focus adjustment k10d

How far out was it originally, Peter?

A point just above the lower 10mm marker was sharply in focus. -40Um adjustment corrected it in landscape. In portrait, after the adjustment, the upper 2mm is more sharply in focus than the centre line is.

If I can correct that, I'll be a happy bunny.

Also, rather than focus on the thick black centre line of the focus chart, I lined the camera up on the left text column and focussed on the middle text string.

I've used my FA 100mm Macro lens at f2.8 for this so far. I'll try again later with a FA 50mm at f1.4, but the images from the DA* 16-50mm are very much sharper.

Damn, I should have done this ages ago.

I have to go out for the morning, back later with more info, hopefully.

Hi all, can I ask the really thick question here? (at least I admit to it)

With the K10D I didn't realise about all this BF type of issues, do they apply to just the autofocus mechanism and focus confirmation or what appears to be focused in the viewfinder when using manual focusing?

For the most part I use older Pentax (A series) fast glass i.e. manual focus together with a soon to be delivered Katz Eye screen. I found it quite hard to get sharp focus every time with standard screen (cos I’m an old bugger, with now new glasses so I can at least find the camera).

Adjusting the autofocus/focus confirmation green hexagon does not change what you will see on the focus screen through the viewfinder. The autofocus mechanism is behind and underneath (i think) the mirror. The light from the image you see in the viewfinder takes a different path. It is entirely possible that an individual camera may have a defect so that the distance from the rear lens element to the focus screen and to the sensor are not identical, but this would be a seperate issue to any auto-focus problems.

For what it is worth my experience with autofocus on close objects with the K10D was that I occasionally got missed focus shots. Unless you have a flat plane directly in the centre of the autofocus sensor this is surely bound to happen. I have since put a split prism screen in the camera, and get great results focussing manually.

Just adding another dimension to this...

I was looking to see which other cameras have user-changeable focus that the K20D has. Looks like it's probably only the Canon 1Ds Mk III
As a result, I dug this method out:
AF Microadjustment for 1D/1Ds III.

There's a ZIP file for the pattern - which is absolutely incredible on the eyes Real migraine stuff! LOL!

Essentially uses an interference pattern to adjust the focus
Think I'm going to have another go at calibrating the K10D later
Matt

What I don't realy understand is why there is a need to test, and potentially adjust, each and every lens. Surely the camera body needs adjustment, not the lenses.

Matt,

Try it in both landscape and portrait and let me know your thoughts.

Peter - when I get a chance, I will try it. The interference pattern test should work the same in both directions. However, I'll try both methods in landscape and portrait with the 50mm f/1.4 lens and see what I can accomplish!
Matt

I've had a few hours playing with this focus adjustment now and have concluded that each lens needs a different focus 'offset'. This is not a problem if you have a K20D, but K10D's seem to have only one register for storing this data, so the one adjustment must be applied to all lenses.

I've used FA 100mm Macro @ 2.8, DA* 16-50 @ 2.8, DA* 50-135 @ 2.8 and Sigma EX DG 180mm Macro @ 3.5 for my tests.

I have been able to make focus improvements as all of my lenses seem to need around -40um of adjustment. The Sigma needs -70um.

Almost a case for leaving the body in debug mode, and labelling all of my lenses with the required offset and make the adjustment as lenses are changed. What a pain in the b**t.

BEWARE: In debug mode, the camera is not turned off when you open the card reader door. This could potentially lead to damaged or corrupted cards. Turn the camera off before removing the card.

I'm beginning to see why Pentax have not issued a firmware upgrade for the K10D. It might not have enough onboard memory to store all of the required values.

Please, please could someone explain why this is so????

Just a guess... could be partly to do with colouration of the lens - the AF mech is known to focus differently in low level tungsten light (ie mainly red wavelength light). I also reckon that the ratio of body motor-to-lens-adjustment may have something to do with it as well. I presume the lens ID enables the body to work out the focus travel for a particular lens (which may explain why some third party lenses mess up - I've seen them "pretend" to be a Pentax lens which isn't even close).

Just some thoughts... will try my focus some time...
Matt

Just a wild guess but could it be how well collimated the light is as it leaves the rear elements of the lens?

When the camera is setting the focus the light is reflecting off the mirror and up to the focusing screen where the focus sensors are, but when the image is actually captured the light is travelling straight from the rear lens element to the image sensor – a different light path. There will always be some error in these two paths which I expect is in part a function of the lens itself. Change lenses and the error (or minor differences) between these paths will change so to get perfect focus the ‘offset’ would need to change also.

I think that the fact that Pentax have allowed this adjustment in the K20D, and that they've allowed various offsets to be stored for different lenses, that this is now a recognised problem.

Why then can we not have a firmware update to the other bodies to allow the same?

I can't even estimate how many DSLR's Pentax have sold since the *istD in 2003, but they have potentially the same number (or similar) upset customers who just might buy Canon or Nikon next time, like Petekd did.

To use this as a tool to coerce customers to move to the K20D just is not on.

Well I guess that if we are talking about adjustments of microns, that not all lenses are made to exact micron tolerances? They would probably have to be within an acceptable range, so perhaps this is why all lenses are different?

I've not done much 'proper' testing, but with just shooting a 'live' subject, the 77mm Limited looks best around -7 on the K20D. Just wish they could come up with a way to automatically adjust the lenses rather than letting us try and get it right ourselves!