[Marc Sabatella]

After reading this thread I'm somewhat surprised. I've never had the slightest trouble with green-button metering on Pentax-M lenses with my K20D. I had (recently sold out all my old glass and bought new stuff) the M 50 1.4, the M 100 f4 Macro, the M 135 f2.5, and I have a Tamron 180 f2.5 with a P/K-M mount on it, and they all expose flawlessly - that is to say, very similar to my Pentax FA and DA lenses.

Did you run those lenses stopped way down? My guess is that you shot mostly wide open or between there and f/8 or so, in which case you wouldn't have noticed it so much. Also, you wouldn't notice it if you had a Katzeye or other focus screen installed. But if you still had any of those lenses, you might try actually running the test all the way through the aperture settings. I suspect you'd find more variation than you expect.

I'm surprised because this would (should!) be the absolute easiest way to meter accurately, as you're measuring actual incident light, not predicting the amount of light that WILL be available.

True, but that light is measured through a focus screen that was not really designed to be metered at anything but wide open. That's why replacing the focus screen solves the problem.

I would have been quite upset to discover that it metered inconsistently with those lenses. Consistent behavior I can compensate for - inconsistent behavior makes that impossible!

Oh, it's consistent in that a given lens at a given aperture always over- or under- exposes by the same amount. But the amount and direction to which a lens is off varies with the lens as well as with the aperture, so you can't just say "all manual lenses always overexposed by 1 stop" or anything like that.

[stevebrot]

(A couple of my lenses vary by 1/2 stop between infinity and close focus on the K10D...also scary!)

That makes sense and is nothing to be scared about.

The f-stop in the lens specs is at infinity focus. At close focus, the lens is a bit longer and thus a bit slower.

Take a macro lens for example, at 1:2 magnification ratio, the f-stop is about 1 stop slower the setting of the aperture ring. At 1:1, 2 stops slower.

Thanks for the good information. Sorry to be an alarmist. I was aware of the light fall-off at greater magnification, but did not realize that it would be measurable in regular shooting. I had only noticed it when doing some exposure tests related to the stop-down metering issue. I found that I had to test all lenses with focus at infinity in order to get meaningful results across lenses.

Steve

[stevebrot]

Did you run those lenses stopped way down? My guess is that you shot mostly wide open or between there and f/8 or so, in which case you wouldn't have noticed it so much. Also, you wouldn't notice it if you had a Katzeye or other focus screen installed. But if you still had any of those lenses, you might try actually running the test all the way through the aperture settings. I suspect you'd find more variation than you expect.

My experience on the K10D is that the Katz Eye screen (both Optibrite and non-Optibrite versions) changes the meter behavior, but does not "fix" the problem. At least it does not fix the problem on my camera. Based on comments from other users with the Katz Eye and other 3rd-party screens, it has not fixed their problems either. The only screens reported to help with the exposure issue have been those originally made to fit the *istD series of cameras (Pentax L_-60 series screens). Too bad none of the *istD screens have focus aids!

Oh, it's consistent in that a given lens at a given aperture always over- or under- exposes by the same amount. But the amount and direction to which a lens is off varies with the lens as well as with the aperture, so you can't just say "all manual lenses always overexposed by 1 stop" or anything like that.

Thank you Marc for repeating that last truth. The behavior for a particular lens is extremely predictable. While the general pattern is pretty consistent, you can't just compute a factor and apply it to all lenses.

Steve