe killer in Photoshop sorts it straight away when it occurs.

Yea good point kirkster501 and atmospheric diffraction is something I’ve looked at in great detail. It is also partly why I dismissed this problem in the first instance. 
 

Although every single image I’ve taken with this scope has shown the same purple stars. I’ve yet to take an image with this scope without the purple fringing and I’ve looking through months of data! 
 

Yet using the same camera/filters I was easily able to image without the colour fringing on my other fsq. 

Edit- also images of M106, M101, M63 and NGC 5907  I’ve taken at or next to zenith show the issue very clearly. This should limit atmospheric issues. 
 

The evidence is compelling really. 

Edited July 3, 2020 by Ken82

[ollypenrice]

You make a very good case.

The idea that the focal lengths per colour are sufficiently different to account for your colour assymetry in stars can be tested using Registar. Registar not only aligns images when co-registering them, it also resizes them. In theory this ought to rectify the problem as defined by your analysis. If you don't have Registar I'd be happy to co-register your RGB files and return them to you once done. Just Dropbox them to my regular email, which I'll send by PM.

Olly 

Edit, regarding occasionally difficult colour registration, it is very, very rare for me in either of the Tak FSQs or TEC140s I use regularly. When it does happen I use Registar to co-register the channels and that sorts it. For some reason I've found it best to make red the reference channel both in AA and Registar. Instinct told me it should be green but red seems the most reliable. I don't know why.

Edited July 3, 2020 by ollypenrice

[andrew s]

1 hour ago, ollypenrice said:

For some reason I've found it best to make red the reference channel both in AA and Registar. Instinct told me it should be green but red seems the most reliable. I don't know why.

Red will be less distorted by atmospheric variations both absolutely and differentially across the field (same reason the sky is blue) may e that's why it works best?

Regards Andrew 

[fireballxl5]

 

3 hours ago, ollypenrice said:

Registar not only aligns images when co-registering them, it also resizes them.

I run a dual Tak setup, RGB using one and Lum on the other. Occasionally combining in PI will result in colour fringed stars, usually red though.  When this occurs I transfer the individual, part-processed channels to MaxImDL for LRGB combination and this works without fail.  I've not found a detailed explanation of how MaxIm works under the hood but maybe similar to Registrar? 

[Ken82]

12 minutes ago, fireballxl5 said:

 

I run a dual Tak setup, RGB using one and Lum on the other. Occasionally combining in PI will result in colour fringed stars, usually red though.  When this occurs I transfer the individual, part-processed channels to MaxImDL for LRGB combination and this works without fail.  I've not found a detailed explanation of how MaxIm works under the hood but maybe similar to Registrar? 

Thanks for that, what your describing is likely atmospheric defraction as you have mentioned this happens occasionally. 

Although with my images as I've previously described I've never been able to take an image without the purple fringing. Every picture i've taken with this scope from months of data has a clear purple fringe. Even images taken at or near zenith have a clear problem M106, M101, M63 and NGC 5907. This should limit atmospheric issues. 

Also my FSQ 85 performed as expected with uniform star colour when tested so this all unequivocally points towards an achromatic aberration problem. 

[Skipper Billy]

I am no expert but it certainly doesn't look right to me.

This is what I get from the same scope - I am really not into pixel peeping and this is just a Jpeg but to my old eyes it looks like it should look and I dont see the same problem that you have Ken.

Good luck - I hope you get it resolved. 

 

[Ken82]

44 minutes ago, Skipper Billy said:

I am no expert but it certainly doesn't look right to me.

This is what I get from the same scope - I am really not into pixel peeping and this is just a Jpeg but to my old eyes it looks like it should look and I dont see the same problem that you have Ken.

Good luck - I hope you get it resolved. 

 

Yea thanks David, this is really what I was expecting. A nice uniform star colour even on the small stars in the image. Very similar to what my fsq 85 produces with this camera. 

[kirkster501]

7 hours ago, ollypenrice said:

You make a very good case.

