Chris - that's exactly what I had in mind

Chris,
If you have nimble fingers you can press the Green Button while still in the OP mode That's how I found the discrepancy between the GB and OP mode.

Both were wrong, but the GB was slightly worse when near to the "critical point" where the viewfinder starts flashing.

Now obviously you can meter wide open and work out the difference in shutter speed, but that's all very fiddly. Additionally, if I had some extension tubes without the electrical contacts I'd try some other tests - I can't yet prove whether the K10D is reading anything back from the lens that's adding to the confusion. Further, it'd be interesting to see how it compares when using, say an fixed f/8 telephoto. I can set up the experiment, but I just don't have the gear anymore...

I did wonder whether it's a case of hitting the lower limit of the metering. But then they specify EV0-21 at ISO100, 50mm f1.4. Since EV0 is 1second at f/1, it still doesn't make a great deal of sense... but I have to check it

Still, I've forwarded my concerns and further evidence - hopefully Pentax will at least come back with some info.

Matt

Yes, you are right. It's a shame the information about this wasn't collated in one place... never mind

So, I compared the Green Button (GB) with Optical Preview (OP) and noticed these variances:

• * When using OP along with small apertures (large f-numbers) and low level lighting the exposure display will flash. If it flashes the exposure will be wrong. The GB doesn't show this.
  
  * When you get near the "flash point" there is a variance between the OP and GB. GB will underexpose by up to a stop. Maybe this is aperture blade timing.
  
  * The "flash point" is directly related to the ambient light and stopped down aperture (i.e. not dependent on ISO rating).
  
  * The *ist-D and *ist-DS have similar behaviour but at different points to the K10D. In addition the *ist-D only flashes in the viewfinder, not on the top LCD during OP. Again, no flashing with GB. So, contrary to some suggestions this probably isn't related to the screen and changing it is unlikely to make a difference on the K10D (though I'm willing to try).
  

I have another theory, which needs some more tests... and it won't be such good news. Will keep you posted

Matt

I don't see that it could. It's only in the "M" mode stop-down with the Green button. Even if it took 1 second to perform a meter reading I shouldn't think that'll be a problem. Using stop down metering is ponderous anyway
(But I reckon they can still do it at that speed, it's just it would seem the sampling is in the wrong position time-wise).
Matt

Jan,
You are absolutely right - metering wide open is not a problem

I strongly suspect that the firmware code for stopdown/wait/meter is wrong. Perhaps it's not waiting for the blades to shut down before metering? (Which would explain why smaller apertures cause more problems as it takes longer to get to that position.)

Either way, it's fixable

Matt

First things first... just found out that my FA24-90 that has been repaired wasn't put back together properly

So, trying again with an F50/f1.7, here are the results:

Taken under the following conditions: tripod mounted, using Expodisc over the lens, centre weighted metering, ISO200, about 1m away from a blank office wall.

So, what comes to light is that the *ist-D had the same problem, but perhaps not as pronounced
Additionally, it may be the K10D is slightly more sensitive at ISO200 than the *ist-D... but that could be down to ambient light changes. Still, you get the general idea

I'm off to report this to through the appropriate channels....
Matt

Okay dokay... trying to get this thread back on track....

There's definitely a problem, it doesn't look like it's anything to do with the screen, and I very strongly suspect it's fixable in firmware

Did a very quick 5 min test with an Expodisc on the FA24-90 and it showed that the stop down metering in M mode with the green button wasn't working at all as expected. (In fact I think I know what they've done... but I'll save that for now.)

I need to double check the details this lunchtime and compare with the *ist-D and the K10D manual and then try and get this sorted once and for all

Matt

But then what about those who have replaced the screen???? Any firmware update would break their K10D metering...
Matt

Thanks for confirming that Chris. At least when I find the time it won't be wasted with my F and FA lenses

I'll try and set this up at work and have a go one lunchtime... will keep you posted...

Matt

George,
Maybe it's down to stop-down-metering (whether that's on FA, F, A, or M). I think I did a quick check when this first came up and found that even the FA 24-90 exposed differently when doing stop down as opposed to wide-open metering.

As I still have my *ist-D, and an *ist-DS I can also try the screens... but it all comes down to time
Matt