Strange behavior of Sigma 40mm F/1.4 Art lens.

[Vroobel]

Hi,

Around two months ago I bought a brand new Sigma 40mm F/1.4 Art and tried it with a full frame Canon 6D. After some test, I found that the F/1.4 gives coma even if it's praised as the best wide-field lens ever, especially for AP. I stopped at F/1.8 which is still fast enough in comparison to my F/4.6, F/5 and F/5.6 scopes. Here are views of the corners and edges of a stacked image respectively without any correction and after applying a BlurXterminator (F/1.8):

 

 

I told to myself: ok, it's acceptable, I should be able to use the F/1.4 again when I buy a ZWO EF filter drawer and move to an ASI2600MC-Pro APS-C camera.

So I bought the adaptor with the filter drawer and screwed all components together. I tried the F/1.4, as the Sigma doesn't have an iris ring and the aperture depends on a body setting. An M42 thread didn't allow me to obtain a horizontal or vertical position of a sensor which I prefer, so I used some thin Baader spacers, but here is what I got:

 

 

I shouldn't cheat myself, it doesn't look good, after all, it's the APS-C and should be almost perfect.

I replaced the plastic spacers with two rings cut out from an 80g/m2 xero paper - they are thinner. But it changed nothing. I got an idea that maybe I should screw the aperture down to the F/1.8, but how to do it? Based on a following CN thread I attached the 6D body, set a BULB mode, pressed the shutter release button and... detached the body - the iris remains in the F/1.8 position. That's a great trick! Well, here is what the F/1.8 with APS-C single sub looks like (sorry for the clouds, I didn't invite them):

https://www.cloudynights.com/topic/837147-sigma-art-40mm-14-canon-mount-no-aperture-ring/

 

 

Nothing better. 

Looking at this rule...

 

 

... I would say that the camera is too close to the lens and it usually makes sense, but not in a case of my Askar FMA230, which behaves just the opposite and I had to reduce the 55mm back focus to 52.5mm (and I'm not the only one).

 

Has anyone had any experience or idea to say what's going on here? I'm completely confused and I don't know what to test next. I'm limited by clouds, so I don't want to make blind tests. I would appreciate any ideas.

 

Edited April 17 by Vroobel

[Nakedgun]

~

Wrong forum?

 

 

.

[Vroobel]

So which is the proper one? I tried to find it. 

[michael8554]

Hi Vroobel

Backfocus is only of concern if you have a FF or FR.

All a lens needs is to be the correct Flange Distance from the sensor, in order to reach infinity focus.

The Canon Flange Distance is 44mm.

So you raise an interesting thought that reducing the lens to sensor distance has similar effects to altering the Backfocus from a FF or FR.

Have you started at 44mm Spacing ?

Michael

[Mr Spock]

Moved to Imaging Discussion so our imagers can take a look  

[Vroobel]

3 hours ago, michael8554 said:

Hi Vroobel

Backfocus is only of concern if you have a FF or FR.

All a lens needs is to be the correct Flange Distance from the sensor, in order to reach infinity focus.

The Canon Flange Distance is 44mm.

So you raise an interesting thought that reducing the lens to sensor distance has similar effects to altering the Backfocus from a FF or FR.

Have you started at 44mm Spacing ?

Michael

Hi Michael, 

The dedicated ZWO Canon EF filter drawer has 26.5mm optical path length which together with the 2600MC's 17.5mm back focus gives 44mm.

After screwing the adaptor to the camera it stops in a random position. The lens is about 1.3kg heavy, so I use dedicated for the Sigma lens ring holding it in the position like after attaching to Canon body. I operate only with horizontal or vertical positions of the sensor, so it's very important to me to keep it. That's why I finally used two paper spacers which shouldn't cause so bad results.

My previous experience with M42 135mm vintage lens says that changing the distance between the lens and the camera affects only a position, where you reach focus at infinity. E.g. for a typical 20mm long ZWO filter drawer v.2 with 1mm long EF/M42 adaptor it appears at around 3m. 

I expected that it will also happen to the Sigma and indeed, the position of focus at infinity looks like in the picture below. 

 

 

If moving the focus position is accompanied by the significant distortion then the Sigma lens + ZWO EF filter drawer + 2600MC is not acceptable for me and I have no idea what to do next. 

Edited April 17 by Vroobel

[Vroobel]

BTW, here is shown where the focus at infinity is located if I work with the Canon 6D. It's not too far, to be honest. 

 

 

[Vroobel]

Sorry for multiple posts. 

