Samyang 135 @f2: understanding potential chromatic aberration

[desmcm]

Hi

Just getting started with my Samyang 135 and noticed that some stars have a noticeable red fringe.

I've set my system up using the Astro Essential M48 adapter plate, and have set the back focus to be as close to 44mm as possible.

Imaging time has been very limited due to clouds. To date I've only tried f2 with a OSC ASI2600MC Pro  and  an L Pro filter, focussed with a Bahtinov mask.

Example image showing close up of stars with a red fringe.

Using PixInsight, I get for the median FWHM and eccentricity in pixels for the different channels typical values of:

R    2.037  0.6071 

G   1.927   0.5880

B    1.857   0.6085

My questions are whether there is an issue with the lens, is it a focus and or back focus issue, or is it a consequence of shooting at f2?

Thanks in advance, Des

 

[geeklee]

Hi Des, I certainly get the same behaviour as you describe (I've been using F2 as well) with the red channel looking a little bloated/out of focus.  The critical focus zone at F2 is so small and I've found that minute changes can show certain aberrations (like tilt) much worse.

If I get a chance to use the lens again this season, I plan to try getting focus with the Bahtinov mask like normal but then vary this slightly either way and check the subs then, see if I can make the red more acceptable. 

I find it more noticeable on certain images, especially some strong Ha targets where the red channel looks poor ... and that's where the Ha is!

Example RGB sub below at 1:1 (just UVIR cut filter)

It's the stacks which generally end up looking poor in the red channel - likely highlighting how critical it is to keep a close eye on focus and each sub before stacking.  Here's an example from a stack using the L-eNhance filter.  That red channel looks so soft.

At 1:2, it's more difficult to spot:

All of these are unprocessed, just Auto STF in PI.

My backfocus is around 45 and a bit with the focus mark in the "L". 

Edited March 19, 2021 by geeklee

[desmcm]

58 minutes ago, geeklee said:

Hi Des, I certainly get the same behaviour as you describe (I've been using F2 as well) with the red channel looking a little bloated/out of focus.  The critical focus zone at F2 is so small and I've found that minute changes can show certain aberrations (like tilt) much worse.

If I get a chance to use the lens again this season, I plan to try getting focus with the Bahtinov mask like normal but then vary this slightly either way and check the subs then, see if I can make the red more acceptable. 

I find it more noticeable on certain images, especially some strong Ha targets where the red channel looks poor ... and that's where the Ha is!

Example RGB sub below at 1:1 (just UVIR cut filter)

It's the stacks which generally end up looking poor in the red channel - likely highlighting how critical it is to keep a close eye on focus and each sub before stacking.  Here's an example from a stack using the L-eNhance filter.  That red channel looks so soft.

At 1:2, it's more difficult to spot:

All of these are unprocessed, just Auto STF in PI.

My backfocus is around 45 and a bit with the focus mark in the "L". 

Hi Lee,

Thanks. That is all very interesting and to the point.

I too have the feeling that stacking exacerbates the issue, but only based on the few hours I have managed on a couple of targets and some half-baked analysis in PI.

I intend to do some more testing clouds permitting.

Next on my list is to fit an electronic focusser - ZWO EAF - to take out the guess work and fumbling.

One quick question: how do you change the sampling in PI to go from 1:1 to 1:2. Do you use Resample at 50%?

Cheers, Des

 

 

 

 

[geeklee]

3 hours ago, desmcm said:

One quick question: how do you change the sampling in PI to go from 1:1 to 1:2. Do you use Resample at 50%?

For the images above, apologies, I should have been clearer - it's just "zoomed out" one level so you're looking at 50%.  If I am actually resampling, I use the IntegerResample Process (rightly or wrongly... probably wrongly!).

3 hours ago, desmcm said:

some half-baked analysis in PI.

You've just described my typical approach (although I do try) 😅

3 hours ago, desmcm said:

Next on my list is to fit an electronic focusser - ZWO EAF - to take out the guess work and fumbling.

It will be interesting how you get on with this.  My current rig doesn't really have a place for an EAF but I'm thinking about alternative options.  @Adreneline has some great images of his setup with an EAF in the "Imaging with Samyang 135" thread.  Can you run autofocus routines when you're so close to the end stop in one direction?  I guess it's tiny steps at F2 so maybe so.

7 hours ago, desmcm said:

with a OSC ASI2600MC Pro

For completeness, the images above were with an ASI533, smaller FOV but same pixel size (3.76 um) so 5.74"PP image scale at 1:1

Edited March 19, 2021 by geeklee

[Adreneline]

7 hours ago, desmcm said:

I've set my system up using the Astro Essential M48 adapter plate, and have set the back focus to be as close to 44mm as possible.

Hi Des,

Experience with the Samyang has shown me that the spacing, as noted by @geeklee is absolutely critical and even changes of a fraction of a millimetre can have a profound effect on the focus position relative to the 'L' on the lens and the shape and aberation of stars.

I have just set my Samyang up with my new ASI294MC having previously used it with the ASI1600 + 8 position EFW (with a mix of Baader (LRGB) and Astronomik filters (L3+NB)). On paper both the 294 and 1600 have their sensor set 6.5mm back from the front face of the camera housing. As mentioned by @geeklee I had my EAF mounted on the Samyang (works a treat) but it is now mounted on my RedCat so I am focussing the Samyang by hand using a BM.

This is the ultra light weight setup with the Samyang and the ASI294 - no filters anywhere. I have removed the EoS fitting and replaced it with the M42 fitting supplied by Astrojolo in Poland. The spacing between the end face of the lens and the front face of the camera is 38.37mm so the spacing from the end face of the lens is 38.37 + 6.5 = 44.87mm.

With that spacing the focus position is right in the middle of the 'L'.

