AF Adjustment

I have tried micro-focus adjustment several times on a number of lenses with my K30 - DA 35 f/2.4, F 35-70mm, DA 18-55mm ALII kit lens, DA 55-300mm - and cannot yet convince myself that I have discovered a method that works reliably and is repeatable.

I have tried taking photos of focus charts; of steel tapes; of book or newspaper text and even tried a method involving interference patterns on a computer monitor. I am not convinced that any of them work reliably and repeatably. Is it that the mechanical tolerances involved do not allow accurate adjustment - or is it simply that my methodology is garbage?

What do you folks do? Have you found the micro-focus holy grail or are you still struggling as I am?

Andrew

There is margin for error when adjusting and also different lighting conditions, different focal lengths, different focus distances all play a part. However, if a lens is consistently 'off' it makes sense that it needs calibrating. See above, what I've said about my Pentax DFA 100mm macro lens. It would not focus correctly in any conditions, at any focus distance, on any body. It needed + 9 adjustment on my K-5iis and +8 on the K-3. That is way different to any other lens I own and has made this lens usable when it was unusable. That may be an extreme example, but I would bet there are a lot of people out there with lenses like this who simply believe the lens is 'a poor copy'.

Regards
David

Can anyone tell if hearing the sound of the elements clink when the lens reaches infinity or zero when I manually focus my da"200 is normal?

Re my focus adjustment. I have been using this focal length regularly for a year now and rarely have I used or felt I needed to us a tripod and I doubt I will use one much with this lens in the future. I now know that if I do use my lens on a tripod it will, adjustment being correct, focus accurately. In my normal use handheld I doubt an adjustment of +3 will make a noticeable difference. My technique is not that good. Now if it had been a + or- 10 that's a different kettle of fish.

Please if anyone can comment re the clinking ie if it is normal or not it would be much appreciated.

Re the clinking noise. I have just tried a da*200 lens in a camera shop and found that it made a similar sound. Panic over

There are permissible tolerances that are very fine. These tolerances are designed to fall within the depth of focus. As long as they do, then all will be well. I'm talking here of Pentax cameras with Pentax lenses. The manufacturer knows how much leeway in manufacture is permissible.

Now if we are talking about Pentax cameras with other makes of lens, then it's quite possible that some will fall outside the tolerances and benefit from adjustment.

I agree, I've owned 5 Pentax DSLRs ( K10D, K20D and three K-5s ) and all 5 of them have required calibration for most of my lenses.


I concur.

You can't set it at the factory

They made it available because the facility needs to be there.

It's called Quality Control. All sorts of parameters on cameras and lenses are set at the factory. They have vastly expensive equipment to calibrate and adjust.

Both Pentax lenses and 3rd party lenses need calibrating with pentax bodies. I

I have several pentax lenses that needed AF calibrating on my two K-5s and the calibration settings are different for each body, and different from the third K-5 body I originally owned.

It can't be calibrated at the factory. that's why Pentax have included the facility for user calibration

I've found that the phase AF on my K-5s isn't consistent in nailing the focus. especially in lower light. Most of the time it is correct.

This is why I take several shot when calibrating to ensure that most of the time it's spot on rather than most of the time it's Front or back focussing.

The K-5II and K-3 has improved the AF system in consistency, speed and Low light so I gather

You can calibrate one body with one lens at the fractory. We can also do this at home.

What you cannot do at the factory is calibrate one particular lens with one particular body as there's no way of second guessing which lens and body will be bought and used by a certain customer. Trying to do so would be an absurd situation. Every K mount body would have to be calibrated with every inastance of every lens and store that calibration against the serial number of those lenses. There must be billions on combinations, given there's thousands of bodies and thousands of lenses.

Not if you stored the lens calibration parameters in the lens itself to be read out by the body - that's the way things are heading.

Hopefully they won't head that way as it will make it really difficult/impossible for third parties to make Pentax compatible lenses, especially for Sigma who traditionally have reverse engineered Pentax mount technology rather than license it. Also it isn't any sort of solution for all the lenses currently in existence.

I would much rather have the camera detect a new lens was being mounted for the first time and offer to calibrate it automatically on that specific body using LV comparison information. That way it could also be re-calibrated after a while if the lens has worn a bit, or if it is being used in extreme hot or cold conditions, all of which probably affect focusing as well.

No, because the calibration settings are different for each body. As I've already stated, I've had THREE K-5 bodies and the same lenses have required different settings on each body.

The settings you store in a lens, such as in the Sigma USB docking system will only work for the body you tested it on.

Unless you can store separate calibration settings in the lens for multiple bodies your suggestion is untenable.



The fact that the Sigma USB dock exists is proof that the manufacturer cannot calibrate their lenses to provide Accurate AF for all bodies. Body Manufacturers cannot calibrate camera AF to prove accurate AF for all lenses either.

This can only be a physical limitation of the Phase AF system, as Contrast AF ( through the sensor ) does not suffer from front or back AF.

The 'calibration' is just a fiddle factor numerical offset added to the phase detect module output to compensate for tolerances in the lens mount thickness - 'global' one for the body and a lens one for the lens flange. The two are added together. Nothing physically moves when you change it.

What do people do if they have two copies of the same lens ?