AF Micro-Adjustment tool

[Bart_van_der_Wolf]

Bart,

Can you perhaps consider editing / supplementing the first posting with the various recommendations and tips that have come up in the forum since your excellent first port over a year ago ?

I have now got lost as to what is true best practise using your LCD chart...

Hi Peter,

In post #68, Nicolas Claris already made a nice summary. Otherwise it's simple for most to use it. Just make sure you display the target at a 100% zoom setting (true size), otherwise it will be hard to get the moiré showing. I suggest to calibrate at the most commonly used shooting distance if possible, because there may be differences between close and far focusing.

The easiest way of determining the amount and direction of offset to be corrected for, depends on the lens model. The ones with an internal focussing mechanism won't show a focus ring rotation, so you'll have to pay close attention to either the lens window with the focus distance indicator, or just note which direction you need to turn the focus ring after AF to achieve maximum moiré.

In general I find it easy to focus manually with Live View and then check which direction the AF thinks is better, then correct until Manual and Autofocus agree. However, when the AF is a bit twitchy before it settles, it may be easier to first AF and then use Live View and note which manual focus offset direction improves the visibility of Moiré. Either way, after adjustment there should on average be no substantial bias towards front- or back-focus anymore. There can still be some variability around the optimal AF setting, but that's normal for the phase based AF-technology. AF is not 100% accurate and repeatable, but on average it's very good.

Just remember, you are trying to reduce the bias in focus offset between AF and manual Live View. Whatever method works for you to achieve that goal, is the best one. Also, don't worry about damaging anything, you won't, and you can always get back to zero adjustment for an out-of-the-box situation.

A final note. I have no personal experience with the newest contrast based type of Live View, so I have not described a procedure to exploit that, if possible. It will probably allow you to see the maximum moiré that can be achieved when it is of the small, mottled, kind (don't know what happens with the color band kind). However, I suspect that it is still easier to check the offset by manually rotating the focus ring in Live View after initial AF (that will give instant feedback as to the direction of the correction needed).

Bart

 

[Rod Witten]

Remarkable tool !!!

Many thanks for your effort.

Rod Witten

 

[Bart_van_der_Wolf]

Remarkable tool !!!

Many thanks for your effort.

Hi Rod,

My pleasure. I needed it myself, because the other tools were not accurate enough, and I decided to share it with everybody. It is based on a solid physical phenomenon which we usually attempt to avoid. I decided to let it work for us, instead of against.

Bart

 

[nicolas claris]

Caution !

Hi folks
Getting prepared for some upcoming important jobs, and founding that the latest shots with the 24-70 were not crystal sharp, I checked the focus using Bart's method and targets and found the settings not to be accurate anymore…
I then checked all my lenses (14 to 500…) made corrections after an overcheck with "natural shots' (not the target but something real).
I must say that the "real life" tests confirmed the accuracy of corrections made with Bart's target :-D

They all needed some adjustments and are all now ready for a new start.

I just wonder what could make the old settings becoming bad (used same method before, and got most of the shot tak sharp…).
Vibrations during transportation? (camera never felt down, nor the lenses).
Any idea?

 

[Asher Kelman]

I just wonder what could make the old settings becoming bad (used same method before, and got most of the shot tak sharp…).
Vibrations during transportation? (camera never felt down, nor the lenses).
Any idea?

Nicolas,

This shift is a big puzzle. Of course, you didn't upgrade the firmware in between then and now?

Asher

 

[nicolas claris]

Nicolas,

This shift is a big puzzle. Of course, you didn't upgrade the firmware in between then and now?

Asher

Hoops! damned, you nailed it!

 

[Bart_van_der_Wolf]

Hoops! damned, you nailed it!

Hi Nicolas,

I'm wondering about that. The latest firmware upgrade for the 1Ds3 supposedly didn't change much other than add some Diagnostic ERR codes. I checked it, and it didn't affect my camera/lenses, or maybe I didn't notice it yet. Anyway, we don't know what other undocumented fixes were in the update, so it is always prudent to check AFMA after updating.

I'll do a more thourough checkup of all my lenses when it is serviced for the ERR 99 and potential Focus issues. Canon organized for a free pickup/service/delivery later in April'09. It's supposed to take less than 1 week.

Bart

 

[nicolas claris]

Bonsoir Bart
have you got any problem with your 1DS3 (err99 or focusing issues…) or do you send it just for "safety"… ?

I neverhad err99 problem, and when the micro focus adjustment is made I find the images to be quite sharp.
I haven't got any problem with the focusing system too… Am I missing something?

