Front focussing issues with Canon lenses

[Sonnar2]

There was a front focussing discussion about the Zeiss ZM 50/1.5 some time ago in the forum. So I've made a quick check with some lenses at minimum focus and highest aperture on my Canon 7 (which I never had any focussing problems with, but not making close focus shots wide open on a regular basis).
On a bright day I've shot some pictures in my kitchen with a measuring tape on the floor and a small object in angular angle.
Shutter speed was about 1/500-1/125 depends on the highest aperture of the lens (checked with my Bessa-R TTL meter).
With the rangefinder I selected my shooting position, focussing at the 1m or 1.1m mark on the tape as accurately as I could, according to shortest focussing of each lens. I was not using a tripod and the shooting-angle to the floor was about 45-55°, not exactly measured, so the whole setup was quite simple and rough. A quick test. Film was Fuji 200 color negative (one film)

Results:
Canon 50/1.4 : perfect
Nikon 85/2 : perfect
Canon 50/1.5: 4-5cm before the mark
Canon 35/1.5: 8-10cm before the mark
Canon 50/0.95: 8-10cm before the mark
Canon 85/1.5: 8-10cm before the mark
Canon 85/1.8: 8-10cm before the mark
Canon 50/1.8 (late version): 10 cm before the mark

The shots were NOT focussed at the small figure. The figure stood level to the focussing line, resulting that the top of the figure which was closer to the camera was sharp and in focus whereas the the food and the 1,00m line wasn't. Notice the very small DOF at most pictures.

I would believe my camera is wrong IF NOT I've noticed earlier some front focussing at very close distance wide open with the 50/1.5 (on other cameras), 50/0.95 and 85/1.5 AND the 50/1.4 was absolutely perfect. I was surprised that the 85/1.8 nearly shows the same amount of front-focussing as the 85/1.5 (but at higher contrast). The front-focussing could be easily detected by low-quality scans (1500x1000). The crops of the pictures following are about 25% (35mm; has to be enlarged from the scan about 1,25x), 35% (50mm) and 60% (85mm)

Of course this does not mean that there are front focussing issues with the same lenses at less close distances, and of couse my individual lenses can be wrong as well.

Anyone has made these kind of observation as well or did some tests?

 

[Sonnar2]

next two pictures

 

[Sonnar2]

last three lenses tested

 

[xayraa33]

the 50/1.5 usually back focuses 4-5 cm after the mark at the 1 metre distance at f1.5.

 

[peterm1]

I had an 85mm Canon - the chrome version (I thought it was f1.8 but was told that the chrome was only in f1.9 and f2.0) Anyway that is beside the point. I found that at middle distances the subject was typically out of focus. I never worked out if it was focussing ahead of or behind the subject. However, I presumed it was focussing behind the subject as I was not shooting wide open and would have thought that if the lens was focussing ahead, the subejct would still be in the zone of focus - at least sometimes.) I eventually sold it without getting to the bottom of this issue. But I think Dante Stella has made mention on his web site that he has come across similar issues with Canon RF lenses. I did notice BTW that at infinity the focus point was out as well. (Cannot recall which way but I suspect it never quite reached true infinity before the focus mount stopped turning.) I suppose this implies it may have been focusing ahead of the correct point so I am not sure where this leaves my original belief about the lens focussing behind the subject.

 

[Charles D. Orze]

All of my classic Canon lenses front focus

All of my classic Canon lenses front focus

Hi,

I have raised this issue before and have noted incidental remarks from others. My 35 f1.5, 50 f1.5, 50 f1.8, 100 f3.5, 50 f1.2 ALL front focus a bit at close distances. Their cams at infinity are also shorter than their Leica and Voitlander counterparts.

Charlie

 

[Sonnar2]

I had 100/3.5 newly acquired and took it to a wedding. It turns out that every single picture taken with this lens was back-focussed badly and I kicked myself for taking an un-approved lens with me. My other 100/3.5 lenses behave nicely, so I don't believe in "errors by design"

 

[John Shriver]

Since the 100/3.5 has a removeable head, someone could have done a head/helical swap on it. which could cause calibration problems.