I'll try this advice tonight (thanks to @Elp) :

 

 

https://optcorp.com/blogs/deep-sky-imaging/how-to-set-the-correct-back-focus

[Elp]

I just tried one of the best, sharp daytime lenses, though a more affordable FE lenses for my APSC Sony, it's an FE lens so built for full frame. Does it coma... It sure does! It's a similar star elongation issue too, you'd think with a crop sensor it wouldn't show it. I've found most lenses are quite poor for astro, especially if you want a sharp flat field.

Regarding the odd BS issue between the Sigma and Askar, I'd presume most companies only do batch QA, so you will find manufacturing variance between items even if they have been made to a BS specification, it'd be within their manufacturing and QA tolerance spec.

Edited April 18 by Elp

[Vroobel]

28 minutes ago, Elp said:

I just tried one of the best, sharp daytime lenses, though a more affordable FE lenses for my APSC Sony, it's an FE lens so built for full frame. Does it coma... It sure does! It's a similar star elongation issue too, you'd think with a crop sensor it wouldn't show it. I've found most lenses are quite poor for astro, especially if you want a sharp flat field.

Regarding the odd BS issue between the Sigma and Askar, I'd presume most companies only do batch QA, so you will find manufacturing variance between items even if they have been made to a BS specification, it'd be within their manufacturing and QA tolerance spec.

 

Of course, you are right, but why my Askar denies the following rule?

 

 

I think it's because it's closer to a lens than to an astro refractor.  Focusing by rotating a ring (OK, WO Cats also have it) while its objective moves in and out.

Anyway, after many attempts I applied around 1.2mm spacer in total and I still have a small coma interchangeably with a radially elongated star in corners. I think it's a point of compromise. They are visible after stacking while an affected by LP background is still there. But the goal is to remove the background, then the distortion isn't visible. I made 30x30s F/1.8 for IDAS LPS-P2 and L-eXtreme filters, now I repeat the same for F/1.4 to see any difference.

 

In the meantime, my main setup captures a Lion Nebula in SHO...

[michael8554]

1.  "why my Askar denies the following rule?"

That rule is for FF and FR, not necessarily for lenses.

You can test whether altering the Backfocus does actually make a difference to the coma, not just the infinity focus.

I doubt it will.

2.   30 sec exposures are showing some central elongation, try 10 secs and stretch the jpegs, they are good enough.

Michael

[Vroobel]

I'm not sure if I know the meaning of FF and FR.

Askar isn't a lens, but it doesn't behave like a typical refractor.

The AberrationInspector shows the 9 areas zoomed. The central elongation is nothing in comparison to the coma in corners, easily correctable by the BX, while the coma is not entirely.

I did the tests. Too thick spacer causes radial, elongated stars in corners, like star trails. Too thin spacer - still coma, but smaller. The optimal distance of 1.2mm is described in my previous post.

 

[Elp]

Unless a manufacturer 100pc checks every part and works to a strict manufacturing tolerance and rejects anything not "perfect" you'll never get any mass product made, it also won't be cheap.

You'll get variance from like for like units, still made within a manufacturers acceptable/repeatable achieving tolerance, but not 100pc at nominal spec with +/- zero variance.

[michael8554]

1.  FF = Field Flattener,  FR = Focal Reducer.

These require the correct Backfocus, or coma will result .

2. "Askar isn't a lens, but it doesn't behave like a typical refractor."

Whatever, interesting that altering Flange Distance while still able to infinity focus does change the coma.

3.  So "Too thin spacer - still coma, but smaller".

If going thinner improves the coma, then try an even thinner spacer, or no spacer ?

Michael

[Vroobel]

The 44mm results in a big coma.

Increasing the distance makes the coma smaller.

When the spacer is 1.2mm thick, some corners have small coma while others have radial elongated stars.

The spacer thicker than the 1.2mm results with elongated stars only. 

Edited April 18 by Vroobel

[R26 oldtimer]

Different flange distance in simple optical design lenses should make a difference only in infinity focus position. But with complex design lenses as the Sigma it will introduce aberrations. So my guess is that you need some spacers / washers so your focus (when focusing at stars) mark is similar as when you had the lens on the DSLR.

[Vroobel]

Right, so here is how the final version looks like with 1.2mm spacer. Before and after BlurXterminator, just 'Correct only' selected.

I nearly forgot: it's F/1.4.

 

 

I'm not going to kill myself, I think it's much better than an average 40mm lens.

