Samyang 135mm f/2.0 – Sensor Tilt or Another Bad Copy?

I had two of this lens, one worse than the other and I returned them both. I didn't keep the images with the coma aberrations, but they were horrific compared to what you have. Nothing you can't easily fix with BlurExterminator, or similar, from what I can see.

There is probably some tilt. But all in all, this is about par for the course for this lens. You have to keep in mind the aperture, price point and that it's a photographic lens and not a telescope.

Thank you all for your insights and the pep talk :-)
The chromatic aberration didn't really bother me. It's more about the star shapes, which, even after applying BXT, were not satisfying, as they left holes in the image.
I know that the following image (crop) is not refined at all, since it might be stretched too much—it was just meant for testing purposes showcasing the result after BXT:



But of course, you most of the time get what you pay for - less than 400 bucks at F2 might be complaining on a high level, though.
So, I will try to mitigate things by experimenting with the tilt. Unfortunately, and even more so after what I’ve read here, there’s a risk that the next copy could be far worse than this one, and then when does it stop?…

Do you use the m48 samyang adapter.   Eg from FLO.   Folks seem the swear by them to improve tilt.

The other thing to check is the backspacing.  There are good treads on this elsewhere on the web but the trick seems to be to make sure that the image plate solve showing a FL of 130mm which is the actual FL of the lens.

Does that adapter  come with tilt-adjustmet screws?

I opted for this solution to also keep the possibility of easily using it on my Canon 6D and kept the EOS adapter ring on the lens. While it may not be the best in terms of flexure, it gives me a very sturdy impression though:

Michael E.:

Do you use the m48 samyang adapter.   Eg from FLO.   Folks seem the swear by them to improve tilt.

The other thing to check is the backspacing.  There are good treads on this elsewhere on the web but the trick seems to be to make sure that the image plate solve showing a FL of 130mm which is the actual FL of the lens.

Does that adapter  come with tilt-adjustmet screws?

I opted for this solution to also keep the possibility of easily using it on my Canon 6D and kept the EOS adapter ring on the lens. While it may not be the best in terms of flexure, it gives me a very sturdy impression though:

Hello Michael.

Problem is not the camera's or sensor's tilt. Is that adapter of yours at speeds bellow f/3. The backfocus need to be super precise, even the switch between filter is enough to change the backfocus and systems to attach the camera such as those, isn't sufficient to keep the optical system free of tilting issues. I'm surprised that you have that tilting ASTAP display too close like above with that system.
The solution that I've found so far was screwing from the back of the lens (using the FLO M48 adapter) up to the camera threads with all of the acessories inclusing a tilt plate to fine tunning things.
Systems fast as that, if you use the ZWO adapter to T2/M42, the simple fact of using the lens EF mount, that little clip with spring load to stop the lens when attached, is enough to put tilt in the system when he's attached to the adapter, by pressure. 

Adjusting backfocus and tilt in systems so fast like that, are one of the biggest challenges for guys like me that have hyperstars, RASA's, and other fast systems. We spend a lot of time under the stars just adjusting the optical trains to near perfection. We spend more time adjusting the optical systems, than shooting actually.

Regards,

Cesar

Not necessarely Michael. The system is even more sensitive when you go low on the f/ number, to the point where sometimes you don't know if you're having tilt or backfocus problem, or even both. The f/ number is the key player. The tilt even could be from the camera sensor, which is a possibility since they are usually mass produced.

I suggest a couple of things. Change the bayonet to the FLO M48 plate insted of using the Canon EF bayonet. Insted of using an attatchable syste, like yours, use screwed rings, a filter drawer, a plate tilt (but from a good manufacture, not one like the ZWO) and using some 0.25mm, 0.5mm and even 1mm spacer rings, and take your time adjusting everything. With the FLO M48 plate you'll need minimum 44mm of back spacing. But that number is very relative, it could be 44.25mm, or 44.5mm, etc. and you need to count also with the filter thickness, which will pull the back spacing even further. Usually the index refraction ir about 1/3 of the filter thickness. For example: a filter of 3mm thickness, will pull further the backspacing about 1mm. A 2mm filter, will be 0.60mm of pulling the backspacing further, etc.. 

