Back focus to heck and back on 60D

[Ezookiel]

Very much "not happy Jan"
Been driven mad trying to work out why all my photos on the 60D with 18-200 are so OOF. Today in desperation I downloaded a focus chart, went through all the rigmarole, and find the 10mm-14mm mark is the most in focus (can't be totally sure till I get them off camera back at home). That's a stupid amount to be out by.
I could imagine some coming out of the factory a little out, but that amount seems ridiculous.
Lucky I took out a "loan camera while being serviced" option on the warranty. I have a big trip this coming weekend.
Grrrr.
At least I now know why I've had to try to use so much sharpen on so many pics, and had to totally discard some that I just couldn't save.

[salty]

Did you check the lens reviews on http://www.dpreview.com/ ?

[arthurking83]

According to DPR the lens seem to offer some very good IQ.

Ezookiel, have you checked the lens at many focal lengths, or just at 200mm?

[larrywen]

I am using this lens on my 7d and it works reasonable well.

is it possible to try your lens on a different body or try another lens on your 60D to find out which part goes wrong?

[Bennymiata]

I have 7 different lenses for my 60D, Canon L's and EF-S, as well as a Sigma and a Tokina, and they all focus perfectly (as far as I can tell at any rate).

Have you tried any other lenses on your camera?

[Ezookiel]

Sorry about the delay getting these up. 28 hour shifts, and the subsequent exhaustion, don't lend themselves to getting much done some days.
Here are the results of the test shots.

[Ezookiel]

Taking the whole lot into the camera store today to see if they can verify the same issue. My test shots on the chart were all done at widest zoom (technically 18mm I'd guess), and following instructions using only centre focus point, 45 degree angle, tripod, remote shutter release, etc, and the photos came out with the rear 10-14mm lines in focus instead of the correct one.
Could explain why I've had a biatch of a time getting flower photos to work unless I manually focus. If I thought I had the stamen in focus, the pic would turn out with something else in focus, often NONE of it in focus. It's made learning a right royal PITA because I kept thinking it was something I was doing.

No other lens to try as I only own one so far. No other body to try for the same reason. Anyway, will let the experts in the shop take a look and see what they think. That reminds me, I need to get the relevant chart pics off the camera to show to them.

[larrywen]

sorry to see the test chart. It's obviously not focusing correctly. Hope the store can do something for this issue. Or have you tried to macroadjust the lens with your camera. I know 7d has this feature, but not sure if 60d has this feature as well.

best wishes

[Lance B]

Hmm. Looking at the posted charts, where the AF point would be, which is the centre of the photo, this is not where the AF point should be. I think you need to re do the test so that your central AF point is aligned with the "focus here" wording on the test chart. In fact, using a ruler from corner to corner on my monitor screen and finding where they intersect, at that point the writing appears to be the area that is best in focus. So, I am wondering if you have focused on the wrong spot?

[agb]

Hmm. Looking at the posted charts, where the AF point would be, which is the centre of the photo, this is not where the AF point should be. I think you need to re do the test so that your central AF point is aligned with the "focus here" wording on the test chart. In fact, using a ruler from corner to corner on my monitor screen and finding where they intersect, at that point the writing appears to be the area that is best in focus. So, I am wondering if you have focused on the wrong spot?

If you have any doubt about where the af point was then look at it in DPP with a quickcheck where you can set it to show the af point, and then be doubly sure that it is where it was supposed to be.Sometimes I am surprised myself that I think the af point is right where I wanted it to be, only to find that I had missed the spot.

[arthurking83]

Hmm. Looking at the posted charts, where the AF point would be, which is the centre of the photo, this is not where the AF point should be. .....

I have a feeling that this is a crop of the original image.

How far from the focus chart did you set up the camera?
What focus mode did you use?

60D has LiveView mode.

a couple of points to note:

Focus mode should be in single shot (with confirm beep) mode. Confirm beep is optional, but make sure you are not in Continuous focus mode(I think what Canon call servo mode).

Focus and confirm and then take the shot sort of workflow.
If you still get the same problem, then try it in LiveView Mode too.
Using liveview, focus in the same manner and does this make any difference?
If not, then there is a very high probability that you do have a defective lens(in focusing).
If there is a difference, what and how much.

Live view will eliminate any possibility that the camera is at fault!(that is the camera's focusing system).

How much difference is there in the focus accuracy at other focal lengths.. eg. 100mm and then 200mm.

Personally I don't particularly like these kinds of charts and methods of focus testing. They can be prone to errors... not always, but there is a possibility.

I prefer to use real targets spaced out fore aft of each other at appropriate distances.

That's not to say that this method is incorrect in any way. it's a quick way to see if there is any problem as a preliminary test, to do further testing.

If you did in fact focus on the Focus Here line, then and there was no possibility of focus error due to the method, then the actual plane of focus looks to be even further behind the 14mm mark, between 14mm and 20mm. DOF seems to be between the 8-10mm mark and the 30-32mm mark.

