35/2.8 short-focusing

[ChrisN]

Earlier this month I sold my 35/2.8 lens to another member, but he has found that the lens focuses short of the proper focus distance, wide open and close up. I had not tested it for this, and it did not show in the rolls I shot with the lens. I guess this can be a problem with old lenses - you can never be certain what you are getting. 😱

If he can't find someone to correct this in Hong Kong, I'll take the lens back and refund the buyer. Can anyone recommend a good repairer/service person for this in HK?

And could anyone explain what is likely to be the problem, and how to fix it? Anyone interested in taking it on? 🙂

 

[rvaubel]

ChrisN
Is this a screw mount lens? If so, the adapter could be to blame. How much is it front focusing? An inch or two at 3ft isn't really that critical if your not working off a tripod shooting pictures of newspapers. If your shooting real people, chances are the focusing error will work for you about as often as against you. More critical is infinity focus. I have a couple of lenses that front or rear focus a little bit and are fine in practice. If your talking about 4 or 5" at three feet, then you have a problem.
Rex

 

[ChrisN]

It is a screw mount. Here's the message from the buyer:

"Hi Chris,

The lens arrived today. Everything is ok except there is a close focus problem. I tested shooting @F2.8 but the focus landed a few inches in front of the subject.

I'll try to see if this could be fixed locally. Pls send me your email address if you wish to see a sample.

Will keep you informed.

Regards,
Johnny

I'm guessing the buyer is testing it with an RD1 (because of the short time factor) and presumably with an adapter that worked ok with other lenses. I haven't pushed on this point because I don't want to appear argumentative. From memory close focus could be about 700mm.

 

[rvaubel]

It is a screw mount. Here's the message from the buyer:

"Hi Chris,

The lens arrived today. Everything is ok except there is a close focus problem. I tested shooting @F2.8 but the focus landed a few inches in front of the subject

I'm guessing the buyer is testing it with an RD1 (because of the short time factor) and presumably with an adapter that worked ok with other lenses. I haven't pushed on this point because I don't want to appear argumentative.

He probably does have an RD-1. I do too and it wasnt till I had one that I was able to review lens test results so quickly. It made me overly anal for awhile. The lens actually sounds OK to me (if its OK at infinity). Their are so many other factors that could come into play especially with the RD-1. But, like you say, you dont want to seem argumentative.
There are actually ways to fine tune screw mount lenses by grinding the adapters but that is the sugject of a long thread. Its actually kind of fun getting the best from your lens. Its amazing how many lenses are slightly off. My recent experience in testing lenses has lead me to a theory that peoples "favorite" lens may have more to do with the chance of getting a good sample, focusing wise, than anything else.
Anyway, if you feel quilty, take it back. Although there is a fix, it requires some dedication to testing that most people arent up for. And you most assuredly need a digital camera for testing unless you think your going to live for another 75 years.

Rex

 

[ChrisN]

... Anyway, if you feel quilty, take it back. ...

Rex - thanks for the comments and thoughts. I'm not feeling guilty (nothing to feel guilty about here!), but if the lens is faulty and the buyer isn't happy, I'm not going to risk hostilities over so little! I can always offer it again as a fixer-upper, here or on ebay. I'm sure the little lens will find a good, loving home eventually!

 

[Kim Coxon]

Hi Chris,
It might not be an RD-1. Such things can be checked using a ground glass screen and a lupe as long as you can get to the film rails (ie no fixed-back bottom loaders). It is how I set up some of the FSU lenses notably the J9s. As to the "problem" it can be down to several things.

First off assuming the lens focuses accurately at infinity (this is the critical one), it could be the camera RF mech. Not only does the vertical and horizontal alignment have to be correct but the "throw" as well. Assuming this is good and inf is good, there is a mismatch between the RF mech and the "actual" focal length of the lens. On FSU lenses, you cannot adjust the caming so you have to adjust the lens groups to match the focal length. On the Canons, it may be possible to adjust the caming.

