Canon 20D close focus problems

[drh681]

right. five lenses(all EOS) "out of adjustment".

think it through, If it were the lenses, would not at least ONE of them been "in adjustment"?

For background, I have been using EOS stuff for seven years(I sold it for five), I know how most of it works and have not, until now, been baffled by a quality control problem.

[Jonathan Wienke]

not much for statistics are you.

or up on manufacturing process.

I have no complaint of these lenses on several different film bodies. so why would they all of a sudden be out of adjustment on a digital body?  and again I ask, why none of them will work with this body? five different lenses, bought over several years, and ALL of them are out of adjustment?
NOT LIKELY.

here is what I guess; Canon has moved the assembly of these bodies to an off shore plant. to facilitate the huge increase in unit volume they are using assembly jigs and sub-assembly of certain components(sensor+mirror box) and someone is passing that sub assembly at one limit of tolerances, while someone else is passing the body assembly at the opposite limit of tolerances. So both sub-assemblies "pass" their inspections; but together wont close focus.

The flaw in your theory is that most other 20D owners would be screaming about focus miscalibration if it had any basis in fact. I still think the most likely cause is user error, especially since you haven't posted any sample images or any explanation of your test methodology. In addition, Canon DSLRs are focus calibrated as a completely assembled unit with a "tool lens" that is used as a known good focus calibration reference. So your guess is pretty far off base.

As to my understanding of statistics, I have a prettty firm grasp of the obvious. Yes, 5 lenses out of adjustment is more unlikely than three bodies, but your argument regarding those lenses working fine on film bodies is totally worthless. Digital is much less forgiving of focus errors than shooting with film because digital sensors record more detail per area unit than film. In addition, people are enlarging digital images much more with digital than with film; when was the last time you made a 24x36 inch print from 35mm film? I do so regularly with 1Ds images, and so do you if you view your images at 100% in Photoshop. The more you magnify an image, the smaller your effective CoC becomes, and DOF shrinks accordingly. So with digital, you're able to appreciate focus miscalibrations of less than an inch that used to be buried in the film grain. It's the same reason that 1Ds owners complain about lenses that used to be highly regarded in film days. The sensor captures so much more detail than film that lens aberrations like CA and coma and suchlike that were smaller than film grain structure are being revealed in all their glory by the unforgiving slab of silicon.

Post your test methods and some sample images.

[drh681]

an inch or less of focus error could be well within the acceptable error tolerance of the 20D AF system design. Except for macro work (which is generally manual focus anyway) and some wide-aperture portraiture

I'm sorry, but you seem to be nuts.
An inch of focus error is not the way any camera manufacturer maintains a reputation for quality.
And this whole thing started because of the poor results I was getting when shooting close ups! You also seem to imply that it is OK for the 20D to not perform certain aspects of photography.

I do indeed say that there is a problem in the late production run of 20D bodies.
A problem that as you say, may go un-noticed in the run of snapshot photography. But it is a problem that should not have shipped.
Snapshots are what I have a Minolta S414 for!

A point of logic:
using manual focus through the view finder will not get a focused image on the sensor; if the sensor is out of the plane of focus! you can focus with 20/20 vision all day and still have front focus. Believe me I have tried.

Auto focus works on sensing the difference between light and dark in your subject. these sensors are not in the film/sensor plane, they are in the periphery of the mirror box. so if the mirror and sensors are correctly positioned, you will see a focused image in the view finder, and get a focus lock signal but still have front focus due to a misalignment of the image sensor.

[heavysigh]

drh681,
     Sounds to me like you're doing everything correctly. You're performing a variety of tests over three bodies and multiple lenses. If you can consistently repeat your findings then the next step may be to contact Canon directly.
     I've worked in the software industry for twenty years and, at this point, I am never surprised when a released product that has been in the field for years suddenly hits a bug by one customer who is doing everything correctly. Actually, it happens all the time. You often find that a slight variance in circumstances (a linkage library definition being slighty changed) can cause a world of difference.
     Also, it may simply be a bad batch. The fact that the serial numbers are so close together seems to lend weight to that argument. As they say in the automotive industry "Don't buy a car made after lunch on a Friday".
     As for forum members who take themselves too seriously, ignore them. You run across insecure people who overcompensate by belittling others and trying to dominate conversations in all aspects of life. The internet just gives these people a permanent home. If you don't believe me, then the next time someone dumps on a person who is just asking a question or is expressing a sincere opinion, check the number of posts for the dumper. You'll usually find a very large number that screams "Love Me Daddy" for all to hear.