 

[Jack_Flesher]

Great tool Bart!

Couple of points worth mentioning though for folks to consider as they adjust their lenses...

1) Aperture choice can cause changes to precise focus; commonly called focus shift.

2) Precise focus can also change with subject distance, though this effect is generally less than aperture shift.

As such, I advise you first choose an aperture nearer where you normally shoot the lens, but still fairly open as DoF often mitigates the net effect of focus shift at apertures beyond f4 (in 35mm and smaller formats). I choose f2 or f2.8 for most of my faster prime lenses and f4 for the zooms. I find that this gets me more than close enough at the wide open aperture. Next, choose a distance that is at the closer end of your normal shooting distances where you want the best critical focus for the lens in question.

For example, with my 50/1.4, I chose f2.8 and 6 feet distance. This gave me a different net focus offset than f1.4 at 3 feet -- so f1.4 at 3 feet slightly back focuses by about 1/2 an inch, but this is preferable for my typical shooting needs than the reverse, where f1.4 at 3 feet is perfect, but f2.8 at 6 feet back focuses more notably by about 1-1/2 inches, since I normally use that lens in the 4 to 8 foot range.

FWIW I do not use live view to confirm, but rather zoom in on the raw review as I find that more reliable to critically evaluate.

Cheers,

 

[Asher Kelman]

Great tool Bart!

Kudos well-deserved!

Couple of points worth mentioning though for folks to consider as they adjust their lenses...

1) Aperture choice can cause changes to precise focus; commonly called focus shift.

2) Precise focus can also change with subject distance, though this effect is generally less than aperture shift.[/QUOTE]

Let me just mention, to be complete, an OFF TOPIC but related pitfall, for the less experienced who focus and recompose. Here, the term "focus shift" is also used where one focuses on say the eyes and then recomposes on the entire face, thus swinging the optical axis of the camera, (say, from one angular position up high to a new position, at right angles to the plane of the subject) to where the distance from lens to the subject is now closer. This, of course, throws the plane of focus backwards and the eye will now be out of focus to a degree. This method of focus, recompose, should only be used when one can't focus well in dim light and one has to choose another object, say a spoon on the table, to grab focus. However, in this circumstance, it's a trial and error labor or persistence and luck.

So, if one focus and recomposes and the image is not sharp, don't recalibrate your camera's focus by adjusting settings!. That back focus, then, is just the consequence of geometry. Any angled line to a surface is always longer than the shortest distance, which is, of course, perpendicular. To avoid this, learn to use selection of the focus point so that it lies over the eye, for example, that you want in focus.

Asher

 

[Jack_Flesher]

Good point Asher!

I meant the optical property of a stationary lens' focus shifting as aperture changes. This is typically most prevalent with fast, spherical lens designs, but also presents to some degree in aspheric designs too.

The big issue here is our DLSRs AF at the lens' max aperture -- and even if that is precisely on, the property I am referring to will shift that exact focus point as the lens stops down -- hence my recommendation for making the settings slightly stopped down from wide open.

Cheers,

 

[Bart_van_der_Wolf]

Bonsoir Bart
have you got any problem with your 1DS3 (err99 or focusing issues…) or do you send it just for "safety"… ?

I never had any problems, ... until the week that Canon offered the possibility to fix, if broken, I got my first ERR 99 ever! As if there was a timer in the camera that set it off. It has happened 3x since, all resolved for the moment by removing the battery and reinserting. The latest one was with the updated firmware, but it was still a 99 instead of something more descriptive.

I filled in the Web form for the ERR 99, and received a phone-call from the Dutch service centre for an appointment to pick up the body at my address by courier. After a small exchange of info they agreed to also look at the AF in addition to the mirror/ERR 99 issue. I didn't run into a focus issue that could be attributed to a malfunction, but then I also don't use the outer AF sensors as much as the center one, or Live View. Now that the occasion presented itself, I'll have them look at it rather than detecting something later and having to sent in the camera again.

Bart

 

[nicolas claris]

I got my first ERR 99 ever! As if there was a timer in the camera that set it off.

Hmmmm… how many shutter actuations ? ;-)

 

[Bart_van_der_Wolf]

Hmmmm… how many shutter actuations ? ;-)

Not that many, a couple of thousands (I lost track due to loading an older card with a higher filenumber). But the modest number of actual actuations is certainly not what caused it (whatever the error actually is, presumably something with the mirror). It was sub-zero Celcius when it first happend, maybe the freezing temperature had something to do with it.