However, in this case the difference is so small that I would not consider it a "fault". In any case, such territory is in the province of SLRs not RFs and at those distances the DOF is so shallow that it isn't really practical to use. If the buyer is not happy, by all means take it back as you suggest. I wouldn't sell it as a fixer upper but as an old lens which is perfectly usable at most distances at all setting but that should be stopped down very close.

Kim

Earlier this month I sold my 35/2.8 lens to another member, but he has found that the lens focuses short of the proper focus distance, wide open and close up. I had not tested it for this, and it did not show in the rolls I shot with the lens. I guess this can be a problem with old lenses - you can never be certain what you are getting. 😱

If he can't find someone to correct this in Hong Kong, I'll take the lens back and refund the buyer. Can anyone recommend a good repairer/service person for this in HK?

And could anyone explain what is likely to be the problem, and how to fix it? Anyone interested in taking it on? 🙂

 

[ChrisN]

More information: from the buyer -

Thanks for the effort. Actually I'm using a RD-1. The camera focuses very well with my Leica 75/1.4 & 50/1.5.

With the Canon, the patch lines up correctly at infinity but around 4" off at 3 feet.

When doing indoor portraits, I'm trying to either step down the lens or focus a little bit further to see if this works.

Also, I'll bring the lens to a shop for checking this weekend. This guy has good reputation & has serviced some of my old lens & body.

Cheers,
Johnny

Now this has me thinking- the camera is ok (as tested with good Leica (and M-mount) lenses). The relationship between lens cam travel and rangefinder focus patch travel has to be linear. If this lens is ok at infinity, but not ok close up, does this indicate the screw-mount adapter is the problem? Or that the cam travel in the Canon lens is non-linear?

Edit to add: ... or has someone "serviced" the lens and gotten the focusing scale out of whack with the correct position of the focusing helicoil? I'll bet that's it - it would not show up at infinity, but it would certainly show close up.

 

[john3]

... or has someone "serviced" the lens and gotten the focusing scale out of whack with the correct position of the focusing helicoil? I'll bet that's it - it would not show up at infinity, but it would certainly show close up.

That's also what I'm thinking. A common problem for old lens.

Johnny

 

[rvaubel]

ChrisN
This "near focus" thing opens a real can of worms. The problem is it is much harder to check infinity focus and many people assume that its OK. Most people test infinity focus outside (naturally!) and forget that to be effective you have to shoot wide open. Not easy to do in sunlight with a minumum ISO of 2OO on the RD. You need neuteral density filters and even then the DOF of a 35mm lens is so deep as to disquise a slight focus error that would easily show up at three feet wide open.
Since your customer thinks his other lenses work OK then that leaves the helicoid or cam in the lens. The heliciod on the wrong thread is very common with some of the Russian longer lenses, but I've never seen it on a canon. However, ever lens is different
I have a 21mm Kobalux that did the same thing and I cured it by grinding off about .010 of the adapter. Thats about 10% of the overall thickness of the adapter! This ws really radical surgery and I think only applicable to the Kobalux. Normally such action would screw up the infinity focus but the Kobalux just seemed to be built about .010 long. Don't try this at home without parental supervision.
Actually, you can make a lifetime pursuit of dialing in rangefinder cameras and lenses. I probably have the largest collection of pictures of newsprint in the known universe.

Rex

 

[akalai]

Hey Chris you might want to have a look at these links, particularly;

"Another thing to worry about: Bad screw/bayo adapters!"
http://www.rangefinderforum.com/forums/showthread.php?t=5438

..also worth scrolling through this thread:

"Jackpot!! Deadstock Avenon 21mm f2.8!"
http://www.rangefinderforum.com/forums/showthread.php?t=5408&page=4

...there was another thread as well which I had listed in my thread subscriptions but lost after one of the recent RFF updates and can't remember what the thread title was, so be interested if anyone reveals some more relevant links...

