Elmar 90 puzzle

[Sean Moran]

Perhaps you can help ...

I've just had an M2 body sorted out after impact damage. It focuses fine with a DR Summicron, even at f/2, even with the goggles at minimum distance.

But ... it is slightly out of focus with an Elmar 90 at f/4 and close range. For example, a portrait subjects's nose is in better focus than her eye.

The strange thing is that on my other M2, both lenses are fine.

Any ideas?

All the best,

Seán in Tipperary.

 

[Sonnar Brian]

Which version of the Elmar is this?

The Elmar, like the Summarit and Sonnar, suffer from "focus-Shift" in which focus will move towards infinity when stopped down. I found the Summarit 5cm F1.5 is optimized for F2.8, and will front-focus at F1.5. That was the decision made by Leica years ago- optimize for stopped-down work when used close-up. The magnitude of the error is about what you describe, "nose to eye".

The Summicron, by contrast, has very little focus shift.

On an LTM 9cm f4 Elmar, I filed the mount down to move the optics in slightly. That optimized the lens for wide-open and close-up.

To solve "front focus" problems, the optics need to be closer to the image plane.

As examples of other lenses that I believe were optimized for stopped-down work, the Canon 50/1.5 appears to be factory optimized for F2.8. That opinion is after taking 3 of them apart. Nikon Leica mount lenses are optimzied for wide-open use. They suffer when used at 2 and 3 stops down from wide-open.

 

[Sean Moran]

Thanks for that quick response, Brian. It's an M-mount Elmar. I wonder if the differences between the two bodies is explicable in terms of tolerances. Perhaps one body slightly back focuses and the other front focuses? When it comes to the Summicron, it makes no difference, but if the errors add (as in the wide-open Elmar), it becomes noticeable. Maybe I should have it optimised for f/4 close up.

 

[Sonnar Brian]

That is my experience with some bodies- within the 0.02mm tolerance, and focus shift is "about" 0.05mm for many lenses.

My M8 and M9 "slightly disagree" with each other. But it works out, lenses that are perfect on the M8 wide-open work best on the M9 stopped down a bit. The M8 is probably off a bit, but I do not mind. I get to use the Summarit wide-open on it.

 

[peterm1]

Its one of the less discussed and even less acknowledged deficiencies of RF cameras that if anything is even slightly out of spec (even if still within tolerance) some lenses will not focus well on some bodies.

A related issue (which is discussed with much more readiness) is that spherical aberations can result in focus shifts as lenses are stopped down meaning that a lens shot wide open (and focussing accurately) can focus less accurately as smaller apertures are selected. (or visa versa)

Thats a different issue but my point is that neither are really a problem for SLRs as you see what you are focussing on through the shooting lens.

I have a few lenses I have issues with. Including some modern ones. I recently noticed that my Voigtlander 75mm f1.8 does not seem to focus fully to infinity on my M8 apparently resulting in some image softness - although I have shot some cracking images with it at closer distances.

 

[Sonnar Brian]

Focus shift is primarily a product of spherical aberration. It is just as much of a problem in SLr's where you focus wide-open through the lens, but it takes the picture stopped-down to the set aperture. Attempting to focus with the SLR lens stopped down is not easy using manual focus, and is not done with AF mode.

Aspherical lenses offer the possibility of much less focus shift.

Note the "none" for focus shift due to spherical aberration for the original F1.2 Noctilux.

 

[charjohncarter]

Focus shift is primarily a product of spherical aberration. It is just as much of a problem in SLr's where you focus wide-open through the lens, but it takes the picture stopped-down to the set aperture. Attempting to focus with the SLR lens stopped down is not easy using manual focus, and is not done with AF mode.

Aspherical lenses offer the possibility of much less focus shift.

Note the "none" for focus shift due to spherical aberration for the original F1.2 Noctilux.

I'm am not an authority on this (for sure) but, I have done some copying with SLR (Pentax) lenses. When it really got done to it I had to buy a macro for this work. Is this what you are talking about? Or am I'm reading a discussion that is off my IQ level.

 

[Sonnar Brian]

for copying work- flatness of field is the important feature of a lens. Lenses such as the Sonnar have high field curvature. Instead of a flat plane being in focus, the distances "weave in and out". So fence slats and brick walls- not a good subject for Sonnars.