I know, it's a bit unsettling, you don't want to experience it mid-session (as I did), although once you know what to do about it (power Off/On) it's less of a total show-stopper but merely an annoyance. I'll let you know what they say they did with my camera (if there is any detail given) when I get it back. Otherwise you might want to consult with your CPS contact if you can have the fix(es) done as a precaution, before the season really starts.

Maybe it's related to the AFMA topic at hand, but I'll certainly check after the repairs for new settings that may be needed. It's interesting, because they will probably need to calibrate the body AF when they replace parts associated with the mirror (which the firmware alone apparently couldn't solve). I therefore expect something to change, we'll see.

Bart

 

[Michael Southcott]

Bart,

I've tried to get this to work on a 50D ... and failed miserably !!
My main lens is a 500F4 tried varying distances upto 40' and still could not see the moire !!

1. Is it possible on a 50D ?
2. How critical is distance and roughly how far away for the 500 ?

Mick..

 

[Bart_van_der_Wolf]

Bart,

I've tried to get this to work on a 50D ... and failed miserably !!
My main lens is a 500F4 tried varying distances upto 40' and still could not see the moire !!

1. Is it possible on a 50D ?
2. How critical is distance and roughly how far away for the 500 ?

Mick..

Hi Mick,

I'm not sure what would be the best distance to try with that (lens/camera) combination. Do make sure that the test target is displayed at 100% zoom setting (=actual pixel size) on screen! Maybe the 500mm (as do the 600 and 800 and 1200) require a modified target, but then they would also need a larger LCD to shoot from a typically large distance (I advise to calibrate close to the common shooting distance, if practical).

For other lenses it seems to work okay if the displayed target fills the central part of the focus screen. Shooting it too close will prohibit the moiré to manifest itself, but an off the cuff calculation suggests at least 20-25 metres (some 80-100ft) could work if at all practical (it would require a very large room ...). Otherwise, just use a distant isolated (flat) surface for calibration of these long FL monsters.

Bart

 

[Marcel Burkhalter]

hi guys

amazing forum and a very interesting thread! i'm really thrilled about microadjustment because i like my images crispy sharp

i had some serious trouble focusing to the point where most moire is visible.
spinning the foucus ring manually by just a fraction of a millimeter made a big difference. this makes it real hard to get comparable results.
so i made some changes to the described procedure:

- manually focusing to infinite
- activating live view with auto focus mode "live view" (not quick!)
- pressing the "AF" button (on the battery grip -> less shaking) until perfect focus is reached (takes approx. 2-4 seconds)
- deactivating live view
- pressing the shutter half way down and watching the focus ring movement closely
- if it moved, setting microadjustment accordingly
- repeat above steps until absolutely no movement is visible
- cross checking with other focal lengths, maybe make a compromise...

with no need to manually finding the perfect focus point (most moire) this procedure is a lot more reliable and reproducible.

i tested the settings under real world conditions (shooting in a nature park) and it looks good. my 70-200mm is definitely a bit sharper (setting: -5). my 24-70mm didn't need microadjustment.

as far as i can tell, the camera uses some sort of software autofocus algorithm when using live view with autofocus mode "live view". there is no mirror flip (as with setting "quick", which is really a lot faster), so the cam MUST be using the image from the sensor to autofocus (and not use the autofocus hardware).
i always had the impression, that the focus obtained this way is *perfect*. the slightest touch on the focus ring in any direction made the moire less visible.

i'm really interested in your opinions as i'm far away from being a professional photographer.
what do you guys think? is this a good way to perform microadjustment or did i miss something?
does the pro line (e.g. 1ds mark3) also have this "live view" AF mode?

@bart: is it possible for you to create different test images (i don't know, e.g. other frequency/amplitude) to compensate the effect, that some screen/lens/distance/focal length combinations produce only a little moire?
my 70-200m was easy with the test image on my macbook. the 24-70mm showed very little moire on the macbook, on my 24" LCD screen the moire was a lot better, but the distance is a problem in my home office...


kind regards
marcel

P.S.: some info on the used equipment: EOS 5D II (latest firmware) with battery grip / 24-70 2.8L / 70-200 2.8L (2x extender adjustment pending).