 

[akalai]

"Fix for front focus with Avenon 21/2.8"
http://www.rangefinderforum.com/forums/showthread.php?t=5799

...still seem to remember another thread on this topic...

EDIT: Here is another useful thread with plenty of pics to help illustrate this problem:

"First of the Avenon's!"
http://www.rangefinderforum.com/forums/showthread.php?t=5732

 

[Kim Coxon]

Hi Chris,
If the lens is focusses correctly at inf and the RF patch is correct, then the RF cam must be in the correct place and the adapter must also be OK. If either of these are wrong, then the lens would be out throughout the range. Furtheremore the focussing scale as such should not affect the relationship between the glass and the cam but purely the read out.

So having eliminated these thoughts, you need to look at the relationship between the effective focal length and cam movement. I haven't examined a 35/2.8 but I have seen most of the other 35's, the 28/2.8 and 25/3.5. The construction of all these is basically the same so I am assuming the 2.8 is similar. Now, if it were 1" out as per your first post, I would put it down to manufacturing tolerance and wear but 3-4" is too much for this. If the lens's point of focus was behind the point that the RF focusses on, it would indicate that the focal length of the lens is slightly too long which could be caused by the rear group being too far back. This tends to be a cemented pair in the Canon's held in place with a collar. If someone has had this out to clean the lens, it might not be seated correctly. If the actual point of focus is in front of the RF patch point, then the focal length is too short and the rear group is effectively too far forward. This would be almost impossible. A much more likely scenario (which could also cause the former case) is that the rear group or an element of it, (if it is not a cemented doublet) has been put in the wrong way round as all these groups are asymetrical.

Kim

Now this has me thinking- the camera is ok (as tested with good Leica (and M-mount) lenses). The relationship between lens cam travel and rangefinder focus patch travel has to be linear. If this lens is ok at infinity, but not ok close up, does this indicate the screw-mount adapter is the problem? Or that the cam travel in the Canon lens is non-linear?

Edit to add: ... or has someone "serviced" the lens and gotten the focusing scale out of whack with the correct position of the focusing helicoil? I'll bet that's it - it would not show up at infinity, but it would certainly show close up.

 

[ChrisN]

Thanks for the links Akalai - I'll check them out.

Kim - just to clarify - "If the actual point of focus is in front of the RF patch point" means between the camera and the subject, right?

 

[Kim Coxon]

Chris,
It is quite logical when you think about it. The RF cam in the camera is set for a "normal" 50mm lens. If you use a longer lens, the glass has to move more than the cam and vice versa. If you put a tape measure on the table and took a picture obliquely along it, then if you had a 50mm glass in a 50 mount and focussed at 1m, then both the glass and the camm should be at 1m. If you were foolish enough to replace the glass with the elements from a 35, the RF would still be matched at 1m but the glass has moved too far. In other words the distance scale on the tape would be foccus much closer to the camera say 80cm. In this case to correct, you need to increase the effective focal length of the glass to 50mm to achieve the correct focus but in all cases the focus would be correct at inf.

In a lens of "normal" construction, you can make small ajustments to the focal length by altering the distance between the front and rear groups. If you reduce the distance, you shorten the focal length. It is this pricinple that Brian Sweeney used on the J3's and I have used on the J9's to correct focus errors. It is not perfect but it gives workable results for small changes. This would not work for wide angles that use the retroficus principle but as most RF wides are not retro lenses, the theory should apply.

I would not expect the Canon to be out unless something has altered since it was adjusted in the factory. Off the top of my head, I can think of 3 things which would affect the apparent focal length. The first is a change in the shape of the glass ie if someone has had the fromnt element repolished and coated badly or has "polished" out any haze. The second is if an element is niot seated correctly in the right place and the third is that an element or group has been replaced the wrong way round.

Kim

Kim - just to clarify - "If the actual point of focus is in front of the RF patch point" means between the camera and the subject, right?