Using this test method, firstly you need a reference point:
To get a reference point of where focus should be, using manual focus and LiveView zoomed in to the level before 100% zoom level, focus on the Focus Here plane.
This is the most accurate way in which to focus.

Once this is done and the DOF zone is then set out, you should then test for AF accuracy.
BUT!! what a lot of people tend not to do, when you start the AF runs, you need to do several runs of focus testing(to eliminate a single or two run statistical error).
So plan on doing say 5 or 6 test runs, in each direction.
What many people don't do is to defocus between each shot, but in both directions of focus travel too.
That is, if you defocus after each shot(as you need too) do so by defocusing the lens in the forward direction and in the rearward direction.
This eliminates one other problem.

That is, if the lens seems to focus fine when AFing when the lens is focused back of the point of focus, but is still and consistently mis focusing when coming from in front, then this could be a mechanical issue.
If it does so from both directions then it could simply be a CPU-Camera communication breakdown.

One last point, Considering that the camera is set to min focal length, at what distance were you from the focus target?
This is important for two reasons.
One is, that the lens may (mis) focus differently, but not necessarily perfectly accurate, just differently!!
Two is, how close were you to MFD(min focus distance)? Don't get too close to MFD for these tests.

Most important point of all of this, is to be armed with info. The 'shop people' may try to wriggle their way out of any problem products coming back.
Armed with as much info as you can muster, this minimises the possibility that they'll turn you away by taking the camera/lens in the back room and coming back saying that their tech says it's all working fine.
The idea is simply to tell them that you are the tech! That is, that you have years of experience and that with your knowledge of all things photographic and AutoFocusing, this lens is not right.
They need not know the absolute truth, just the truth that they need to know!

[Ezookiel]

These are cropped quite drastically, so the centre of the image is not indicative of where I focussed the camera.
I used the camera on a tripod, next to a kitchen bench, so that the camera was around 50-60cms from the chart.
I didn't use liveview, but want to try again and try some of the suggestions above.
I had the camera set to it's widest angle. I didn't try testing any others. I probably should, but wasn't sure the lens wouldn't slowly extend itself given the downward angle it's tilted at.
I set the camera to AV as the instructions required, and dialled up the lowest f/stop number I could get with the 18-200mm lens.
I then used the viewfinder to check I had the central focus dot dead centre on the chart's central black line, then used the 2second self-timer to activate it (I should have dragged out the remote just to be really really sure).

To get the 45 degree angle, I set the tripod tilt meter to zero, attached the camera, and then adjusted the legs to get the spirit level on it dead centre. This way when I used the tilt meter on the tripod to measure the 45 degree angle, it should have been correct (provided the degrees tilt gauge printed on the Manfrotto is accurate).
I'd like to try it again with some actual angle measurements just to be absolutely sure, but not sure that the angle would make a huge difference, as arthurking83 points out, he prefers real targets set up, which would show up the error even if the camera was dead horizontal, so I'm assuming this chart is still going to show up a focus error even if the angle isn't a perfect 45, but that only the measurements in mms will be out if the angle isn't perfect.

I took the camera to the shop today, and yes there is a loan camera included in the warranty, but only for specific events (you can borrow a camera if you have a major event to shoot, but it's not loaned out for the entire period the camera is being repaired, only as you need it) and worse, they only have a 1000-D or something to lend out anyway, so I'll keep it for this big coming weekend and try to compensate in other ways, and since I'll be with a mass of photographers, maybe get them to test my camera as well.
The guy in the store did take a few quick shots in store of things like the keyboard on the computer, and then zoomed them right in, and he claimed it was focussing ok, but I have no idea where on the keyboard he focussed, and then on which letters he zoomed in, but it probably would show up on a keyboard too if you did it right.

Anyway, after the weekend I'll have a better idea if it's the lens or the body, because I may get a chance to do some lens swap tests with the togs that are coming, and if others that know more about this than I do agree that there's a problem, I'll take it in on Monday to be sent off for repair.

And the 60D apparently doesn't have a micro adjust feature like some of the other models ... bummer!

Thanks for the suggestions everyone, I'll let you know how I go with it all.

[fillum]

I didn't use liveview, but want to try again and try some of the suggestions above.

I'm not familiar with the 60D, but assume that it uses phase-detection for normal AF and contrast-detection for Liveview. I think contrast-detection is usually performed on the image sensor (rather than a separate AF-sensor as for phase-detection). What this means is that you might get accurate focus using Liveview, but still have a real problem with normal AF.



Cheers.

[KeeFy]

These are cropped quite drastically, so the centre of the image is not indicative of where I focussed the camera.
I used the camera on a tripod, next to a kitchen bench, so that the camera was around 50-60cms from the chart.
I didn't use liveview, but want to try again and try some of the suggestions above.
I had the camera set to it's widest angle. I didn't try testing any others. I probably should, but wasn't sure the lens wouldn't slowly extend itself given the downward angle it's tilted at.
I set the camera to AV as the instructions required, and dialled up the lowest f/stop number I could get with the 18-200mm lens.
I then used the viewfinder to check I had the central focus dot dead centre on the chart's central black line, then used the 2second self-timer to activate it (I should have dragged out the remote just to be really really sure).