The idea that the focal lengths per colour are sufficiently different to account for your colour assymetry in stars can be tested using Registar. Registar not only aligns images when co-registering them, it also resizes them. In theory this ought to rectify the problem as defined by your analysis. If you don't have Registar I'd be happy to co-register your RGB files and return them to you once done. Just Dropbox them to my regular email, which I'll send by PM.

Olly 

Edit, regarding occasionally difficult colour registration, it is very, very rare for me in either of the Tak FSQs or TEC140s I use regularly. When it does happen I use Registar to co-register the channels and that sorts it. For some reason I've found it best to make red the reference channel both in AA and Registar. Instinct told me it should be green but red seems the most reliable. I don't know why.

Olly,  so Registar can register and *resize* all the filters' stars to a nominated master's stars - luminance I'd guess???  That is a mightily powerful trick if so and I do not think PixInsight can do that in its Star Alignment process.

[ollypenrice]

16 minutes ago, kirkster501 said:

Olly,  so Registar can register and *resize* all the filters' stars to a nominated master's stars - luminance I'd guess???  That is a mightily powerful trick if so and I do not think PixInsight can do that in its Star Alignment process.

Registar certainly does it, yes. That's why it can co-register vastly different focal lengths for composite imaging. My poor brain can only cope with small doses of Pixinsight but I thought it could do the same, not that I've tried. What the Pixinsight team should study is just how easy it is in Registar. (Open two images. Click to register one to the other. Done.)

Which image you use as master is up to you. A NB imager wanting to combine Ha, OIII and SII while using Ha as luminance at the end would surely align onto Ha. As you suggest, the logical choice for LRGB might be L.

Olly

[tomato]

Astro Pixel Processor does a great job of combining images from different telescope/camera set ups, again with no fuss, might be worth trying, it has a free trial.

But  based on the evidence you have presented, I agree you have a case for a hardware rather than a software fix.

[Ken82]

7 minutes ago, tomato said:

Astro Pixel Processor does a great job of combining images from different telescope/camera set ups, again with no fuss, might be worth trying, it has a free trial.

But  based on the evidence you have presented, I agree you have a case for a hardware rather than a software fix.

Yea thanks for the responses all.

Regarding different software and correction/adjustment of focal lengths in RGB the point is if the telescope was truly apochromatic i wouldn't have to mess around with this. 

Differences between the RGB focus points in my  FSQ 85 (0.07mm) and my FSQ 106 (0.51mm) is a massive 629% increase. The FSQ 106 isnt focusing RGB at the same point. These focus points have also been measured on multiple nights over the course of months.  

[Andysea]

I had some correspondence with Ken regarding  this issue and my experience with the FSQ. I live in the US so I'm dealing with Takahashi America (TNR)

I responded to a post on CN and I am going to share it on here as well. So here it goes.

I have pretty much given up on my FSQ106. It was showing a very similar issue to yours (Ken's), coupled with a non flat field. 

It was not under warranty anymore but I sent it back to TNR at the beginning of September 2019.

In December, TNR said that they weren't able to flatten the field so they would send it back to Japan. Fast forward seven months and I still don't have the scope. Quite frankly I doubt that I will get it back either collimated or not.

I don't really care either way as I'm finally getting the AP GTX in October.

My suggestion is to have the scope replaced because if they send it to Takahashi Japan you might end up waiting several months to a year to get it back.

[Whistlin Bob]

Oh no! This is really surprising- the measured difference in focal length seems a slam dunk in demonstrating CA.  I really thought the trip to France would solve the problem. How can they possibly have come to the conclusion that it's all fine?

I would assume that if they're that confident it's a 'fine example' then they'll be happy to swap it for another one, as this one would still have a strong value as a proven and tested scope. 

[gorann]

On 03/07/2020 at 10:57, kirkster501 said:

Ken I sometimes get this with my FSQ85 and with my TEC140.  Note, sometimes, (i.e. intermittently).     I am sure it is something to do with light refracting in the atmosphere before it reaches the lens and the light being split already by the atmosphere.

Fringe killer in Photoshop sorts it straight away when it occurs.

I had not heard of the Fringe killer action. Is it this one?

https://blog.lorrifreedman.com/hdr-and-the-purple-fringe/

I downloaded it but when I run it the whole colour balance of the image is altered (but I do see less purple fringes). Is there a better action?