Suggestions for a tilt plate: Gerd Neumman like these: https://www.astroshop.eu/tilting-adaptors/gerd-neumann-jr-camera-tilting-unit-ctu-xt-m48/p,51573
Or the Baader one: https://www.astroshop.eu/filter-wheels-filter-sliders/baader-ufc-tilter/p,75225

Why one of those? They are super precise with the adjustment's scews, and the are adjustable from the side, not in the front or the back. This will help you a lot. The Gerd Neumman is my choice since they are 11.3mm of thickness in the optical path.

Suggesting for the spacer rings: https://www.astroshop.eu/extension-tubes/astoptics-extension-tube-m48-fine-tuning-kit-0-5mm-0-8mm-1mm/p,57847  - I bought for me 4 of 0.5mm and 4 of 1mm of those rings. 
Some ZWO cameras usually sends with them 0.25mm T2/M42 back plastic of those rings as well 0.5mm. This helps fine tunning things also.

That's the FLO M48 Plate: https://www.firstlightoptics.com/adapters/astro-essentials-samyang-lens-to-m48-adapter.html

Those are the spacer rings that I mentioned above: https://www.astroshop.eu/extension-tubes/explore-scientific-extension-tube-m48-set/p,69009 - But you can find in Amazon more cheaper ones like this: https://www.amazon.es/dp/B0C3ZJWG3B/?coliid=I2841AW6ZNQ03M&colid=FFGC41KBZX3C&psc=1&ref_=list_c_wl_lv_ov_lig_dp_it

For your QHY camera the sensor backfocus is 17.5mm, so from those 44mm minimum, you have 26.5mm, imagine that you use the Gerd Neumman tilt plate (11.3mm) gives you 15.2mm for filling with those spacer rings above or a filter drawer that have the maximum of 15mm thickness. I don't know any that shorten. I use the Altair Astro filter drawers, which are 17mm, but in my case I use ZWO cameras, which I can manage more easy the backspacing regarding the QHY or even playerOne, Touptek, etc.

Hope this information helps you achive your results. Keep us posted.

Regards,

Cesar

Do you use the m48 samyang adapter.   Eg from FLO.   Folks seem the swear by them to improve tilt.

The other thing to check is the backspacing.  There are good treads on this elsewhere on the web but the trick seems to be to make sure that the image plate solve showing a FL of 130mm which is the actual FL of the lens. (Not 135mm)

I was doing exactlt this but could not fathom why I was a full 5mm off. As my stars weren't too bad I just left it. Delighted to know that the "real FL" should be 130mm.

I assume I'm right in thinking that any gap needs to be reversed? e.g. if the platesolved focal length under-reads (say, 129.5mm), then I need to reduce spacing?

Hello John,

I didn’t care very much regarding the FL in the platesolve, since I’ve worked for the shape and quality of the stars, instead of the 135mm of the image.
I’m not sure why it shows like that (the 130mm) instead of the 135mm, but I imagine it’s related with the crop factor, since it’s a full frame lens. But like I said, I’m not sure about this explanation.

Best regard,

Cesar

I'd be happy with that. I really think this corner pixel peeping goes a bit far sometimes. I would be more concerned with how the image presents as a whole and in your case, it presents very well.

But at 2.8 the aperature is 48mm. Compared to being wide open, you are going to need to double your exposure time for a similar image. It might be worth considering taking a few frames to do a star overlay at 2.8 but the bulk of your imaging session at 2.0 would be the best bang for your buck.

Thanks for your advice. You’re correct that at F/2.8, it would require doubling the total integration time to be comparable to F/2.0.
If I understand you correctly, you're suggesting using the stars from the F/2.8 integration and the starless image from F/2.0. The point is that BXT has no problem correcting the star shapes per se, but rather struggles with filling the gaps in the starless image. So, I’m afraid this approach won’t help much, especially when there is a lot of nebula present. 
There is a cropped image shot at F/2.2 illustrating this effect further up. Maybe someone has an idea on how to mitigate this effect other than just dialing the stretch down? Usually my worflow starts with SPCC followed by BXT and SXT (all in lienar state)

I don't know why people stop down the lens rather than using it at native aperture i mean native f, with narrowband i understand about the shifting, but for RGB is that also a case?

Hi Tareq, this is done to increase the f/ (f/ = focal lenght / diameter), thus to reduce/control the size of optical aberrations in the corners.