EOS 20D - focus error

I've seen in a few other forums mension of 20D focus errors.
I've noticed a lack of focus in most shots, when using all my L lenses wide open.

I did a little test and it seems 5mm too short, at 70mm on the 24-70

Which is why when I put the centre AF point on the eye (my wife and cat are bored of that game now) the iris isn't sharply focussed.

Can I ask for a second opinion please?
Let me know if you want the full resolution multi-megabyte versions of these reduced images ...

#1 a simple tool !
#2 the ruler is at 45 degrees (so is the tab) focus seems best at 7mm on the ruler which is 4.9mm in real life
#3 it looks like about 5mm to me in the head-on shot too

Paul, looks like a bit of back focusing to me.

thanks, I thought so. I assume it's a factory adjustment.
Don't want the body going back to Canon now because I have a media pass for the NEC car show this week and then F1 nest week.
I'll just have to try to manual focus for driver portraits.
Then see if I can get the body swapped at the shop afterwards.

They were good with the 10-22 EF-S that wouldn't go to manual focus when you work the switch.
Exchanged it on the spot in about 10 minutes (writing down serial numbers, boxing the bad lens).

If it is anything like what happened in the beginning with the 10D and back focusing, it took, taking the cameras into Canon with the lenses and having them adjust based on the camera, to get them to focus properly. It really did cause a real headache.

One thing I would suggested maybe is increasing the light on your test, using the centre focus point only and see what your results are based on that as well, before getting too carried away with trips to Canon Repairs.

I was using the centre focus point only, so I did one thing right :)
I really hope it's not a light issue, as this weekend is an indoor event (car show).
I specifically bought 2.8 L to give me a good chance of working indoors.

A friend has a 20D as well, and his seems to be focussing about 1cm too close with the kit lens. When we get together this weekend we can do some tests with all our lenses on both cameras.
10-22 L, 24-70 2.8 L, 70-200 2.8 L, 18-55 kit lens and two bodies should give us a good set of photos for assessing the problem.
The fact that I can get better focus on the screen (despite poor accomodation and double vision) means it's a sensor problem I think, as it affects all lenses. Unless there is an interface from camera to lens that is analog (not digital) and there is a problem on the body contact.

Quote:

---

Originally Posted by Flashram_Peter_AUS

If it is anything like what happened in the beginning with the 10D and back focusing, it took, taking the cameras into Canon with the lenses and having them adjust based on the camera, to get them to focus properly. It really did cause a real headache.

---

Not to start an arguement or hijack the thread, I don't really think this is a good solution. Reason I say this is that what do you do when you put the lenses on another body? Are they now "off" with that other body? I'm curious, I've heard of the problem before but never heard what the fix was (or never really thought about it if I did).

First I want to know if they are off with that body, or not.
If the focus is fine on his body, then it's my body and not all 3 lenses.
If his kit lens is OK on my body, then I have 3 off-focus L lenses.
If his kit lens is off-focus by the same amount, it's my body.

I need to get more measurements, in brighter light, to see if Peter is right about the light level.
But since the sensor is reliably getting focus lock, I don't think it's a light problem.
But I'll prove it, rather than guess :)

Listen, I'm going to sound like a jerk here so brace yourself....

If you have a concern about the performance of your camera, please send it to Canon for evaluation. Provide them with whatever information and photos you can to describe your problem (like you've posted here).

All we're doing in this thread is making assumptions. Sure, they make sense. But that doesn't mean they're correct. Canon is the one that's supposed to know things for sure.

Expect to send in the lenses that exhibit the problem along with the camera, I've had to do this with my 10D. May I also suggest you send in your entire line of lenses at the same time for focus evaluation with the camera. I need to do that myself sometime as the focus of my 10D is horrible with almost all my lenses.

When I sent mine in a saw improvement afterwards. But mine was much much much worse than what you're showing here.

Oh, while I'm at it, there's another cople of popular assumption we make which is that the focusing "eye" is PERFECTLY in line with the focusing square shown in the viewfinder (and in the screenshots you show here). Couple with that is the assumption that the autofocus uses an area the exact same size as the little box which evidence suggests is highly inaccurate. It seems to use an area much larger than the box. And who's to say that it's going to use the exact center of the box and not one of the outer corners? See what I mean?

I hate Canon's stupid autofocus system.....there, you busted me with my big axe to grind :)

Quote:

---

Originally Posted by another view

Not to start an arguement or hijack the thread, I don't really think this is a good solution. Reason I say this is that what do you do when you put the lenses on another body? Are they now "off" with that other body? I'm curious, I've heard of the problem before but never heard what the fix was (or never really thought about it if I did).

---

I agree with Peter and think he should send them all in if he's really concerned.

And you might very well be right that those lenses would then perform differently on another camera. It's a completely valid concern that you brought up. For me, I need the lenses to work on my main camera so I'd send them in with my main camera. I'd worry about focus problems one another camera if I ever get to that point of having another camera.

PS. I hate canon's autofocus system :)

PSS. Please place emphasis on my previous statement :p