To get the 45 degree angle, I set the tripod tilt meter to zero, attached the camera, and then adjusted the legs to get the spirit level on it dead centre. This way when I used the tilt meter on the tripod to measure the 45 degree angle, it should have been correct (provided the degrees tilt gauge printed on the Manfrotto is accurate).
I'd like to try it again with some actual angle measurements just to be absolutely sure, but not sure that the angle would make a huge difference, as arthurking83 points out, he prefers real targets set up, which would show up the error even if the camera was dead horizontal, so I'm assuming this chart is still going to show up a focus error even if the angle isn't a perfect 45, but that only the measurements in mms will be out if the angle isn't perfect.

I took the camera to the shop today, and yes there is a loan camera included in the warranty, but only for specific events (you can borrow a camera if you have a major event to shoot, but it's not loaned out for the entire period the camera is being repaired, only as you need it) and worse, they only have a 1000-D or something to lend out anyway, so I'll keep it for this big coming weekend and try to compensate in other ways, and since I'll be with a mass of photographers, maybe get them to test my camera as well.
The guy in the store did take a few quick shots in store of things like the keyboard on the computer, and then zoomed them right in, and he claimed it was focussing ok, but I have no idea where on the keyboard he focussed, and then on which letters he zoomed in, but it probably would show up on a keyboard too if you did it right.

Anyway, after the weekend I'll have a better idea if it's the lens or the body, because I may get a chance to do some lens swap tests with the togs that are coming, and if others that know more about this than I do agree that there's a problem, I'll take it in on Monday to be sent off for repair.

And the 60D apparently doesn't have a micro adjust feature like some of the other models ... bummer!

Thanks for the suggestions everyone, I'll let you know how I go with it all.

One point you didn't reply to. Focusing mode? Set to one shot , ai servo or servo?

[Mark L]

One point you didn't reply to. Focusing mode? Set to one shot , ai servo or servo?

I've been wondering about focusing mode lately, so interested to know why this makes a difference when photographing a static object.

[arthurking83]

I've been wondering about focusing mode lately, so interested to know why this makes a difference when photographing a static object.

Even tho the subject is static, there is always the possibility that if you're using continuous mode AF and continuously focusing, the camera can be misled into believing that it need to focus again.. and again, and again.....

This can lead to very small shifts in focus as the focus system is trying to please the user in getting a 'sharper' image.

This doesn't mean to say that it will happen, only that it can happen and when doing tests(such as these) you want zero room for untrackable error possiblities.

I love AF-C focus mode. I can't think of any situation where I've needed to set the camera back to AF-S mode for any reason, other than testing, or setting it up for someone else to use.
eg. because I use AF-On button, when I give the D300 to my son to use, he can't understand why you need to press this to get that to work, so I have a 'General' memory bank set up in the camera for his exclusive use where the camera is set to work as per normal, and I'll set the focus mode button to AF-S.

[Lance B]

You might want to check your minimum focus distance for that lens as it is 450mm and you are getting close to it on your test.

the 45 degree thing is juts a guide and need not be all that exact.

As Arthur says, he is not so keen on these tests and neither am I. It is for indication purposes only. Anyway, Arthur has written a good dissertation on the matter and I can't really add more other than to say that you should really do the test at say 50mm, 100mm and 200mm as 18mm can be a little too wide for accurate focusing. No zoom lens has perfect focus at all zoom focal lengths and may BF (back focus) at wide end and FF (front focus) at the long end or visa versa or be both the same at each end and correct in the middle. This is the rice you pay for such a long zoom range.

[kiwi]

.... This is the rice you pay for such a long zoom range.

he didnt pay rice he paid peanuts


to the op - are you able to replicate his keyboard test while there at the shop ?

[Ezookiel]

Sorry, pretty sure it would have been one shot. That's what it's mostly set to anyway. I'll have a look and see if it's in the exif data just to be sure.

[Ezookiel]

Well. Now that the 4wd photography trip is over, and my daughter's Formal is over, I finally decided I can do without the camera while they fix it.
The woman at the camera store that sold it to me was quite blown away by the focus chart results. She sent the charts in with the camera, along with 4 or 5 photos still on the camera that show quite clearly the amount the focus is out - one of a tiny little spider on a wooden table, where you can see the centre focus point is on the spider, but the grain in the wood is clearly in focus several cm behind the very OOF spider, and one of 4 drops of water on a branch, where the focus point is clearly shown as the branch, but nothing is in focus because the camera focussed somewhere behind the branch.
Estimate it will be six weeks to get it back at normal times, and probably longer than that over Christmas. Sure hope it's less than that. I'd kind of like to have my own camera rather than a loan camera for the last week in January for the second 4wd photography trip.