 

 

 

Edited April 18 by Vroobel

[michael8554]

I guess then that we need to hear from other Sigma 40mm F/1.4 Ar users if they get coma like that.

Such as:

https://www.cloudynights.com/topic/824952-sigma-40mm-a-f14-star-test/

Looks like you have a lens at the far end of Sigma's QC range.

Michael

[Vroobel]

On 19/04/2024 at 18:10, michael8554 said:

I guess then that we need to hear from other Sigma 40mm F/1.4 Ar users if they get coma like that.

Such as:

https://www.cloudynights.com/topic/824952-sigma-40mm-a-f14-star-test/

Looks like you have a lens at the far end of Sigma's QC range.

Michael

 

Thank you, @michael8554, you inspired me to a more intense analysis. 

I realised that I used filters from the beginning which falsified the flange distance by mentioned 1/3 of their thickness (L-Pro Canon FF, 2" IDAS LPS-Pd and 2" L-eXtreme). I couldn't sleep well, so decided to remove the filters and any spacers and make another comparison with the picture from the CN thread (Canon 6D, F/1.4). The results are as follows.

 

 

A starfield from the CN thread:

 

 

 

My own Canon 6D at F/1.4 - that means it's the same setup:

 

 

 

ASI 2600MC Pro with ZWO Canon EF/M42 filter drawer, exactly 26.5mm of an optical path + 17.5mm of the APS-C camera back focus = 44mm of the flange distance:

 

 

 

I couldn't capture so many stars, but I would risk and say that my copy of the Sigma 40mm F/1.4 Art looks slightly better than the CN user's one.  It's also possible that the difference is caused by an astro-modiffication performed on my Canon 6D.

The question remains why the stars captured by APS-C are so distorted, although they should look better than with the full frame sensor. The goal is to use the ASI2600MC Pro camera with filters: IDAS LPS-P2 and Optolon L-eXtreme which make the distortion bigger. Enlarging the flange distance with spacers caused the coma to turn into radial elongation in some corners which seems to be a compromise. 

Is the ZWO EF/M42 filter drawer designed wrong? Or the camera back focus isn't equal to 17.5mm? I would appreciate any constructive conclusion.

 

[Elp]

If you're using it at F1.4 I think you're going to struggle immensely. All my fast lenses need to be stopped down to F3 or F4 to get anywhere near acceptable stars at the edges, maybe even F5.6. The apsc sensor makes this worse as it's a large sensor, I could probably get away with it with my 183 sensor. There could also be a contributing issue of not using fast filters.

[Vroobel]

1 minute ago, Elp said:

If you're using it at F1.4 I think you're going to struggle immensely. All my fast lenses need to be stopped down to F3 or F4 to get anywhere near acceptable stars at the edges, maybe even F5.6. The apsc sensor makes this worse as it's a large sensor, I could probably get away with it with my 183 sensor. There could also be a contributing issue of not using fast filters.

I'm sorry, I don't understand. My full frame Canon 6d gives nearly pinpoint stars at F/1.4 while the smaller APS-C gives the coma. Shouldn't it be the opposite?

BTW, I bought the 40mm F/1.4 Art with an assumption to stop it down to F/1.8 or F/2.0, if needed, but rather not to F/5.6.  

[Vroobel]

10 minutes ago, Elp said:

 There could also be a contributing issue of not using fast filters.

 

I have to admit that I didn't think about it. But today I didn't use any filter.

[Elp]

Yes it should be the opposite. I'm surprised it works pin point to the edge of a FF sensor, most reviews of lenses where they state they're good for astro and you look at the edges they look awful, from experience all the lenses I've used distort at the edges even on FF lenses, though I've never used an Art.

How have you measured your distance?

[Vroobel]

Originally, I wanted to buy a 50mm lens. When I made some research I found that the 50mm lenses are as you wrote above. Then my attention focused on a Tamron 35mm and the Sigma 40mm Art. I had to read a lot to find that the Sigma is outstanding even at F/1.4, better than the Tamron. It's worth the price.  

I measured the optical path of the ZWO EF/M42 filter drawer using a vernier caliper.

Edited Wednesday at 22:26 by Vroobel

[Elp]

I would note the ZWO Ef adaptor likely isn't mounted square so will cause an element of tilt, I get it on all my SY135s, the weight of the lens doesn't help. A lot of people have to shim to get it right.

Make sure you're measuring with a good quality Vernier caliper too either a manual one or digital, measure each component separately, then again when assembled.

Edited Wednesday at 22:28 by Elp