More Canon 85 EF 1.8 Woes - SLR Experts?

[Tom Stanworth]

My original 85 1.8 went back to the dealer some months back due to massive back focus. We are talking about 50m at 50m, despite the image looking sharp through the VF. Close subjects at, say, 1m were completely blurry on film despite being sharp in the VF. I tested the lens on two digi bodies and same issue. I am talking wildly not even remotely in focus...

Now I have my second copy from the same dealer. It is smack on at portrait distances and surprisingly sharp at 1.8, gaining contract and resolution a stop down. However, the further the focus distance the more back focus starts to creep in. This is the same as the lat example but not nearly so bad!

I tried comparing manual focus and AF and found that results, oddly, were a hair sharper in MF; however, this did not change the appearance of growing back focus with distance.

At distance, say 15m, the lens will show sharp on centre in the VF but in fact be obviously very soft on film at 1.8 and still soft at f2.8/4, with actual focus somewhere behind (at about two metres behind the subject it is RAZOR sharp). The left side of the frame also appears sharper than the right at distance at the wider apertures. I have shot some subjects where - and I cannot get my head round this - sharpness appears at various points in the frame that appear unrelated to the plane of focus (i.e. focus a 50m. Centre soft as hell even at a stop down. Right side (parallel to focal point) even softer. Left centre (once again parallel to focus point, pretty sharp), lower left not bad, lower right, soft as hell, infinity about the same across the frame an much better than the point of focus! I am baffled. I have never had an issue with a canon lens before (apart from the last 85 1.8). I have many other brands for all formats and apart from a dodgy CV 21P which was massively decentered in a single plane, have not seen such gross errors with lens peformance. In this case its like a LF camera with swings and tilts I cannot make sense of (like subtle, multiple decentering AND back focus at distance)

How can the lens be pin sharp at 1m and then back focusing terribly at distance? Is this normal? At about 3m and wide open, I estimate about six inches of back focus, which is only nullified by stopping down to about F8!

Do you think there was just a bad batch of lenses? I assume the lens should perform very well at distance, despite being primarily used for portraits up close. I have not had such issues with their cheaper zooms, or L lenses. All have been great, focusing accurately throughout the range. This one is amazingly good at 1m but unusable at less than f5.6/f8 at medium and longer distances, requiring DOF to make up for back focus.

I am very confident it is not the body. This body performs spectacularly with my 70-200 f4, 135 f2L etc at ALL distances.

Any thoughts? Rapidly losing confidence in Canon here. I desperately need this FL/speed for a project (in progress) and have already missed shots because I had to send the last one back. Am I really going to have to spend $1900 on the 85 1.2 to get a lens that works as it is supposed to?

Sorry for the rant... V frustrated, annoyed, confused...

 

[Leighgion]

I don't know if your individual lens issue falls under this, but Canon has apparently been having some major issues both in design and quality control in both lenses and camera bodies. The bulk of the problems are falling on the AF with no tangible answers in sight.

While I am a Nikon shooter, I take no joy in this. I haven't seen the Canon 85mm 1.8 mentioned specifically, so hopefully your problem is isolated and not endemic.

 

[grompirx]

Never had a focus issue with my EF 85 f 1.8. It is always fast and accurate.

 

[Chazzy]

I've been thinking about acquiring this lens, so I will follow this thread with interest.

 

[Ian Grant]

Is the focus screen in correct register, a very slight issue there will have big ramifications on focussing.

Iam

 

[df cardwell]

Q #1 Film or Digital ?
Q#2 Which body ?
Q#3 Is you manual focus ability accurate ? See Q #2.

WHAT are you trying to shoot ?

"Canon has apparently been having some major issues " Document that, please.

It is practically impossible for this to be due to 'bad lenses'. The 85/1.8 is a time proven standout.
Most 'back-focus' issues can be traced to digital issues, or inherent AF behavior.

Finally, what Ian asked.

 

[Chazzy]

Doesn't the focusing screen of an SLR in proper adjustment see exactly the same focus as the film? I don't understand how the focusing screen could show something different from what wound up on the film, unless there is a camera problem.

 

[eddym]

I assume this is not happening with other lenses on the same camera? If not, I can't help.

But if the same thing happens with all lenses, then it could be that the mirror stop needs adjustment. The mirror only needs to be a little out of line to throw the film focus way off, whilst the image appears sharp in the viewfinder.

If it's only this lens, then I don't get it. If it's sharp on the screen, it should be sharp on the film, whatever the lens.

 

[Mark Fisher]

Mine is perfect and I've shot wide open quite a bit. It is my favorite Canon lens. Two bad ones in a row, however, would make me suspicious of the current production run. I'd ask your dealer to order two of three of them to increase your odds. If that doesn't work, perhaps Canon could service the one you have.

