Forgot Password
Pentax Camera Forums Home
 

Reply
Show Printable Version Search this Thread
09-20-2011, 02:52 PM   #1
Junior Member




Join Date: Nov 2008
Location: Taunton, UK
Posts: 48
K10D GX10 CCD Position Adjust (AF adjust)

CCD Position Adjust (AF adjust)

I'm sure you're all fed up with the run of the mill back and front focus issues that people have with their K10Ds and GX10s.
I had such a problem on my GX10 and now that is mostly corrected through an offset entered in debug mode.

But on looking closely at the distance beyond or in front of the focus confirmation point, I have suddenly realized why some parts of my photos are not as sharp as they might be.
While there was an overall back focus that would ruin many a shot, the problem was worse on one side of the pictures than the other.
Having corrected the main back focus I now see that the focus plane is not square to the camera.
The focus point on the right of my pictures is closer than on the left.
Now that I look for this I see it in many photos. Often it is very subtle and not really an issue, but sometimes it is a problem.
Presumably this is a mechanical issue with either lens, camera or both.
However I have now noticed this same issue on 2 different lenses (but both Sigma telephotos).

My thinking is that maybe the CCD is not 100% square to the light path from the lenses.
The service manual (page 70) mentions an adjustment that can be made to the CCD called 'CCD Position Adjust' that is described in the manual like this:

1) Remove the Bottom Cover (Don't remove the Power and USB port cover)
2) Remove 3 screws
3) Lift the Front cover to the side and rotate the adjustment screw.

My question is whether anyone knows what this adjustment actually does.
I guess it physically moves the CCD, but in what direction? Can this straighten the CCD if not quite square?

09-20-2011, 03:03 PM   #2
Junior Member




Join Date: Nov 2008
Location: Taunton, UK
Posts: 48
Original Poster
Sample picture taken at 45% to pavement using Sigma 70-300mm lens set to 70mm at f5.6.

Take a look at left and right sides of the picture just above center.
I used auto focus at the center.

Similar effect with 400mm Sigma lens where I first noticed this.
Attached Images
View Picture EXIF
SAMSUNG GX10  Photo 
09-20-2011, 04:02 PM   #3
Junior Member




Join Date: Nov 2008
Location: Taunton, UK
Posts: 48
Original Poster
The approximate line of sharpest focus I have indicated here.
The fact that it is curved is probably down to the lens, but I'm thinking that the difference between one side and the other may be down to the CCD.
Attached Images
View Picture EXIF
SAMSUNG GX10  Photo 
09-21-2011, 07:58 PM   #4
Veteran Member




Join Date: Jul 2008
Location: Detroit MI, USA
Posts: 508
Because the sensors plate is part of the non-serviceable shake reduction assembly the sensor can't really be moved. Think about it, its free floating between two metal plates; and you don't want to tilt the whole SR platter in any direction during assembly. The SR (as a whole system) platters can be moved up/down however and there is an alignment the tech can do to the whole plate but its only a vertical alignment and not anything that would help your example above.

The three screws you refer to really adjust the position of the AF module. Take a look at any of the K10D or K20D service manuals and you will see the AF module being held by three springs that are adjustable from the bottom (then glued). This enables corrections for the AF module.

Now other than the obvious (to make small corrections) I don't know how that could correct a sensor being tilted to the side a bit. But lets assume the sensor is installed during manufacture to withen very tight specs and that its not tilted but perpendicular to any light coming in. I suppose putting in the AF module with the same exact alignment as the sensor in every camera is pretty hard. So the camera is bolted (front lens mount) to a special machine and the tech turns those screws until the software says its good. Mine was adjusted and it does have a major impact on AF performance. But you can't do it yourself and don't try please... unless you have the gear shown in the link to the manuals below

Downloads

Its a question I asked myself. I spent a bit of time researching those screws and can write a little about what they do. They do not move the sensor, they need a machine to be adjusted correct, they move the AF module. Those are the only facts I have.

09-22-2011, 01:52 AM   #5
Senior Member
karma mechanic's Avatar

Join Date: Sep 2007
Location: Hampshire, England
Posts: 192
James is correct, that adjustment although called CCD_POSITION_ADJUST does in fact change the relationship between the AF module and the sensor by moving the AF module. It has limited effect, and for larger adjustments it is possible to use the AF adjustment in debug mode. That's safer in any case since it can be reset easily. You've already done that.

That's not what is causing this problem. It looks to me as if your sensor is not exactly parallel to the lens. Or possibly the lens is not exactly centred.

The sensor block is mounted on three bolts into spring-loaded pillars, and those govern the distance from the lens mount to the sensor block and whether the sensor block is parallel to the lens mount. The specification of that adjustment, assuming you have the correct measuring equipment, is +/- 0.1 mm. So even if you had the camera opened up you'd have no clear idea of how to adjust it and how to verify the result.

On the K20D those sensor mount bolts are accessible without taking the main board out, and I assume the same is true for the K10D. It would be possible for a suitably skilled person to make very minor adjustments to those bolts using a flat test subject. If you are not suitably skilled then you risk breaking the camera.

I've just converted a K20D to IR, and that involves removing the sensor entirely. When I opened the camera the three bolts were all 1.75 turns from tight. For correct IR focus at infinity with the Sigma 10-20mm I ended up with them at 3 turns from tight, and the AF offset I ended up with using the kit lens was then +400 um (0.4mm). I would say that a small amount of turn on the horizontal adjuster would get your lens and sensor exactly parallel. I'm not offering to break your camera for you though!
09-22-2011, 07:31 AM   #6
Junior Member




Join Date: Nov 2008
Location: Taunton, UK
Posts: 48
Original Poster
Thanks Karma,

Your description sounds spot on. I'm not too sure about doing this, but if I did, is it obvious which is the horizontal adjuster?

