[JohnBee]

For those of you who were following I finally concluded my investigation of the infamous MF, AF confirm issue with my K-5.

And so here's what I did.

Today I went out to my local camera pusher and got to try several MF lenses on two other K-5's to test. And low and behold... they all did the exact same thing!

This is/was great news in some way.
Because I knew my unit was unlikely defective.
or... at the very least... not alone

Either way, I was more than happy to conclude that the issue is not a defect per say and that I could rest easy knowing my camera wasn't defective.
Therefore, I will content myself with installing a Katzeye in a second K-5 which we will dedicated as a MF shooter.

Oh... and we picked-up our second K-5 today also to complete our working kit.
Needless to say, the wife was pretty happy about her new unit as she giggled while entering her name in the menu and took her first shots. All in all, I am very happy with the K-5's performance as I continue to work with the unit. And I am quite happy to retire our K20's with the K-5 this year.

And so now, all that's left to do is call the dealer in hopes of getting my 32GB SDHC back from one of the units I was testing

---

Last edited by JohnBee; 04-17-2011 at 08:28 AM.

[Catalana]

Thanks for your update. I want to make sure that my next course of action will solve my FF issue with manual lenses, so here are a couple of questions.
Do you think that installing a Katzeye focusing screen alone will fix the issue? I am under the impression that it is rather an issue of a correct focusing screen "shim" that will fix it. The Katzeye would make focusing easier, but if it itself is not set to spec in relation to the mirror and prism, you would still be focusing on the wrong spot. Does the Katzeye come with an assortment of shims?

Thanks...

[kheldour]

Thanks for the update. Will update for the K-5 later this year and hope Pentax will fix the issue soon.
And I hope you don't loose your 32GB card.

[Kenn100D]

I believe all K10, K200D, K20D, K7 have the AF confirmation even if you set the switch to MF.
I use this as a guide to my eye using Manual Focus Lens.

[JohnBee]

I believe all K10, K200D, K20D, K7 have the AF confirmation even if you set the switch to MF.
I use this as a guide to my eye using Manual Focus Lens.

Yes that's true,
But the problem is where the AF confirmation(on the K-5) looses its accuracy at wider apertures. By 1.4, it's useless.
For what its worth, the K20's do this too(to a degree) but they seem less prone than the K-5. But in either case, the results are unreliable. And so my resolve is to install a split screen or use LV to get the job done.

[jstevewhite]

Thanks for your update. I want to make sure that my next course of action will solve my FF issue with manual lenses, so here are a couple of questions.
Do you think that installing a Katzeye focusing screen alone will fix the issue? I am under the impression that it is rather an issue of a correct focusing screen "shim" that will fix it. The Katzeye would make focusing easier, but if it itself is not set to spec in relation to the mirror and prism, you would still be focusing on the wrong spot. Does the Katzeye come with an assortment of shims?

Thanks...

A split rangefinder screen will not resolve the issue with the AF confirmation. It will allow you to manually focus accurately, though; even still, you may have to change the shim to calibrate your focusing screen properly.

[joe.penn]

Yes that's true,
But the problem is where the AF confirmation(on the K-5) looses its accuracy at wider apertures. By 1.4, it's useless.
For what its worth, the K20's do this too(to a degree) but they seem less prone than the K-5. But in either case, the results are unreliable. And so my resolve is to install a split screen or use LV to get the job done

This is interesting. I posted in the other thread as I was running into issues with the MF on the K5, MF seemed to be way off while AF was spot on, and this was using a few different MF lenses.

The strange thing is, it seems (for me at least) that when I finally had a chance to drop the super tak 1.4 on it, the MF was spot on with that lens, and that was through all the stops.

[JohnBee]

There's no doubt inconsistencies exist at this level, but apparently... the K-5 AF module does not work well with MF lenses. However... if you use a lens with contacts, the camera can compensate and remain accurate.

For what its worth, I tried this with 3 other K-5's(my own being a 4'th) and the results were identical on all units. So I was relieved in knowing it wasn't a malfunction per say.

[MightyMike]

I just posted about this on dpreview... I have noticed the issues dating back to the K10D, and it wasn't always consistent leading me to believe that slower lenses with their larger DOF weren't such an issue.

My fix was to set the K-5's global AF adjustment to -10 when using a manual lens, all of a sudden its spot on.

interestingly the K-x with a similar lens (both being 135mm F1.8s) has this issues to a lesser degree, sure its there most of the time but i tend to get closer to accurate with the K-x over the K-5 (keep in mind a different lens with the same F.L. and aperture)

Mike

[padmk3]

Just to answer an earlier question, the Katzeye focus screen doesn't come with any shims. Just fitted one to my K5 and luckily I don't need any, it's spot on. I believe Katzeye have very good support if you do have any problems though. As for the Katzeye, fantastic, totally transforms the image quality in the VF and makes MF instantaneous.

[bjdavey]

I believe Katzeye have very good support

I can testify to that. I've just received my K-5 back from Katzeye, where they installed and calibrated my new focus screen. Their service is of the highest order, and I am delighted with the result.

