[Alex Zhang]

Hi guys,

I recently discovered that my beloved K-5 had developed a strange mis-focus problem: it will either front-focus or back-focus depending on the orientation of the camera relative to the object to be focused on. More specifically, when I am shooting a focus chart lay flat on the desk with an elevation angle of approximately -45 degrees, the camera will front-focus if I hold the camera horizontally and back-focus if I hold it vertically. Here are the test shots I made:

center part of the horizontal shot:


center part of the vertical shot:


All shots were taken with the center point focusing mode. Full uncropped images can be found in #10 reply below.

This same phenomena can be observed on a FA 31m and two DA 35 lenses wide open. All three lenses focus fine on other cameras.

I bought the K-5 about half a year ago. I did focusing tests at the time and it did not have this problem.

I sent the unit to the Pentax repair shop in China and they don't seem to know how to fix it.

Has anyone run into similar problems or know what's going on here?

I'd appreciate any help.

Alex

---

Last edited by Alex Zhang; 10-22-2012 at 06:30 AM.

[Na Horuk]

Ha, that is odd. Did you try the basic first steps, like charging battery, removing battery for a while, resetting to factory defaults, turning off auto-rotate and things that have to do with camera orientation, updating firmware?
The only way hardware could be responsible is if there is a part a little loose in the AF mechanism that tilts when upright or something.

[Tanzer]

The focus points in the K-5 are annoyingly large, larger than the red focus point indicators imply. Your description "center part of the horizontal shot" or "center part of the vertical shot" is a clue to me, that you were probably far back from the focus chart, such that the focus target could see the whole box, and thus picked whichever edge of the box it wanted to. From my experience with focus calibration, you need to be about as close to your target as the images you posted, without any cropping, if I said that clearly.

You must be close enough to the target, of use a large enough focus chart, so that when pointing at a blank area of the center box, you can verify that the camera can NOT acquire focus. If you can't get the camera to focus hunt on the white area, then you are too far away, and/or using too small of a focus target.

The other purpose of this step is to discover exactly where the real center focus target actually is, relative to the red center focus indicator. You'll find that the red indicator does not align with the actual focus point -- they are actually superimposed from different ends of the pentaprism, so they are typically not well aligned. Once you know all this, then you can be sure exactly where your focus target is, and how large it is.

[Alex Zhang]

No, I did not try these steps. Since the camera is still at the repair shop, I can't try them now. I upgraded the firmware to 1.13 a while ago. Don't know whether this has anything to do with the problem. I will give your suggestion about loose part to the repair guy, see whether he can find something along this line. Thanks for the help.

---

Last edited by Alex Zhang; 10-21-2012 at 07:41 AM.

[Alex Zhang]

Thanks for the info, Tanzer. I don't have the K-5 with me now. I have a K-r with me. Does it have similar large focus points? I can give it a try to better understand what you are saying. From my previous tests, it appears the mis-focusing only happens to K-5, but not K-r.

Also FYI, these test shots I uploaded were taken from approximately half meter away. I did try to focus at minimum distance of the lenses, but only horizontally. I also saw severe front-focusing at these close distances. K-r had not problem at any distance with the same lenses.

I am attaching a test shot made at minimum focus distance of the FA 31 lens. This one has no cropping.

[IchabodCrane]

To get a better understanding of what you're dealing with, can you tell us if the problem shots taken with the K-5 are cropped? If yes, can you post the full images? If they are cropped, then you could be dealing with a case of not enough "material" at the center for the focus system to lock onto. Instead, it could be finding a spot of bigger material and locking onto that. I assume you have the camera set for center point focus on these tests?

[benjikan]

My situation as well. I am exhausted with the paranoia of having over 50 percent of al my stuff out of focus especially when shot under f5.6. I certainly hope that Pentax take back my almost new K5's and give me a couple of K5ll's...

Thank goodness I purchased a split screen focussing screen. I am now focusing most of my stuff manually.

[Tanzer]

Thanks for the info, Tanzer. I don't have the K-5 with me now. I have a K-r with me. Does it have similar large focus points? I can give it a try to better understand what you are saying. From my previous tests, it appears the mis-focusing only happens to K-5, but not K-r.

Also FYI, these test shots I uploaded were taken from approximately half meter away. I did try to focus at minimum distance of the lenses, but only horizontally. I also saw severe front-focusing at these close distances. K-r had not problem at any distance with the same lenses.

I am attaching a test shot made at minimum focus distance of the FA 31 lens. This one has no cropping.

Unfortunately I don't have any experience with the K-r, so I can't speak for relative size of its focus points.

This last shot does seem to be close enough that the camera should have been focusing on the center line, in which case the front focus seems pretty severe. But when you get the K-5 back in your possession, give the "blank area" a test just to be sure where the focus point really is, and repeat the test. Good luck!

[123jippo]

turning off auto-rotate and things that have to do with camera orientation

These can also effect to AF?

[Alex Zhang]

To get a better understanding of what you're dealing with, can you tell us if the problem shots taken with the K-5 are cropped? If yes, can you post the full images? If they are cropped, then you could be dealing with a case of not enough "material" at the center for the focus system to lock onto. Instead, it could be finding a spot of bigger material and locking onto that. I assume you have the camera set for center point focus on these tests?

Hi IchabodCrane, the images in the original post are cropped. Yes, all images are taken with single point center focus mode. I am attaching the uncropped images for reference:

Horizontal image without cropping


Vertical image without cropping

[Alex Zhang]

Hi Tanzer, I guess if the center focus point has shifted I would indeed observe this problem. Will give it try once I get the camera back. Thanks for the advice.

[Alex Zhang]

My situation as well. I am exhausted with the paranoia of having over 50 percent of al my stuff out of focus especially when shot under f5.6. I certainly hope that Pentax take back my almost new K5's and give me a couple of K5ll's...

Thank goodness I purchased a split screen focussing screen. I am now focusing most of my stuff manually.

Sorry to hear that. I can understand how frustrate focusing problems can be and hope you find a cure for your camera.

[Alex Zhang]

These can also effect to AF?

I believe the focusing problem is not really related with the orientation of the camera relative to the ground. I remember did some testing on targets on a wall while hold the camera horizontally. I observed back-focusing instead of front-focusing during the table chart tests with horizontal orientation. This led me to think that maybe Tanzer is right in guessing that the problem is caused by focusing point shifted off the center. If that is the case then I would observe back or front focusing depending on where the real focusing point lies inside the scence.

And even this is just a wild guess, because I don't seem to have problem locking focus with anything I point the center focus point to. If it is shifted, I would guess that I will have problem locking focus for certain scences.

[IchabodCrane]

I believe the focusing problem is not really related with the orientation of the camera relative to the ground. I remember did some testing on targets on a wall while hold the camera horizontally. I observed back-focusing instead of front-focusing during the table chart tests with horizontal orientation. This led me to think that maybe Tanzer is right in guessing that the problem is caused by focusing point shifted off the center. If that is the case then I would observe back or front focusing depending on where the real focusing point lies inside the scence.

And even this is just a wild guess, because I don't seem to have problem locking focus with anything I point the center focus point to. If it is shifted, I would guess that I will have problem locking focus for certain scences.

Alex, if there's any way you can repeat the horizontal shot with the camera upside down, that will probably tell you if the actual center focus spot is really below center.

Alternately (and easier!) -- maybe you can use a rigid item with fine detail that can be propped with top toward the back (as in the shots you've taken) and also with the top toward the camera (image taken looking from underneath). If you get front-focus in the first case but back-focus in the second case then it would confirm misplacement of the center AF spot.

[RyanW]

As an aside, does DCU show the focus point on photos?