 

[Tom Stanworth]

OK some answers to questions:

My eyesight is perfect and manual focus actually resulted in better sharpness on the close in tests. Fractional, but my eyeball is better than the AF.

Tested on an Eos 3 in perfect condition and was shooting perfectly with all other lenses, incl Canon 135 F2 wide open at distance. I will need to test this again but the 135 and 70-200 are in a different country now.

If mirror were out of alignment, would this not affect close distances too? The lens focuses perfectly in VF and on film at 1m. Spot on using a chart I use to check my RF cameras/lenses are aligned. The error creeps in at longer distances and is proportionate to distance.

At distance, say 50m, the lens does not appear to be performing uniformly across the frame. Right side is clearly worse than left. On film, focus is well beyond suject and only stopping well down sorts this issue out (same iss as shooting at 3-5m, where film focus is 6+ inches to rear)

I have checked the mirror and it is seated as it should be. The camera has had no impacts or knocks.

The first copy of this lens, used on same body, was so bad that every single frame from near to far was MILES out on film compared to what I saw in the VF. Were there a problem with the body, it would not explain how the first copy turned every frame to perfect bokeh on film, despite being sharp in the VF and the second copy very much better (the first being unusable in every regard, at every distance and every aperture i.e. shoot a brick wall at 5m and you cannot make out the bricks, only a haze with what might be fuzzy mortar lines. At 1m,focuising on the nose, on film you would have the back of their head sharp!) Shot at 50m, the focal point looked like the bokeh and 50-100m behind was pin sharp). The second copy; however, is still not usable without very obvious back focus, beyond the short range i.e 1m, but night and day better than the first. There are clearly serious differences between the two examples and the camera body has not left the cupboard between lenses.

Mirror issues would also not explain that focus does not appear consistent across the frame at distance. The right side is much worse than the left at wider apertures suggesting decentering.

I guess I will have to accept that this is a problem I wont be able to solve in a rush. I have a 12 month warranty so can shoot close up portraits (for which it is spot on) until getting a second body out here to check the lens against for longer distances. I am 90% sure the lens is the issue (after all I have had two lenses through with a common issue, just to very different extents), or, so looks like a second one will be going back to B&H once this is done....

 

[polyglot]

Sounds like it has either a tilted element or bent mount, causing a similar effect to using a tilt lens, i.e. a plane of focus not parallel to the film. That would cause significant confusion to the AF module because of the way it samples rays from each side of the exit aperture to determine the focus offset.

The other thing is that SLR viewfinders generally do not receive all the light that the film does - they are effectively vignetted. That means you see a DOF equivalent to something between f/3.5 and f/5.6 depending on the camera, and you won't really perceive small focus errors. You can check this by putting on a f/1.4 lens, setting it to f/2.8 or so, toggling DOF-preview and noting that nothing really changes at all in the viewfinder.

There's also the matter of how the VF glass is ground. For brightness reasons, it's often receptive only to a narrow range of incident angles (this is a big contributor to the looks-more-stopped-down-than-it-really-is effect). That means the VF doesn't get rays from the far edges of the lens' exit aperture, which means you not only can't really see a focus error with a good lens, you can't see gross tilted/misfocus effects on a defective lens. You can MF with a good lens wide open because you pick the midpoint between where you see it go out of focus in each direction - with a misaligned lens, that all goes to pot. You might find that if you MF with a subject on one side of the frame, take the shot and them MF with a subject on the other side of the frame, the focus errors are significantly different.

Since you have access to digital bodies, I'd do some rigorous testing with those, merely because it's free. Set up some test targets at 1m, 5m and near-infinity, do AF and MF tests with the targets at centre, top, left, bottom and right. The changes in the focus errors will tell you a lot about how the lens is misbehaving. Of course, do some control tests with some other lens just to make sure you've not suddenly gone blind

 

[df cardwell]

Tom

No clue. You've suffered enough.

If the EOS 3 is focusing the other lenses, replace the lens.

 

[Tom Stanworth]

For those following this, I ran yest another roll thru the set up and will be returning this second copy. The test was better planned to test for all issues and revealed:

The lens does begin to backfocus beyond about 1m. It is already noticeable at 1.5m and a real problem at 3m.
The lens also suffers from decentering. Pin sharp on centre at 1m, also on left side. Really impressive levels of detail here. Right side softer, particularly lower right, which is terrible. Left side and centre hardly improve stopping down. Top right very sharp by f4. Lower right still obviously less sharp than rest of the frame at f8. Not talking extreme corner either, but quite a portion of the frame. This would make portraits with the eyes etc in the top 1/3 of the frame a real problem as the right side would be used.