Thanks again for your expert input.
09-22-2011, 08:49 AM   #7
Junior Member




Join Date: Nov 2008
Location: Taunton, UK
Posts: 48
Original Poster
I see that there are 3 adjuster screws and not one of them is really the horizontal adjuster. They just all have to be in exactly the right place. Easier to introduce problems than fix them I think without specialized equipment.

09-22-2011, 08:59 AM   #8
Senior Member
karma mechanic's Avatar

Join Date: Sep 2007
Location: Hampshire, England
Posts: 192
Here's a shot I took before the original conversion (note this is the K20D), you can see the three Allan bolts through the cutouts in the main board. The one just to the right of the viewfinder is partly obscured by a flip-type ribbon cable. You can get to this point at very low risk if you follow the service manual already posted. I use a sheet of paper for making notes as to which screws are which, and place the screws and notations on a wide strip of masking tape so they don't roll away. To get the back off you'll need to remove quite a few screws and note their positions since there are several different sizes. The top cover can be lifted slightly as in the picture, and the bottom cover can be removed with 5 small screws and 8 slightly larger ones.

To remove the back there is a ribbon cable connector underneath, then the back can be gently pulled away from the body - you'll find that you may need to ease off the front panel as well to free the back.



I'd say that the bolts towards the bottom of the body will jointly take care of horizontal adjustment, but exactly how much adjustment is needed and in which direction is unknown. You can leave the top one, since the cable that's in the way is quite a fiddly one. Bear in mind that the image on the sensor is actually upside down, so it may be possible to infer from your sample shots whether the sensor needs to be moved towards or away from the lens and which bolt is resonsible. Or you may ruin it. Note also that the bolts are surrounded by some glue so they don't move. A 1.5mm Allan key turns them fine.

I accept no responsibility for your camera, etc etc. I am simply providing some information to allow you to assess the problem. Please don't embark on this unless you really understand the risks of messing up the camera. Ask me more questions if you aren't sure.
09-22-2011, 09:47 AM   #9
Junior Member




Join Date: Nov 2008
Location: Taunton, UK
Posts: 48
Original Poster
Thanks Karma, that's so helpful.

I can see from photos of the K10D that the bottom 2 screws are obscured by the PCB (no special holes), but it looks like it comes off easily enough. Thanks for the tips about screws etc.
09-22-2011, 10:01 AM   #10
Senior Member
karma mechanic's Avatar

Join Date: Sep 2007
Location: Hampshire, England
Posts: 192
QuoteOriginally posted by kevbirder Quote
Thanks Karma, that's so helpful.

I can see from photos of the K10D that the bottom 2 screws are obscured by the PCB (no special holes), but it looks like it comes off easily enough. Thanks for the tips about screws etc.
It looks like the K20D could be used to take test shots while the innards are exposed, thus allowing trial-and-error adjustment of those screws.

Unfortunately removing the main board is quite fiddly. The service manual does explain how, but there are quite a few wires to detach and reattach, plus some of the ribbon cables are quite short and have to be sworn at. The problem then is that your adjustment can't be tested until you put it all together again, so the trial-and-error approach is going to be impossibly long-winded and high risk. I've taken several to bits, but I wouldn't go down that path myself.
09-22-2011, 02:35 PM   #11
Junior Member




Join Date: Nov 2008
Location: Taunton, UK
Posts: 48
Original Poster
Karma,

Thanks for your wise words. I've just done a test with my 'standard' 50mm lens that I rarely use and I can't honestly say that I'm seeing the same issue. So although I have 2 Sigma lenses showing the same effect, I think it would be unwise to deliberately misalign the sensor to suit these lenses.

Thanks yet again for your great help. I do think now that the lenses are not sitting quite square to the camera rather than the CCD being off. Probably that is a problem with the lenses that I may just have to live with. When you look too hard you can always find something that isn't perfect.
Reply

Bookmarks
  • Submit Thread to Facebook Facebook
  • Submit Thread to Twitter Twitter
  • Submit Thread to Digg Digg
Tags - Make this thread easier to find by adding keywords to it!
adjustment, camera, ccd, dslr, focus, front, issue, lenses, photography, position, square

Similar Threads
Thread Thread Starter Forum Replies Last Post
K10D (and other bodies) AF adjust values dantekgeek Pentax DSLR Discussion 4 08-27-2010 04:52 AM
Broken K10D Date Adjust?!? DurDur Pentax DSLR Discussion 3 01-01-2010 04:45 PM
Need help with AF adjust on K10D mutedphotos Pentax DSLR Discussion 8 04-01-2009 08:31 PM
K10D - So debug gone at 1.3? Any way to adjust AF? ksignorini Pentax DSLR Discussion 4 04-13-2008 08:24 AM
Adjust the SR strenght in K10D Voe Pentax DSLR Discussion 33 03-25-2008 03:38 PM



All times are GMT -7. The time now is 03:34 PM. | See also: NikonForums.com, CanonForums.com part of our network of photo forums!
  • Red (Default)
  • Green
  • Gray
  • Dark
  • Dark Yellow
  • Dark Blue
  • Old Red
  • Old Green
  • Old Gray
  • Dial-Up Style
Hello! It's great to see you back on the forum! Have you considered joining the community?
register
Creating a FREE ACCOUNT takes under a minute, removes ads, and lets you post! [Dismiss]
Top