[dofmaster]

I have a 31mm Ltd that seems to have it's focus out quite a bit. At +10 it's almost there but my reading of the chart is that it should be adjusted to about +12/13.

I assume that I would have to send this in under warranty?

Can anyone tell me what they do if the lens' focus can't be saved by the in-camera adjustment? Do they do a physical realignment of some of the elements?

I do not want to send in my camera for the 4 weeks that it will take, and I'm also thinking about upgrading the K1 later this year when it would lose it's AF adjustments anyhow.

[alamo5000]

What kind of light are you doing the test under? You should not do that test under artificial light.

Yes, it matters.

Once you are sure that you did the test right if it's still a problem then send it off.

[dofmaster]

What kind of light are you doing the test under? You should not do that test under artificial light.

Yes, it matters.

Once you are sure that you did the test right if it's still a problem then send it off.

Normally I do the test in my lounge with either (mornings) the target outside on the balcony in direct sun, or I reverse and have it inside in a relatively well-lit room if it's later in the day (outside it would be backlit due to the sun's position).

I have done this test two times so I'm reasonably confident, though the amount of fringing I get at f1.8 does make it a little hard. I plan to do the test a third time across the full AF adjustment range (i.e. 10 photos) for full confidence.

But still, do they rejig the lens, do I need to send it off with my camera?

[stevebrot]

It seems you have had your share of problems with focus and sharpness. (Did Pentax figure out what was wrong with your D FA 24-70?) A lens that falls out of range for AF fine adjust is pretty rare and when it happens, it is usually a good indication that there is a serious optical problem with the lens that makes it difficult for the PDAF system to attain accurate focus (decentered or defective element). A few questions:

• Are you using a slant-ruler device? If so, is the target portion parallel to the sensor or is it at an angle?
• What distance are you using for your tests?
• Are you using other than the center AF point for your focus adjustments?
• Are you able to attain acceptable center focus using full manual technique (manual focus using magnified live view)? You said there is fringing full open. Is the fringing bad enough to make manual focus difficult?
• Is the camera able to attain CDAF focus under live view without excessive hunting or inaccuracy?

I call particular attention to the fringing wide open. The FA 31/1.8 should be free of both LoCA and LaCA at center at all apertures. CA at its core is degraded focus and the PDAF sensor has a difficult time with it in much the same way as you or I do. Given your difficulties with the 24-70, I hate to suggest that your FA 31 has problems, but it very well might. If it is on warranty, now might be the time to have it checked out.


Steve

[leekil]

It seems you have had your share of problems with focus and sharpness. (Did Pentax figure out what was wrong with your D FA 24-70?) A lens that falls out of range for AF fine adjust is pretty rare and when it happens, it is usually a good indication that there is a serious optical problem with the lens that makes it difficult for the PDAF system to attain accurate focus (decentered or defective element).

I assume you have ruled out that the camera is not significantly out of AF calibration, if you have adjustment problems with several lenses?

[pasorro]

Dont do the test at 1.8. repeat the test at 2.8. Your camera AF central point is designed to work at 2.8 as Max Aperture.

Sorry bad english

[stevebrot]

Dont do the test at 1.8. repeat the test at 2.8. Your camera AF central point is designed to work at 2.8 as Max Aperture.

Sorry bad english

It still works at f/1.8 and with the same efficiency as at f/2.8. The results will be the same.


Steve

[stevebrot]

I assume you have ruled out that the camera is not significantly out of AF calibration, if you have adjustment problems with several lenses?

The reason why I am barking up the "bad lens" tree is the comment about significant fringing wide open making focus difficult. While a good copy of the FA 31/1.8 will have LaCA at the margins wide open, the center should be sharp and clear of aberrations. Significant LaCA has the potential to bias the PDAF sensor.

@dofmaster has other lenses, one of which had demonstrated centering issues. That problem was not fully figured out until a controlled test was done and adjustment was never an issue.* That aside, I guess it would be fair to ask how much fine adjustment has been needed on those to see if the there might be an issue with sensor positioning. (I.e. If all required significant "+" adjustment.)


Steve

[dofmaster]

It seems you have had your share of problems with focus and sharpness. (Did Pentax figure out what was wrong with your D FA 24-70?) A lens that falls out of range for AF fine adjust is pretty rare and when it happens, it is usually a good indication that there is a serious optical problem with the lens that makes it difficult for the PDAF system to attain accurate focus (decentered or defective element). A few questions:
[*]Are you using a slant-ruler device? If so, is the target portion parallel to the sensor or is it at an angle?

For the tests I'm using a Spyderlenscal tool. The target portion is parallel to the sensor

Datacolor SpyderLENSCAL
[*]What distance are you using for your tests?

About 3 metres
[*]Are you using other than the center AF point for your focus adjustments?

No
[*]Are you able to attain acceptable center focus using full manual technique (manual focus using magnified live view)? You said there is fringing full open. Is the fringing bad enough to make manual focus difficult?