"my mental image"

Spherical Aberration: rays from the outer portion of the objective focus slightly in front or behind (under/over corrected) the actual focal point. There is a "best focus" where most of the rays meet. As you stop down the lens, the outer rays are cut-out, and the best focus appears to shift.

 

[Steve M.]

Sean, for what it's worth, I've had this problem w/ two of the three 90 Elmars I've owned. One was tack sharp at all distances, the other two were sharp only at infinity and mid-distances. Close up, those two back focused. All were on different cameras.

I never did sort out what the problem was (and truthfully, I disliked the way the lenses stuck out like a piece of pipe hung on the front of the camera so much that I sold all of them, even the sharp one).

Issues like this, along w/ my ability to focus an SLR better, convinced me to sell my RF gear and go back to shooting an SLR w/ R lenses. The difference in size and weight is pretty noticeable, but so are the results from the R glass, so I can live w/ the compromises. An SLR. assuming your chosen camera offers this feature, allows you to focus wide open, then stop down and use the DOF preview to confirm what you're going to get on the film or sensor. If you're shooting macro (or micro for the Nikon shooters here), you're probably shooting wide open anyway.

 

[Sonnar Brian]

I'll try some "fence Post" tests with my Elmars, I have a 9cm f4 collapsible and a pre-war 9cm f4.

This is with the Collapsible 9cm F4 Elmar on the M8, wide-open.

On the M9, it would probably be best at F5.6.

 

[Sean Moran]

Thanks for all this amazing detail, folks. I think perhaps the best thing I can do is to know which lenses are calibrated to which bodies.

Seán.

(Currently slightly browned off that I won't be going to Pakistan for Christmas, January and February. My visa was refused on a minor paperwork technicality. Ah well, there's always 2012)

 

[Roger Hicks]

Thanks for all this amazing detail, folks. I think perhaps the best thing I can do is to know which lenses are calibrated to which bodies.

Seán.

(Currently slightly browned off that I won't be going to Pakistan for Christmas, January and February. My visa was refused on a minor paperwork technicality. Ah well, there's always 2012)

Dear Sean,

I'm equally p'd off that Frances and I won't be going to India at around the same time, because they have now outsourced their visa service so that there are more hoops to jump through than I care to contemplate: a classic example of "in order to improve our service to you, we have made the service worse."

It's Maggie Thacher's fault. When I first went to India in the early 80s, I didn't need a visa. Then, the Maggon imposed frankly racist transit visa requirements on travellers from the sub-continent, and their governments retaliated. As long as visas were 'in house', I could get through to the High Commission. Now, with 'outsourcing', as you say, it's ridiculous tiny technicalities by Jobsworths ("Nah, more than my job's worth, that is"). If I didn't need a visa, I'd have spent quite a few more lakhs of rupees in India than I have -- far more than their piddly visas cost, and far more to the benefit of the Indian economy.

Cheers,

R.

 

[Sean Moran]

I feel your pain, Roger. To ease the blow in our case, I've arranged a refund for our flights (minus the £100 each refund fee deducted by PIA) to be spent on a couple of weeks in Istanbul over Christmas and the New Year. I'll be taking a 1956 Leica M3, a DR Summicron 50 (with goggles), a meter, plenty of hp5+ and nothing else. Is anything else (photographically) needed for Istanbul? It's a city I've never visited before.

All the best,

Seán in Tipperary.

 

[Sonnar Brian]

It's been raining for the last two days, I will post some "focus shift" tests when the sun pops out on the fence Post.

 

[Sonnar Brian]

Got a little break from the clouds right before sunset- Took the 1954 9cm F4 Elmar out on the M8.

focus shift is NOT a problem with this lens. Makes sense as it is a long focal length lens. The 9cm Elmar is not a telephoto formula, but an anastigmat with longer than normal focal length. The Bokeh appears neutral, so it would line up with the top diagram in the figure.

So what is going on with your camera: one is probably slightly out of tolerance, which shows up with a longer focal length lens.

I'll have to walk around with the 5cm F3.5 Elmar and 5cm F2.8 Tessar. With the F1.5 Sonnar, the focus shift is very evident.

But don't worry about it, even if you first learned about it on the Internet.

 

[Sean Moran]

Thanks Brian. Your theory sounds highly plausible. I mainly use the 50, so as long as I remember which bodies work well with the 90, there's no issue. All the best, Sean.