I got to run the tests today, and I found that:

1. Auto Exposure OK from f/8 to f/16. Overexposed at f/5.6 and f/22 to f/45.
2. Manual Exposure, similar to Auto, ok from f/8 to f/16, overexposed outside the range.
3. Sunny 16 worked at ALL apertures! Surprise here!
4. Sunny 16 was metered at f/8 to check the meter, and was close enough.

Summary: Manually meter at f/8, (one stop down), then just shift f/stop and shutter speeds to maintain the same exposure. I got five good, accurate, even exposures with the 400 on the k10d using these exposures at ASA 400:

1/3000 @ f/5.6
1/1500 @ f/8
1/750 @ f/11
1/350 @ f/16
1/180 @ f/22

I don't find f/32 and f/45 all that useful, so I did not bother. Chasing sparrows at 1/45 sec with a 400mm lens will not give me very many satisfactory images, I think.

Can you do a set of shots, not just reading shutter speed, of a block wall, at each F stop and let the camera pick shutter, then give us the grey scale average for the center of the frame?

That is the only way to understand what the camera is doing

Thanks

I found the auto mode completely unreliable outside the limited f/stop range. I don't know how the camera does this - but it is the metering that is inaccurate. If I use the metering from the "good" apertures, and then simply switch the aperture for shutter speed to keep the exposure constant, it holds on the 400. I have to retest the 100 now, and might get to it this weekend, and might not. I have a bridge tournament all day Saturday.

If you want to have a shot at this with an M or K lens, you can do it by just trying the green button. The exposures will not change after a certain f/stop, no matter how you do it. I was testing today to see if I could, in fact use the other f/stops if I just used sunny 16, and I can.

There was discussion on the forums that it might have been the stop down lever not arriving at the correct point due to a difference between linear and progressive stop down, but that did not happen with the 400.

The five accurate shots ( the apertures I will use ) are on my Flickr site as of a few minutes ago. I added the link to my signature. It's a really, really, really, really big wall to the North East of my house. but it is not brick.

Your really big wall is not exactly what I had intended, i use something a little more boring for tests, the exposed blocks on my house foundation.

What I do is put the camera in M and use green button to set exposure at each apature, then in PSP X2 measure the grey scale in the center 10% of the frame.

It would be very interesting to compare your results done that way, to my 400mm F5.6 Vivitar. I tested it, and found at F5.6 the grey scale value was between 140 and 160 (3 tests 3 different days) it climbed to between 170 and 180 at F22 and then dropped 10 points at F32

Note the metering should give a grey scale of 125 for correct exposure. as a result about 1-1.5 stops over exposed rising slightly from wide open

You should also set the exposure steps to .5 stops so that the camera logic matches the lens increments. I have found that if the camera is set to .3 the exposure can seem inconsistent stop to stop on the lens, especially if the lighting is just at the threshold, because the .3 stop steps of the camera don;'t match the .5 stops of the lens as you cclick through apatures.

For the purposes of my testing, I only used whole stops, so the 1/3 vs 1/2 was of no import. It will be extremely difficult for me to find an exposed brick wall here, particularly since all the foundations are still hidden by a meter of snow or more.

The other thing is that I really don't care about pictures of bricks. I want to know if I can obtain consistent exposures of the real thing. 1/10 stop is not going to bother me. 1/2 stop will bother me seriously and whole stops are intolerable. The mountain shot contained a wide range of subject exposure from dark (the trees) to light (the snow on top), and that, combined with the histograms, tells me that the exposure is one that is usable. I now know that all I need to do is to meter once at f/8 and use that exposure using the wet ware calculator in my head, or counting stops on shutter speed and aperture.

Another thing that came out of this testing is that I am going to reverse the edials. Holding the green button or dof preview over while trying to adjust a shutter speed with the front dial just does not work for me. Another project - change all the edial assignments. Maybe by August I will be ready for April's spring birds.

PS. I did a set of shots of a red flag with snow and trees in the background when I tested the auto exposures. At f/5.6, one stop over. f/8 - f/16, very nice. f/22 - f/45 one stop over. The camera is set to use 1/2 stop increments. This has to be something to do with the way the exposure meter sees the M lens.

I understand, in some respects, but it makes for an easier characterization. again I understand, I have 1 meter exposed in summer, right now in toronto it is also all covered
again I understand, and I made the mistake of asking someone the other day for a "calibrated' brick, but by using the same sopt or subject over and over again I can get comparative and repeatible measurements, i check the metering performance with all my lensesthis is why I test over the whole range I know this but I would prefer when I meter to know in advance the shot will be good without checking the histogram, I find I am spending a lot more time on the K10D checking the histogram than I think I should be. the *istD meters perfectly with manual apature lenses.I find the green button better for me, but those without K10D or *istD have only DOF, maybe by august your snow will be gone!

Not likely. That little "wall" is around 3,000 m. My house is somewhere over 1,300 m. (those are 6,500 ft and 4,000 ft in old style).