[hjb981]

This was an interesting thread, and I have a couple of questions:

1: Why would the autofocus work less well with large aperture lenses? Shouldn't that make it more accurate, if anything? I might add that I am using a Pentax M 50/1.7 (standard lens from the late 70s) with my K-5, and the autofocus confirmation during manual focusing does not work well with it. Autofocus does work normally on my DA 18 - 135, so I agree that it works less well with this manual focus, large aperture lenses, I just don't understand why.

2: With one of those Katzeye focusing screens, will the camera work exactly like normal when I use it with a modern lens in automatic mode, but with the addition of a split screen to confirm focus with autofocus lenses? That would be great, because then I could use my old 50 mm manual lens much easier than today. Today I use it for extreme low-light photography at parties etc, and focus by using the scale on the lens, take a picture, check it, adjust, take another picture, check, adjust and so on. I am getting pretty good at judging distances from this . Using live view is another option, but often to slow to capture the pictures I want.

[bjdavey]

With one of those Katzeye focusing screens, will the camera work exactly like normal when I use it with a modern lens in automatic mode, but with the addition of a split screen to confirm focus with autofocus lenses?

I've noticed no difference at all with my AF lenses, other than it is much easier to see that the AF has nailed it (or not). It shows me exactly when I need to fine tune or readjust the focus. It's not just the split screen - I find that the surrounding ring of microprisms is often even more useful for seeing the focus pop point. With my aging eyes, and often difficulty matching my specs to the viewfinder, the Katzeye screen has made the whole focusing process remarkably easier.

[Skymist]

I spent some time this evening trying to take indoor light MF shots with my K-5 and the Zeiss 50mm F/1.4 at about F/2 or less. I have a focusing screen from focusingscreen.com, a model with a split center and a ring. (FSB)

It can be quite frustrating. Using the focus confirm feature, it is rarely right on best focus. Under the best conditions, I can adjust by turning the lens a bit past the confirm point, and get a good focus. But then when I pick a different subject, the adjustment amount is different. The screen is only a help when used with a high contrast subject in enough light to matter. Used on a face in dim light, it is hard for my old eyes to accurately spot the convergence.

After quite a bit of experimentation, I began trying a new tack. I centered the subject, and swept the focus through the confirm point while taking 4 or 5 exposures in slow continuous mode. Then I would quickly review the images, choose the one which had the best focus, and mark it with the "protect" flag. After collecting a lot of images this way, I did "delete all" and the protected images, comprising only 10% or so of the total, were the only ones left. I know it sounds complicated, and it took a bit of practice, but I could be sure of getting results.

There's no question that manual focus in dim light is difficult, especially when using large aperture lenses. It's no sure thing even in bright light, but at least you can stop down and get a more forgiving situation. What I don't know, is how different the K-5 is from similar cameras, such as the D7000, in this respect. The nice thing is that the flexibility of the K-5 controls and features can help make up for the problem.

In firmware 1.03, which I am running, the focus light does not come on as often. When it does come on, it is quite effective. I don't see yet what determines whether it is on or off.

[jstevewhite]

I spent some time this evening trying to take indoor light MF shots with my K-5 and the Zeiss 50mm F/1.4 at about F/2 or less. I have a focusing screen from focusingscreen.com, a model with a split center and a ring. (FSB)

It can be quite frustrating. Using the focus confirm feature, it is rarely right on best focus. Under the best conditions, I can adjust by turning the lens a bit past the confirm point, and get a good focus. But then when I pick a different subject, the adjustment amount is different. The screen is only a help when used with a high contrast subject in enough light to matter. Used on a face in dim light, it is hard for my old eyes to accurately spot the convergence.

After quite a bit of experimentation, I began trying a new tack. I centered the subject, and swept the focus through the confirm point while taking 4 or 5 exposures in slow continuous mode. Then I would quickly review the images, choose the one which had the best focus, and mark it with the "protect" flag. After collecting a lot of images this way, I did "delete all" and the protected images, comprising only 10% or so of the total, were the only ones left. I know it sounds complicated, and it took a bit of practice, but I could be sure of getting results.

There's no question that manual focus in dim light is difficult, especially when using large aperture lenses. It's no sure thing even in bright light, but at least you can stop down and get a more forgiving situation. What I don't know, is how different the K-5 is from similar cameras, such as the D7000, in this respect. The nice thing is that the flexibility of the K-5 controls and features can help make up for the problem.

In firmware 1.03, which I am running, the focus light does not come on as often. When it does come on, it is quite effective. I don't see yet what determines whether it is on or off.

I've had really good luck using the microprism ring around the center split rangefinder. The rangefinder is great for vertical things, but faces, as you say, can be tricky. So I put something small and contrasty (like the reflection on my daughter's glasses frame, or my wife's black earring) in the microprism circle and turn the focusing ring until it's smallest and clearest. This gives me good consistency even with my 50mm 1.4 and my 100mm f2.8. It's a testament to the sensor that I'm even trying to manual focus (because the autofocus has given up ) at the light levels I am