So far as I can tell, there's no systemic error with my Canon lenses compared to the Leica ones. The focusing cam comes out the same 7.5mm at inifinity on all of them. I just mated my Canon 50/1.8 back-to-back with my Summicron 50/2, and they match up perfectly (cam to flange between lenses).

Now, someone may have opened up Sonnar2's Canon lenses for cleaning (it's easy), and not screwed in a group to the same level of tightness. (I noticed a prior repairman had scribed this on my 50/1.5.) Not tightening the same amount would lead to a different focal length, and thus a focusing error.

One check to do. See if the focusing error is consistent for all focal lengths, or if it's worse as the focal length of the lens is shorter. If it's the latter, it's not a cam problem, it's a collimation problem, probably in the body. (Wrong flange to film distance.) You can fudge the "gain" of the rangefinder (cam arm length) to correct for this collimation error -- but only for ONE focal length. For other focal lengths you would be off.

Properly calibrating the focus on a Canon RF (according to the service manual) required a gauge block and depth gage to check the flange-film distance, and a reference dummy lens with a very precise 7.5mm cam. Then you also need to adjust the rangefinder "gain", I don't remember how they specify to do that.

 

[dexdog]

Sonnar2, your recent test of the Canon lenses intrigued me, so I collected all of my fast (less than f:2) LTM lenses, and photographed a carpenters ruler on a table top. Distance from film plane to point of focus was approximately 36 in/91 cm. Point of focus was on 36-inch mark on ruler. I cannot post any scans right now (7 old computer is dying, does not recognize scanner). Test camera was a Canon 7sZ overhauled in 2006 by DAG. The results are as follows; a minus sign in front of number means focus is closer than indicated.

Canon 50/0.95, f: 0.95, - 4 in/10cm
Canon 50/0.95, f:1.4. - 4 in/10cm
Nikkor 50/1.4, f: 1.4, - 1 in/ 2.5cm
Nikkor 50/2, f: 2, - 1 in/2.5 cm
Canon 50/1.2, f: 1.2, - 1 in/2.5 cm
Canon 50/1.8 black, f: 1.8, zero, dead-on focus
CZJ 50/1.5, f: 1.5, -2 in/5cm
Canon 50/1.4, f: 1.4, -1 in/2.5 cm
Canon 50/1.5, f: 1.5, - 1in/ 2.5 cm, lens #1
Canon 50/1.5, f: 1.5, -1 in/2.5 cm, lens #2
Canon 35/2, f: 2, -1 in/2.5cm, lens #1
Canon 35/2, f:2, zero, dead-on focus, lens #2
Canon 35/1.8, f: 1.8, -1 in/2.5cm
Canon 85/1.8, f: 1.8, zero, dead-on focus
Canon 100/2, f:2, - 1 in/2.5cm
Canon 35/1.5, f: 1.5, - 1 in/2.5 cm

Overall the results were pretty consistent. Most of the lenses in this sample have a near-focus limit of about 3 feet/1 meter (Nikkors excepted). The only lens that significantly close-focused was the 50/0.95. The Canon 35/2 lens #2 is the one that Raid used in his wide-angle test in Spring 2007. Those of you who followed that test may recall that this particular lens performed poorly, and Roland and other suggested that it needed to be collimated. Anyway, I had the 35/2 cleaned and collimated by Peter Smith recently, and now it appears to be very good. The CZJ 50/1.5 is the R-Sonnar that I obtained from Roland last year.

Lastly, just for laughs, I used a Canon 50/1.8 in FD mount with a Canon FD to LTM adapter. This particular combo close-focused about 7 inches. I guess that means one should not mount SLR lenses on RF cameras and expect to use them wide open.

 

[Sonnar2]

thanks dexdog, happy to read I'm not completely wrong.
We all know statements going like "Canon RF's aren't corrected for close focus" Maybe this is what stands behind it. On the other hand the whole forum is full of great pictures of the Canon lenses at middle distances (say, 2m-5m)
I haven't made a test with my Nikon S2 (my only lens for it is a black 50/1.4) but I 90% of the close distance shots taken with it were perfectly in focus whereas with the Canon 50/1.5 LTM just about the half. And I had very sharp results with it at f/5.6 in about 3-5m...
my result so far: horses for courses..

cheers, Frank