 

[Bart_van_der_Wolf]

hi guys

amazing forum and a very interesting thread! i'm really thrilled about microadjustment because i like my images crispy sharp

i had some serious trouble focusing to the point where most moire is visible.
spinning the foucus ring manually by just a fraction of a millimeter made a big difference. this makes it real hard to get comparable results.
so i made some changes to the described procedure:

- manually focusing to infinite
- activating live view with auto focus mode "live view" (not quick!)
- pressing the "AF" button (on the battery grip -> less shaking) until perfect focus is reached (takes approx. 2-4 seconds)
- deactivating live view
- pressing the shutter half way down and watching the focus ring movement closely
- if it moved, setting microadjustment accordingly
- repeat above steps until absolutely no movement is visible
- cross checking with other focal lengths, maybe make a compromise...

with no need to manually finding the perfect focus point (most moire) this procedure is a lot more reliable and reproducible.

Hi Marcel,

Yes, with more recent incarnations of LiveView it is possible to perform an AF in LiveView mode by using a slower 'contrast detection' method. The earlier versions of LiveView didn't have that option, so it wasn't included in the initial AFMA 'workflow' suggestions. Thanks for adding that suggestion.

i tested the settings under real world conditions (shooting in a nature park) and it looks good. my 70-200mm is definitely a bit sharper (setting: -5). my 24-70mm didn't need microadjustment.

Contrast detection AF is the most accurate method of determining focus, but it is also relatively slow. So it would be less suited for dynamic scenes, but very well suited for stationary scenes. However, for AF microadjustment it is important to use a suitable (=Flat) surface of adequate dimensions as a target. A non-flat target may confuse the phase AF, while the contrast AF may arrive at a correct setting. The difference would in that case be misleading, and correcting for it may lead to sub-optimal AFMA results.

as far as i can tell, the camera uses some sort of software autofocus algorithm when using live view with autofocus mode "live view". there is no mirror flip (as with setting "quick", which is really a lot faster), so the cam MUST be using the image from the sensor to autofocus (and not use the autofocus hardware).
i always had the impression, that the focus obtained this way is *perfect*. the slightest touch on the focus ring in any direction made the moire less visible.

Yes, it's the most accurate focussing method, but not all camera models have that option.

i'm really interested in your opinions as i'm far away from being a professional photographer.
what do you guys think? is this a good way to perform microadjustment or did i miss something?
does the pro line (e.g. 1ds mark3) also have this "live view" AF mode?

There's nothing wrong with your method, just make sure that you use a flat target surface, perpendicular to the optical axis, and it should cover an area in excess of just the central red 'active' AF marker in the viewer. The 1Ds3 (which I use) doesn't have a contrast detection AF mode.

@bart: is it possible for you to create different test images (i don't know, e.g. other frequency/amplitude) to compensate the effect, that some screen/lens/distance/focal length combinations produce only a little moire?
my 70-200m was easy with the test image on my macbook. the 24-70mm showed very little moire on the macbook, on my 24" LCD screen the moire was a lot better, but the distance is a problem in my home office...

I could make other versions of the original target, but the difficulty with the 24-70mm is that it is probably less sharp @ 70mm than the 70-200mm is. The moiré only shows up clearly with sufficiently sharp lenses. The 24-70 after postprocessing/sharpening is quite good, but that won't help at the initial capture or calibration stage. What you could try, is using the alternative target I mentioned early in this thread (post #16).

The long distance calibration will remain a problem unless one lives in a large palace with huge LCD screens ;-)
In that case use the contrast AF method on a flat wall with lots of detail. It's preferable to calibrate at a common working distance, which is usually far away with long telelenses and closer by with shorter lenses (but there are exceptions).

Bart

 

[nicolas claris]

The long distance calibration will remain a problem unless one lives in a large palace with huge LCD screens ;-)
In that case use the contrast AF method on a flat wall with lots of detail. It's preferable to calibrate at a common working distance, which is usually far away with long telelenses and closer by with shorter lenses (but there are exceptions).

Bart

LoL!
I have a huge (well 30 inches only ;-) monitor but not a large palace!

I received yesterday, back from Canon France my IDs3, as they have changed to my request the mirror box. Yes I had some AF issues…

As a matter of fact, I had to redo all micro adjustments, none of the previous were accurate anymore…

So my findings with your method Bart, is that it works beautifully with wide angles, but I get more consistent results (because of my small palace!) when shooting the parking lot in front of our office, (then compare + and - focus on target) for the 24-70 (at 70 of course!) 70-200 and 500mm…

The camera is still sharp, I just hope that now I won't get that AF issue with fast moving and shaked objects such as yachts at sea…

 

[Michael Durisseau]

Mr. Van der Wolf,

What about some kind of hardware adjustment? I know the lens mount has some kind of shimming that, if those parts are readily available, could help focus, yes? It would be better if you had to remove some of the shims rather than add...