I give up. I have a stunning 135 f2 in the UK and will collect that next time I am back. Really I wanted to work with 85mm, but $1900 for a 85 1.2 II is beyond me right now and I have no faith that a third copy of the 1.8 EF will be any better. Such a shame. up close on centre and on the left, it is obvious what this this lens can do when it is put together right. I'm very disappointed with Canon and have no other SLR system to work with. For me, RFs and 90mm lenses is not going to work. I need quick AF for this job.

 

[df cardwell]

Tom

Canon has been making the 85/1.8 for ages, and it has been a great lens.

The 3 I have (different camera kits stashed in different places to make travel easier !) are dazzling.

Faith shouldn't have anything to do with this: you've has had an amazing run of bad probability !
Go for a third.

Are you in the US now ?

 

[Andy38]

My first EF 85:1,8 was old , but sharp , with no AF problem ; but it has been stolen about one year ago .
I have bought another and this new is not perfect , with short and long distance AF problems , on Eos 3 , 33 (Elan 7) and DSLRs .
The first was my favourite lens , often used at full aperture ; now , I don't use this new one any more...

 

[lilmsmaggie]

I've noticed similar issues using this lens on a Canon EOS 7N. VF will show sharp image but when printed, images are sometimes soft. I especially noticed with head shots that there will be areas of the image that are razor sharp but softness in odd places.

For example, I took a shot of tree bark lying on the ground that gave the impression of Bullwinkle the moose's head (I used this for an abstract composition). The outer edges and pretty much all of the image when printed was razor sharp except for a small portion in the center that was soft. Weird.

I've stopped using the AF feature on the lens and just use MF which helps but IMHO is still inconsistent.
I haven't tried this lens on a digital body yet. I'm hoping I get better results.

 

[Tom Stanworth]

I'm in Afghanistan and buy from the US through military mail. These are the only two new lenses I have ever rejected in my life! Its such a pig to get to the military base and send mail out that returning faulty items is a real chore. I could exchange it but worry it is a batch issue or that the container full of lenses was dropped from a few metres at the docks! Might well not be Canon's fault, but something as simple as that!

I'll decide in the next few days whether to exchange it or just send it back, but looking at the centre and left of those negs is torture because the lens evidently has the potential to be spectacular and at a decent price One thing really missing from my current project is portraits and with time dwindling the loss of two faulty lenses costing months of opportunity is particularly galling. At this point, I will do whatever it takes. If I have to fork out for a 85 1.2 then I will because I absolutely cannot let this project finish without some portrait work. Got to think on this...

Thanks all for your help and listening to my rant!

 

[polyglot]

Can you get a refund and buy one elsewhere that you know is from a different batch? Say a known-good second-hand one.

 

[df cardwell]

cannot let this project finish without some portrait work

Got it. Can you swing the 70-200 zoom (non USM) ? Yeah, I know almost the same as the f/1.2,
and the f/1.8 is supposed to be working !



Good pictures, Tom.

 

[modafoto]

Never had a focus issue with my EF 85 f 1.8. It is always fast and accurate.

Same here. Great lens!

 

[Tom Stanworth]

cannot let this project finish without some portrait work

Got it. Can you swing the 70-200 zoom (non USM) ? Yeah, I know almost the same as the f/1.2,
and the f/1.8 is supposed to be working !

Good pictures, Tom.

I have a 70-200 F4 and so dont want to double up here (it is also in the UK). I really wanted a smallish and light 85mm. No doubt the lens is a cracker when working right - I was amazed at the wide open performance up close. I will either have to wait for the 135 when I go home, get another 85 1.8 from elsewhere or plumb for the 85 1.2...

I already have a few portraits but they are the 50mm variety that are more upper/full body. Gotta make a decision soon, but am whincing at the cost of a 85 1.2. I could get a 100m macro, but I dont shoot macro and its very close to the 135 in FL.

 

[Tom Stanworth]

I've ordered a 85 1.2 II using my kidneys as collateral. I did some portraits using the duff 1.8 that I will be returning, by turning it upside down to that the soft lower right is not on the face hopefully they will be OK! This way if the 85 1.8s were subject to a terrible batch, I will be OK. Alternatively if B&H's entire stock came in on a container that got dropped from a crane I will be no better off with the 1.2. And there I was debating whether to bring my 135L out here and decided not to...

 

[df cardwell]

Merry Christmas, Tom !

The 85/1.2 was your destiny !

d

 

[elekm]

Maybe there was some extreme heat from global warming that caused one of the plastic lens elements to melt or cause lubricant to get into the electronics or something else bizarre. Or maybe that production line (in China or Indonesia or Japan) didn't do adequate testing on a batch of lens. One of those things that you'll never know, as I believe spot checks are done but not tests on individual lenses -- at least for non-L lenses.

 

[Tom Stanworth]

I just hope the 85 1.2 works out and is worth the loss of internal organs. Assuming no problems with the 85 1.2, I need to make sure I produce the goods while wishing I did not have a real job (one of those things that always gets in the way).... work is so inconvenient.