Thanks,
Heavysigh.

[Ray]

Jeez, is that the long way of saying that the lens turns until the camera tells it the image is sharp?

Not only that, but it's a good method of analyzing how much manual adjustment is required to bring the target into precise focus, without having to download the images and peer at them on the screen. If you have more than one lens of the same focal length, which is quite likely if you have a number of zooms, you can quickly see if if they all behave the same at a particular FL and distance from the target.

[drh681]

Oh Wow!

 :laugh:

[drh681]

I want to thank everyone for their input.
This unit is going to Canon for "adjustment"
and is not being used for any critical work until it comes back and passes the test.
For those wanting to lay the problem on the lenses, I made some of the same images on film last week, and had no signs of the front focus problem.

It is easy to presume a user interface problem with newly aquired equipment. Indeed the few things I have used it for have pointed up a slew of "doofus factor" points. (mostly in remembering to reset white balance, etc.) I was doubting my camera handing skills myself till I decded I had done everthing as usual with the only variant being the camera body.
These discussions have helped me sort out my thoughts.

[drh681]

the camera and 100mm are covered, the fifty is going to be 40$.
no offence taken, I understood your reply.
I have noticed that on many boards there is a tendency for the first responders to presume that the poster is "not doing something right".(a fair presumption often enough) I had that go-round with Jonathon. In fact the only offence I took was when someone tried to pick a fight with me for standing my ground then.  
I welcome comments and wont take offence to didactic questioning.(I do it myself some time, but I try to keep it more in the nature of a check list)

[Eric Myrvaagnes]

Great! Looks to me as if they moved it from really yucky to really usable. Glad it worked out well.

Eric

[Ray]

The question arises of course, just how many cameras (and not just the 20D) are slightly out of adjustment to a degree that's not as obvious as in the above example, giving the impression to the owners that slight misfocussing is normal and unavoidable with their quality of camera.

Not that I'm trying to create a panic.....   .

[drh681]

I have been having problems with front focusing when the subject is within six feet or so.

the first two cameras were only one s/n apart. this third body is weel off that sequence but still has a front focus of one half to one-and-a half inches in front of the selected focus point.

Am I missing some info about a "design feature"?

After two returns, I agreed with the seller that I would deal with this through Canon if it showed up in the third. So tomorrow is for Canon techhelp; but it really gripes me that I need to send a camera for "adjustments" right out of the box!

[Jonathan Wienke]

That's not much more off the wall than 3 consecutive bodies "out of adjustment". What's even more likely at this point is some type of flaw in the test methodology, or other user error.

[drh681]

when was the last time you made a 24x36 inch print from 35mm film?

well... 20x30s about 3 years ago. some from ISO 800 and 1600, I know what film grain is and what can be made out of it.

as for the method, I used the oft posted 'target on a 45° rule' method. in filtered daylight and under flash with a tripoded camera. for a "real life" test I used a set of  three staggered film boxes and focused on the center box. Center focus point selected.
plus a few dozen flowers and tiny things.
it was the flowers that actually tipped me off. I had no Idea that this was an issue. as you say, small errors in focus point do not show up in longer veiws with their much greater DOF let alone the consequent smaller apertures. even close up the zone of appearent sharpness will extend past the target at small apertures. BUT the actual focus point is still in front of the target.
 
trust me, it is not a 'sharpness issue' either, I know my lenses and what I can expect of them, and what not to expect. They are not L lenses, but they are not the mediocre kit zooms either.

I have a prettty firm grasp of the obvious. Yes, 5 lenses out of adjustment is more unlikely than three bodies,/I]  
you miss the point that the three 20D bodies were all made in the same place and in the same time frame.
to have all the lenses, manufactured over many years, and possibly in different places, showing "focusing errors" would indicate a VERY long term Q.C. problem at Canon. again, not likely.

why are you busting my chops?
you dont even use the camera!