I have found changes of as little as 0.2mm have a big effect on the final focus position, in fact you can see a ZWO 0.2mm spacer adjacent to the camera body.

Adding filters of any sort between the lens and the sensor will increase the spacing by nominally 1/3 of the thickness of the filter, so in the case of my Astronomik that is about 0.3mm. My experience is that a change of 0.3mm will move the focus position outside the foot of the 'L'.

I've seen it stated that this lens will "focus beyond infinity" - wherever that is!! I also read that the spacing should be adjusted so that the optinum focus is achieved with the marker in the foot of the 'L'.

Finally I obtained much better result using the Samyang with my ASI1600 + NB filters when I used a Hoya Pro UV cut filter on the front of the lens. The OIII images were much improved and I was achieving consistent FWHM results for Ha, SII and OIII - and far less 'blue bloat' when using the Baader RGB filters.

Hope this helps.

Adrian

 

 

[geeklee]

12 minutes ago, Adreneline said:

no filters anywhere

Nice write up of the new setup Adrian.  Do you still have the Hoya in front of everything with the 294 setup or is the window of the 294 UVIR cut alongside AR?

[Adreneline]

9 minutes ago, geeklee said:

Do you still have the Hoya in front of everything with the 294 setup or is the window of the 294 UVIR cut alongside AR?

Hi Lee. I just left it in place. I've never noticed any sort of reflection issues with the filter so I figured it would do no harm to leave it.

The weather has not been great since combining the 294 and Samyang but I did manage 60 x 180s of the M81/M82 region in the hope of getting some IFN. I'm not sure whether I've got it or not! My initial processing of the osc images has not revealed anything, despite following advice from @AbsolutelyN ; what I can see may be just wishful thinking. I think my processing is letting me down

Adrian

[geeklee]

10 minutes ago, Adreneline said:

I did manage 60 x 180s of the M81/M82 region in the hope of getting some IFN. I'm not sure whether I've got it or not!

I grabbed this area back in December and got about the same but 180 X 60s and F2. There seemed to be some there when the data was worked hard. Unless I'm mistaken and it's dust.

https://astrob.in/wudss6/B/

Edited March 19, 2021 by geeklee

[Adreneline]

2 minutes ago, geeklee said:

There was some there when the data was worked hard.

That is off the scale impressive compared with mine! Would you like my .fit master to play with?  🤞

Adrian

P.S. Apologies Des @desmcm for sidetracking you thread.

[geeklee]

1 hour ago, Adreneline said:

Would you like my .fit master to play with?  🤞

I'm intrigued, go for it!

[desmcm]

Dear Adrian and Lee,

Thanks for sharing your experiences. There's a lot of helpful info to guide next steps.

Tuning the back focus is obviously the thing to concentrate on accepting at the outset that this will  not be a joyful process. However, my approach will be to do this iteratively while imaging, as I still get a deep satisfaction in obtaining any kind of image, even one with bloaty red stars as in the attached!

One very useful hint that might nudge me in the right direction is the answer to the following:

If the optimum focus as measured by a BM is found to the left of the L (as looking skywards along the lens) does this mean that I need to increase or decrease the distance between the lens and the sensor? Currently at nominal 44mm, my focus is to the left of the L.

Regarding the EAF, I like the look of this implementation:

https://www.thingiverse.com/thing:4761672

or the one now available from deepskydad.com. I too am concerned about running out of adjustment if the focus is close to the hard stop.

Best, Des

 

 

 

 

 

 

 

[geeklee]

12 hours ago, desmcm said:

not be a joyful process

quite 😅 

11 hours ago, desmcm said:

If the optimum focus as measured by a BM is found to the left of the L (as looking skywards along the lens) does this mean that I need to increase or decrease the distance between the lens and the sensor? Currently at nominal 44mm, my focus is to the left of the L.

I believe - from my own testing - you would need to add a little spacing.  If you're the opposite and at the end stop "past" infinity then you'd need to take away.  Sure someone will chime in if this is incorrect ( @Adreneline )  When I say a little spacing, it will likely be less than 1mm.

11 hours ago, desmcm said:

Regarding the EAF, I like the look of this implementation:

https://www.thingiverse.com/thing:4761672

Me too!

Lovely image above, what a vast FOV that camera gives with the Samyang.

[Adreneline]

22 minutes ago, geeklee said:

you would need to add a little spacing

I agree! - I think

As Lee says whatever you add/subtract it needs to be in very small amounts. I am sure it does pay off to get it "bang on" as star shapes into the corners are much better.

Adrian

 

[desmcm]

29 minutes ago, geeklee said:

quite 😅 

I believe - from my own testing - you would need to add a little spacing.  If you're the opposite and at the end stop "past" infinity then you'd need to take away.  Sure someone will chime in if this is incorrect ( @Adreneline )  When I say a little spacing, it will likely be less than 1mm.

Me too!

Lovely image above, what a vast FOV that camera gives with the Samyang.

Thanks.

Forgot the details. 30 x 120s lights, gain 100 -10C @f2 ASI2600MC Pro, L Pro filter. 50 x flats. 50 x Dark flats. Heavily processed in PixInsight, tricky because of strong light pollution, including heavy handed clipping. 

[alacant]

18 hours ago, desmcm said:

focus as measured by a BM

Ssme issue. Different lens.

The BM focus is for the dominant g and b, leaving the red fuzzy. The best we can do with our Zeiss is a compromise; none of the colours is at its optimum focus.  Compromised Infinity for RGB is the BM position with a tiny rotation of the focus barrel away from infinity toward closer distances. A similar trick can be used with refractors where usually it's the blue which is compromised.

Here is the compromised infinity on the Zeiss and a similar fov to the above using the technique:

HTH

 

Edited March 20, 2021 by alacant