 

[Bart_van_der_Wolf]

Mr. Van der Wolf,

What about some kind of hardware adjustment? I know the lens mount has some kind of shimming that, if those parts are readily available, could help focus, yes? It would be better if you had to remove some of the shims rather than add...

Hi Michael,

You can call me Bart!

A hardware adjustment would help, whether by shims or by tweaking the mirror path lengths, but it's not necessary on cameras that offer an AF Micro-Adjustment feature. In fact, it is pretty hard to do by adjusting hardware alone.

The camera sends commands to the lens focusing motors, and if we can intercept those commands and adjust them before execution, then the focus will be correct afterall. The only shortcoming of the single offset adjustment is that it usually requires an offset and a slope adjustment to calibrate an otherwise linear process. So the current AFMA may or may not work well at different distances, it would require a trip to the official service point to get that calibrated (to within their tolerance range).

Cheers,
Bart

 

[Michael Durisseau]

Hi Michael,

You can call me Bart!

A hardware adjustment would help, whether by shims or by tweaking the mirror path lengths, but it's not necessary on cameras that offer an AF Micro-Adjustment feature. In fact, it is pretty hard to do by adjusting hardware alone.

The camera sends commands to the lens focusing motors, and if we can intercept those commands and adjust them before execution, then the focus will be correct afterall. The only shortcoming of the single offset adjustment is that it usually requires an offset and a slope adjustment to calibrate an otherwise linear process. So the current AFMA may or may not work well at different distances, it would require a trip to the official service point to get that calibrated (to within their tolerance range).

Cheers,
Bart

Hey, Bart,

That't why I was asking about the hardware adjustment...some of us have come to the end of our adjustment with the M/A. I thought it would be another do-it-yourself option.

Thanks for the response!

p. s. Are you aware of a service manual for the 1d classic? A broken one (as long as the dial on the back still works!) would be OK...

 

[Bart_van_der_Wolf]

Hey, Bart,

That't why I was asking about the hardware adjustment...some of us have come to the end of our adjustment with the M/A. I thought it would be another do-it-yourself option.

Thanks for the response!

p. s. Are you aware of a service manual for the 1d classic? A broken one (as long as the dial on the back still works!) would be OK...

Hi Michael,

I've seen some descriptions on the internet about a small hex key being needed to adjust the angle of the main mirror (with a small bolt in the back of the mirror box). That would adjust the total path length for the secondary AF mirror as well. I don't have a (link to a) service manual for the 1D, maybe others have an idea. A full service calibration is supposed to also tweak the lens settings table (in camera and lens firmware).

Cheers,
Bart

 

[Isaac Burdeinick]

Trying on Tamron 18-270mm

Hi Bart

Im going to try The Micro Adj. on the far end of Tamron 270mm @F6.3. For that need to do it at least at 13.5 Meters! that is a hassle... any suggestion? can it be done at less distance? I know that Canon tips say at 50 x the focal distance.
What would you do?
Keep it at -5 for the 18-50mm end? or change it to +3 for the Far end, Go back & forward...

 

[Isaac Burdeinick]

Here's some very good news. This test with a 1MIII with the focus fix from Canon shows tack sharp auto focus with the three lenses I have with me (16-35L, 24-105L, and 70-200L). Earlier tests with other means showed some adjustment needed before the camera was sent in for repair.

Conduction the test as described produced such fantastic results that I felt a more rigorous test should be done. I devised a two step test:

Step 1: Do exactly what has been described here.

Step 2: Do not move camera.

2.1: While not in Live View, but in Auto Focus mode, manually rotate focus ring to de-focus the lens.

2.3: Press shutter half way down to actuate auto focus.

2.4 Without moving anything, enter Live View.

If the moire lines are still there, the auto focus returned the lens to tack sharp focus.

This part of the test should only be done after the user is satisfied with any adjustments made.

I added this test to see if the lens motor and or auto focus circuit has any hysteresis (slop in returning to same point). I am delighted to report that all works fine.

The only question I have is that Canon recomends that the tests be made at a "distance similar to actual subject matter you shoot . . ." But the argument against that is that Canon doesn't and can't do it that way themselves in their service center.

Time will tell how all this works out.

Wil

Hi Will

As I Understand you have to Set The AF in Manual and then change it to Auto, Otherwise you could damage the fine moving mechanism in the Lens.