Yes
[*]Is the camera able to attain CDAF focus under live view without excessive hunting or inaccuracy?

Yes

I call particular attention to the fringing wide open. The FA 31/1.8 should be free of both LoCA and LaCA at center at all apertures. CA at its core is degraded focus and the PDAF sensor has a difficult time with it in much the same way as you or I do. Given your difficulties with the 24-70, I hate to suggest that your FA 31 has problems, but it very well might. If it is on warranty, now might be the time to have it checked out.
Steve

(I answered some of your questions in the quote)

I have an email that my 24-70 is ready to be picked up from the dealer.... which I will do later this week (currently on holiday). It's been away for about 4.5 weeks, but I don't know if that means anything.

I will redo the tests on the 31mm when I have time - next 5 days or so, and share some of the photos here. Warranty is still good for another 2-3 months.

---------- Post added 07-10-18 at 12:22 PM ----------

Dont do the test at 1.8. repeat the test at 2.8. Your camera AF central point is designed to work at 2.8 as Max Aperture.

Sorry bad english

What does that mean in practice? E.g. for the new 1.4 lens...

---------- Post added 07-10-18 at 12:25 PM ----------

I assume you have ruled out that the camera is not significantly out of AF calibration, if you have adjustment problems with several lenses?

My DFA 70-200 and DFA 28-105 are sharp and don't need adjustment. (28-105 is less sharp).

How do I 'rule out' that the camera is not significantly out of AF calibration, just if all my lenses need strong AF adjustment?

[stevebrot]

My DFA 70-200 and DFA 28-105 are sharp and don't need adjustment. (28-105 is less sharp).

How do I 'rule out' that the camera is not significantly out of AF calibration, just if all my lenses need strong AF adjustment?

The fact that only one lens out of your current three needs adjustment establishes that it is probably a problem with physical positioning of the AF sensor.

Thanks for answering all those questions. About the only thing I would suggest is to decrease the distance from 3 meters to something closer. The LENSCAL instructions differ on distance in the several variants of that document I found (anywhere from 5x to 50x the focal length), but none suggest 100x. That is simply too far from the perspective of making it easy for the camera to attain focus with any degree of precision. OTOH, the distance should be far enough that the image of the slant scale will not be affected by field curvature. I suggest a distance where the slant scale is still in the center 1/3 of the frame when the center AF point is on the intended part of the target. This may be easier if using the small square at far right as the target. I found three versions of the user guide online. One said center of the main square, another said small square, and a third said main on one page and small on another. If using the small square, take care that AF point does run over the edge of the target to the slant scale. Oh, and make sure the target is parallel to the sensor. One of the users guides says to make sure both camera and target are level. That is not the same.


Steve

* I don't own the LENSCAL and could not find any version of the instructions that was helpful in terms of where to place the AF point for best results. All those intersecting lines in various parts of the target, but with no clue as the purpose of any.

[leekil]

Originally posted by leekil
I assume you have ruled out that the camera is not significantly out of AF calibration, if you have adjustment problems with several lenses?
The reason why I am barking up the "bad lens" tree is the comment about significant fringing wide open making focus difficult. While a good copy of the FA 31/1.8 will have LaCA at the margins wide open, the center should be sharp and clear of aberrations. Significant LaCA has the potential to bias the PDAF sensor.

@dofmaster has other lenses, one of which had demonstrated centering issues. That problem was not fully figured out until a controlled test was done and adjustment was never an issue.* That aside, I guess it would be fair to ask how much fine adjustment has been needed on those to see if the there might be an issue with sensor positioning. (I.e. If all required significant "+" adjustment.)

Just to clarify, I was responding to your comment, but asking the question of @dofmaster. I wasn't questioning your suggestions, but offering another one that hadn't been mentioned.

[stevebrot]

Just to clarify, I was responding to your comment, but asking the question of @dofmaster. I wasn't questioning your suggestions, but offering another one that hadn't been mentioned.

Your question was quite reasonable.


Steve

[dofmaster]

Here are the shots from my 31mm ltd. You can see zero adjust, and plus 9 and plus 10. To me the + 10 still looks out.

Good news is that it's not decentered.

Still 2 weeks left on this lens' warranty....

RAWs of the same:

https://drive.google.com/open?id=1UBSAhjT2XMq7AkER01-XG74Hs7vvbOHf

[rogerstg]

I think you have to be much closer. If I haven't messed up my calculations, at 3m your depth of field is about 1m. But at a half meter its about an inch.

Maybe try to use the lowest ISO instead of 200 so that you can squeeze the last bit of sharpness out of the lens.

[ryan63]

this adjustment can be done in software via the debug menu.
My DA*16 - 50 f2.8 needed more than +10 points of focus adjustment on my k-3.
Via the debug menus I added +60um of af adjust then re tested the lens.
It now focuses spot on with only +7 points.
All my other lenses then Needed re-calibration and had shifted by roughly the same amount, in the same direction