[jani]

To me it sounds very unlikely that you would be the only 20D owner on the forum to have this problem unless it is some problem with your procedure. Yet you sound like a competent and experienced photographer, so that seems unlikely.

Eric, I've also had similar problems, but I choose to believe that it's a sum of the inherent lack of precision in the focusing system of the 20D, combined with a viewer that is lacking (at least compared to my old 650) in terms of showing when I've got focus, and again combined with lack of precision in wetware*.

My most notorious personal example was a series of test shots of the EF 50mm f/1.4 USM. I attempted to take shots of a cueball on a pool billiard table under excellent lighting conditions, using a tripod and autofocus (followed by locking to manual focus).

The camera focused a couple of inches behind the cueball, and I didn't notice. It was impossible to see on the LCD, too.

I'm not claiming that this is the case for drh681, though it certainly can happen.

Other embarassing moments have occurred when a finger has changed the diopter adjustment ...

* Wetware is, for those who aren't familiar with the term, basically the human neural system, including the brain.

[drh681]

unrealistic expectations that are more appropriate for a 1-series camera with a much faster and more accurate AF system that functions well in much lower light levels than the consumer-grade AF system of the 20D.

uh... real proud of that 1Ds aren't you.

Canon rates the 20D AF sensitivity at a -0.5 EV and it seemed to focus fast enough tonight at a Colt League field hereabouts. I have NO complaint on the AF speed. just the accuracy. at full aperture. and close up.
As a side note the focus point seemed to be a few inches in front of the players at the plate.I guess six to twelve. hard to say accurately from thiry feet away. this works okay when it is  pitcher moving into the focus point. As you say at 30 feet the DOF is more accomodating. but the focus point still aint where I locked it.

So, Jonathon is the only person that knows how to properly operate and EOS system camera? My only real "expectation" for this camera, is that it would perform as well as my Film EOS cameras, such as focus where I put the active focus point.
You keep harping on adjusting the lenses, I dont see an adjustment screw on any of them. My guess is that any "adjustments" will be made to the sensor mount in the camera. something that should have been done before shipping!
as for the QC problem it may only be one container load or the units from a particular bench. If you knew anything of manufacturing, you might get the idea.

Boku: if you have something germain to this discussion, add it in.
If you are just trying to pick a fight with me about how I characterized a fallacious point by someone else, butt out.

[drh681]

hmm...
compare the zooms at 50mm...

okay, but why a magic number?

they,the three that I have checked, all front focus at both ends of the zoom range.

:unclesam:

have a good'n!

[Ray]

Very nice! But this has nothing to do with your current focussing problems, does it!  

You can download the charts of B&W lines here

[timp7910]

Hey drh681,
  Take heart in that i had exactly the same problem with my 20D with the IS kit lens. It only took me one return to get it right.

You can tell the focus point by shooting a page of newspaper text at 45 degrees with tripod,  mirror lock, f9,.... the usual.  I can post my tests if necessary, both the first and the second camera.  

I feel for you after seeing this post.  MANY others have posted similar problems in other forums for the 20D and the EOS1D.  However, you will get many replies in the categories of: 1) Here is how my 20D works and it is great (you must be an idiot), or 2) It is a digital feature that only a newbie would be bothered by since you obviously don't use USM (you must be an idiot)  3) You just don't know how to take a photo, hold a camera.... (you must be an idiot).    

[ The parenthetical comments are what the repliers are thinking.  I think your question is a sensible one. ]

Good luck to you.  I love my SECOND 20D. Especially, the IS lens.  This is my first really good DSLR, if you don't count the D70 i had for a while.  I just wish i could get my closet of Nikon ED lenses to work with it!  

Tim

[lester_wareham]

I am amazed at the presumptions people make in response to reported problems with "their" favorite brand of whatever!

EOS 20D-c is at Canon now for "adjustment" I also left a brand new 100ƒ2.0 and year old 50ƒ1.8 for "calibration" (which I think they recommend to cover the cost of the "adjustment"! )

Like I say no offence.

Are Canon charging you for this? I thought they did this sort of thing under warranty.

[drh681]

here is one of my 'field test' images from before the adjustment by Canon. center AF selected and placed on center box:



and here is after the adjustment:



and a 100% crop of the after: