I've noticed on my 17mm TSE and a metabones that it can't reach infinity focus with shifted corners. The field curvature is just too great. So I increased the flange distance by 0.3mm and they come into focus, but now I have huge field curvature problem with close in focusing. Tomorrow I'm going to test 0.1mm and 0.2mm, because with 0.3mm infinity focus is reached even before infinity mark. (I'm hoping with a more precise flange, field curvature will reduce while the extreme corners at infinity will still be in range)
These TSE lenses seem to be affected by changes of even 0.1mm in a big way. Maybe the Sony sensor glass is affecting the extreme corners as well.
As a note, when I'm talking about sharp shifted corners, I'm talking at f8 or f1. Below that the extreme corners aren't sharp on high MP cameras.
It seems to me that, on balance, the A7r2 has to be at least partly at fault because the A7r does not show the problem with or without the adapters. The different adapters may contribute too but they cannot be the whole problem. Nor can floating focus lens elements, as the lenses should work with or without an adapter so long as the distance from the sensor is the same.
Are the Sony lens mounts less robust than the Canon and others ? Could they be more easily distorted by hanging heavy Sony lenses from the camera instead of hanging the camera from the heavy lens ?
Maybe there is a slight error in the sensor placement that puts it too close to allowed tolerances.
Just so you know, I have never used the Sony gear. I mostly use Nikon gear now and I intended to get an A7r2 to use with my TS-E and other Canon lenses so that I could avoid buying a 1DX2 at twice the price. Now it's definitely in my too hard basket and I'm quite disappointed, as I imagine most of you are too
The more I hear, the more I suspect that the addition to the register between the outer surface of the flange and the sensor face is the culprit and that is down to sony having changed it. This fits with the inability to reach infinity focus encountered by some.
No one seems to have had the problems on the a7r but only on the a7r2. When I encountered the problem I wrote to kolari to see if the sensor glass had changed thickness between the a7r and the a7rii and they responded that it had not changed and that both the a7r and the a7r2 use the same sensor glass.
Perhaps someone should adjust the flange distance though I'm not sure how that would be done. Maybe a simple replacement of the r2 flange with the r flange or maybe milling down the r2 flange?
I guess that milling the flange on the adaptor might also help.
Has anyone tried the mc-11? Sigma might have picked up on any change in focal register.
I'm really excited that I've cracked this issue. I guess I'll be keeping my TSEs! Yay!
0.4mm increase in flange distance has fixed the issue for my 24mm TSE, metabones 3 and A7R2. 0.3mm works well for the 17mm TSE, but I'm also going to try with 0.2mm.
They still need to be stopped down to f11 to reach good sharpness, but I think I've got performance like you expect on a high MP Canon body.
TakenWild wrote:
I'm really excited that I've cracked this issue. I guess I'll be keeping my TSEs! Yay!
0.4mm increase in flange distance has fixed the issue for my 24mm TSE, metabones 3 and A7R2. 0.3mm works well for the 17mm TSE, but I'm also going to try with 0.2mm.
They still need to be stopped down to f11 to reach good sharpness, but I think I've got performance like you expect on a high MP Canon body.
Can you post a photo of the shimmed adaptor? I keep reading about shimmed adaptors and can't visualise where/how they are being done. I've sold my TSE since starting this thread but im still interested im the possibility of a fix
You have to take off the rear mounting plate of the adapter and put thin pieces of metal there. Then you attach the mounting plate on top of those slices.
Correction I used 0.05mm stainless steel foil. So I used 0.2mm (4 layers) on the 24mm TSE and 0.15mm on the 17mm TSE (I have a metabones 3 for each).
I'm currently testing putting 2 layers on th rear on the metabones and 2 layers on the front mount as I'm having problems with tightness when mounting the lens.
I might even increase the layers to 5 to see what happens. It is amazing that such tiny distances have such a huge impact on the focus. If 0.05mm makes a difference, then there is no hope in metabones getting it exactly right for each lens.
You have to take off the rear mounting plate of the adapter and put thin pieces of metal there. Then you attach the mounting plate on top of those slices.
I don't recommend opening up the rear of the metabones. It's a nightmare to get the shim in there. I got one layer in there which is all I needed. 3 layers on the front is the maximum it can take and keep getting the lens of and on, I needed 4 layers in total for the 24mm TSE.
TakenWild wrote:
I might even increase the layers to 5 to see what happens. It is amazing that such tiny distances have such a huge impact on the focus. If 0.05mm makes a difference, then there is no hope in metabones getting it exactly right for each lens.
If canon did not need to change anything in the mount for different lenses why would metabones have to? All Metabones should have to do is make the adapter the correct length to make up for the difference between the register of the sony and the canon. On the a7r they did this. What worked on the a7r does not work on the a7r2. Conclusion: sony changed something on a7r2. Conclusion 2: whatever they changed should have the same impact on all lenses but may well be immaterial on lenses with slop at the long end of the focus range which is probably most lenses in this day and age where they are designed to rack past infinity then correct focus by racking back.
I'm surprised that shimming is helping because I had assumed that the register was slightly longer on the a7r2 but apparently not. I think sony should recognize this as a deficiency if it turns out to be the case and offer a remedy.
Well I guess there's no point to test my A7r as this seems to be a A7r2 issue. I don't tilt, i just shift but perhaps this means I should stick with the TSE lenses on the 5dsR.
masimo wrote:
If canon did not need to change anything in the mount for different lenses why would metabones have to? All Metabones should have to do is make the adapter the correct length to make up for the difference between the register of the sony and the canon. On the a7r they did this. What worked on the a7r does not work on the a7r2. Conclusion: sony changed something on a7r2. Conclusion 2: whatever they changed should have the same impact on all lenses but may well be immaterial on lenses with slop at the long end of the focus range which is probably most lenses in this day and age where they are designed to rack past infinity then correct focus by racking back.
I'm surprised that shimming is helping because I had assumed that the register was slightly longer on the a7r2 but apparently not. I think sony should recognize this as a deficiency if it turns out to be the case and offer a remedy.
I don't think this is a Sony or Canon problem. It's the adapters, Metabones adapters do not seem to have tolerances in the 0.05-0.1mm range. These tolerances really matter for the Canon 24mm TSE. My bet is Metabones tolerance rounds down to the nearest mm rather than up. Rounding up will cause more problems than down with most lenses.
As for the differences people are reporting between the A7R and A7R2 I don't know, I've got both those cameras and I say well maybe there is a difference, but I've hardly noticed it. They were both bad and they have vastly improved with my metabones shimming. I'm kinda of doubtful of people's claims about how good the 24mm TSE is with the A7R. I don't know, I can only go by what I have and my tests.
Here are unshifted, unsharpened crops from my TSE24II on my A7R @ f8 in all 4 extreme corners. This is at least 300 feet from my balcony. I get the same results on either my MBIII or MBIV adapter.
I could nitpick but the corner sharpness is similar and in my opinion pretty darn good. My Canon 16-35 f4 IS was sharper in the corners than my FE 16-35 and this is even sharper than the zoom.
Chris, my issues I've been trying to correct is for fully shifted corners. That is where I've been having problems with metabones on both my A7R and A7R2.
TakenWild wrote:
I don't think this is a Sony or Canon problem. It's the adapters, Metabones adapters do not seem to have tolerances in the 0.05-0.1mm range. These tolerances really matter for the Canon 24mm TSE. My bet is Metabones tolerance rounds down to the nearest mm rather than up. Rounding up will cause more problems than down with most lenses.
As for the differences people are reporting between the A7R and A7R2 I don't know, I've got both those cameras and I say well maybe there is a difference, but I've hardly noticed it. They were both bad and they have vastly improved with my metabones shimming. I'm kinda of doubtful of people's claims about how good the 24mm TSE is with the A7R. I don't know, I can only go by what I have and my tests. ...Show more →
Those are good points but...
In my situation the test was done with the a7r and a7r2 along with a Metabones version 3 and my tse17. The only variable was the two sony cameras. The same adapter worked on one with the same lens and didn't on the other. That leads me to a difference in the camera. I also used a leica 560 telyt on its own tripod mount to verify and found the same problem.
I plan to measure the register on the two cameras later today and will post my results.
Just to refresh: my test was done by focusing the lenses on the a7r with the adapter, then removing the sony an adapter and replacing it with a sigma sd1 body with canon mount. With the a7r the register between the two systems, the a7r with metabones III and the sigma sd1 was exactly the same and the lens remained in focus on the sigma. When I repeated with the a7r2 the lens was no longer in focus on the sigma indicating a difference in register. Multiple tests with the 17mm tse and a leica 560 telyt had the same results.
TakenWild wrote:
I don't think this is a Sony or Canon problem. It's the adapters, Metabones adapters do not seem to have tolerances in the 0.05-0.1mm range. These tolerances really matter for the Canon 24mm TSE. My bet is Metabones tolerance rounds down to the nearest mm rather than up. Rounding up will cause more problems than down with most lenses.
As for the differences people are reporting between the A7R and A7R2 I don't know, I've got both those cameras and I say well maybe there is a difference, but I've hardly noticed it. They were both bad and they have vastly improved with my metabones shimming. I'm kinda of doubtful of people's claims about how good the 24mm TSE is with the A7R. I don't know, I can only go by what I have and my tests. ...Show more →
Are you able (and willing) to take your lens and adapter and try them out on another A7r2 at a shop ? Success proves nothing conclusively because you might just be lucky, but a bad result would surely imply that the camera variation is at least part of the problem.
OK, here is my 24 TSE at f8. shifted extreme corner, 100% crop, no sharping. First is no shift, then +6, +10, and +12. +12 is not great but even +10 is pretty good.
EDIT: I didn't change exposure and color balance was auto so that probably explains a lot of the color shift. I also don't have a 5DsR to compare to or I would. I am probably going to pick one up soon for my trip to Iceland in June.