[kirkster501]

2 hours ago, gorann said:

I had not heard of the Fringe killer action. Is it this one?

https://blog.lorrifreedman.com/hdr-and-the-purple-fringe/

I downloaded it but when I run it the whole colour balance of the image is altered (but I do see less purple fringes). Is there a better action?

No it comes as standard in Camera Raw Goran.

[kirkster501]

16 hours ago, Ken82 said:

Yea thanks for the responses all.

Regarding different software and correction/adjustment of focal lengths in RGB the point is if the telescope was truly apochromatic i wouldn't have to mess around with this. 

Differences between the RGB focus points in my  FSQ 85 (0.07mm) and my FSQ 106 (0.51mm) is a massive 629% increase. The FSQ 106 isnt focusing RGB at the same point. These focus points have also been measured on multiple nights over the course of months.  

I'm really surprised ay this.  Do you have the reducer connected Ken or is the scope at native FL?  You may or may not be aware that with the reducer the distance to the chip is crucial that it matches that of the reducer.  To the mm.

I am sure you have checked but there is nothing, absolutely nothing that could be reflecting the light in the path in some way?  A lens hood, dew shield or something?  Back to basics Ken but don't overlook them mate.

Edited July 4, 2020 by kirkster501

[Stub Mandrel]

I'm cautious of making a definitive statement but everyone who has seen the stacked images seems to agree that the images in different colours are coming out at slightly different scales giving radial colour fringes, which suggests chromatic aberration.

Olly (I think) pointed  astigmatism making stars in the corners seem squished and Vlaiv noticed the star shapes varied between R G and B.

It's hard to believe with so many independent voices saying that the fault is with the optics, not the stacking. Also, I think I've never had misaligned stars more than once or twice and have always been able to improve them by getting rid of bad subs or changing stacking settings.

I know you can get rid of artefacts like this through processing, but they will rob the image of small high-contrast details.

More importantly, when you pay the premium on a multi-element apochromatic scope isn't getting rid of these effects exactly what you are paying for?

 

[kirkster501]

11 minutes ago, Stub Mandrel said:

More importantly, when you pay the premium on a multi-element apochromatic scope isn't getting rid of these effects exactly what you are paying for?

Most definitely - a five grand refractor!!!

[steppenwolf]

Registar is a /one trick pony. BUT it does that trick fantastically well.

2 hours ago, Stub Mandrel said:

More importantly, when you pay the premium on a multi-element apochromatic scope isn't getting rid of these effects exactly what you are paying for?

Absolutely!

[ollypenrice]

1 hour ago, steppenwolf said:

Registar is a /one trick pony. BUT it does that trick fantastically well.

Absolutely!

...and it does interesting variations of the trick! However you want the trick doing, with whatever data you have, it does it. 

Olly

[Ken82]

7 hours ago, kirkster501 said:

I'm really surprised ay this.  Do you have the reducer connected Ken or is the scope at native FL?  You may or may not be aware that with the reducer the distance to the chip is crucial that it matches that of the reducer.  To the mm.

I’m not using the reducer kirkster501, all images are at native F5.

I now have it on very good authority from a leading optics engineer that there is likely a spacing or radius error/s with the scope. 

[gorann]

7 hours ago, kirkster501 said:

No it comes as standard in Camera Raw Goran.

Aha, but Camera Raw is only for DSLR, isn't it?

[Adam J]

On 03/07/2020 at 21:05, Ken82 said:

 

Differences between the RGB focus points in my  FSQ 85 (0.07mm) and my FSQ 106 (0.51mm) is a massive 629% increase. The FSQ 106 isnt focusing RGB at the same point. These focus points have also been measured on multiple nights over the course of months.  

The total difference in focus position between all filters on my Esprit 100 is 0.04mm meaning I don't really have to use filter offsets. So that is horrendous. I assume you are using the same filters as used on the FSQ85? 

Maybe take a set of LRGB without refocusing just focus on L and then show us what you get in each channel. 

Adam.