 

[Doug Kerr]

Hi, Isaac,

As I Understand you have to Set The AF in Manual and then change it to Auto, Otherwise you could damage the fine moving mechanism in the Lens.

Not on any of the lenses Wil mentioned, which all have "full-time manual focus".

 

[Doug Kerr]

Hi, Bart,

I've seen some descriptions on the internet about a small hex key being needed to adjust the angle of the main mirror (with a small bolt in the back of the mirror box). That would adjust the total path length for the secondary AF mirror as well.

The adjustment I think you are speaking of isn't intended for correction of AF errors, and moving it will shift the alignment between the finder image and the taken image (which is what it is intended to adjust).

Indeed moving it will also (as a side effect) change the AF calibration, but differently for the different AF points.

It's best to reserve it for its intended purpose.

Best regards,

Doug

 

[Bart_van_der_Wolf]

Hi Bart

Im going to try The Micro Adj. on the far end of Tamron 270mm @F6.3. For that need to do it at least at 13.5 Meters! that is a hassle... any suggestion? can it be done at less distance? I know that Canon tips say at 50 x the focal distance.

Hi Isaac,

I'd do it at 270mm, and AFMA at something like 4 or 5 meters if possible indoors. The suggestions to shoot at 50x FL are old school. The number mentioned in the Canon tips by Chuck Westfall were based on shooting resolution targets on film (and maybe to calibrate a lens by itself). Film has a much more gradual and dithered fall-off towards limiting resolution. The absolute resolution of most DSLRs is lower than of fine grained film, and drops of fast once it approaches the limiting resolution (Nyquist). Also, when using my AFMA target moiré method, we exploit some properties of sampled imaging at 2 different sampling densities, so that's something different alltogether.

The only uncertainty with the Tamron is that at smaller apertures (e.g. f/6.7) there may be too little modulation to generate a useful moiré pattern. But since it's free, there is no harm in trying.

As always, I do suggest to check actual focus performance afterwards at longer distances (if you use the lens for such work) on a flat surface with some surrounding slanted areas to detect front/back focus. Always remember that the focus sensors are larger than the indicators in the viewfinder. When lenses use internal focusing, there could theoretically be different AFMA settings at different distances for a given focal lenght, so checking for that won't hurt and strengthens confidence.

What would you do?
Keep it at -5 for the 18-50mm end? or change it to +3 for the Far end, Go back & forward...

I use the settings for the more critical longer focal length myself (because DOF increases towards short FLs anyway), but when there is a significant difference throughout the range, I'd either choose the most used FL, or select a compromise setting.

Cheers,
Bart

 

[Bart_van_der_Wolf]

The adjustment I think you are speaking of isn't intended for correction of AF errors, and moving it will shift the alignment between the finder image and the taken image (which is what it is intended to adjust).

Indeed moving it will also (as a side effect) change the AF calibration, but differently for the different AF points.

It's best to reserve it for its intended purpose.

Hi Doug,

Sure tinkering with these things is something that is not for the faint of heart, but if it gets the job done ...
I don't know if the adjustment only changes the angle of the mirror, or if it changes the fore/aft-position of the hinge. Surely Canon service should be able to adjust these things, maybe in the same way.

BTW, this is the article I was referring to: http://astrosurf.com/buil/autofocus/adjust.htm

Cheers,
Bart

 

[Doug Kerr]

Hi, Bart,

Sure tinkering with these things is something that is not for the faint of heart, but if it gets the job done ...

My concern is that a change made to the mirror stops may well compensate for another problem and give improved AF accuracy (at least for one of the AF detectors), but might leave the user with an entirely different problem that may not be noticed at first (namely vertical inconsistency between the finder image and the taken image).

I don't know if the adjustment only changes the angle of the mirror, or if it changes the fore/aft-position of the hinge.

In the cameras I am actually familiar with, this adjustment (sometimes there is a pair, one on each side) controls the resting angle of the mirror assembly.

Surely Canon service should be able to adjust these things, maybe in the same way.

Well, they certainly can adjust the mirror stops (presumably to get the resting mirror angle correct).

There are various mechanical "adjustments" (involving changeout of spacers and shims) that will have an effect on AF behavior (as well as on the consistency of finder image vs. taken image focus).

I have no idea how common it is for a Canon Service Center to actually have to change any of those. I'm fairly sure that the most common work done to normalize AF operation with respect to the camera proper is done by "software" changes made to tables in the camera.

Best regards,

Doug