Ok, I have new detail on the matter.
First, I got the chance to try out this Tamron 1.4x MC4 version with 7 pins (variant #2) I saw earlier today for a short trial.
So I walked to one of the three places around me who carry Pentax gear and have DA* lenses.
The helpful store people let me try my K20D with the Tamron TC and the DA* 50-135 f/2.8.
I used it at 135mm with the light available in the store (sample attached is a
100% crop of it -- so you can judge AF accuracy and EXIF for light conditions). Light (EXIF) was ISO800 1/125 f/2.8 (LV6 -- "light value", i just learn that EV "exposure value" doesn't include iso... -- 1 step already subtracted due to TC).
Now come my findings:
- AF converges to focus within about 1 sec.
- At focus, it confirms focus about 70% of cases.
- At 30%, it just won't confirm focus and won't fire.
In case #3, it doesn't hunt, just doesn't confirm and fire. If it happens, it helps to focus elsewhere and come back once or twice. Also, it usually happened at low contrast rather than high contrast. My feeling was it wouldn't happen outdoor at daylight. However, it also didn't happen w/o the TC. The sample attached is for case #2.
One more detail: when focussing, I got the impression that the AF over- or undershot at each step. It still converged but may have needed more AF operations than usual.
So maybe, the body told the lens "move focus by x/f" but of course used a wrong assumption what f is...
If this is the case I doubt that the number of pins or K10D vs. K20D (or even SDM vs. screw AF) could influence this at all...
So my question is this:
Do we call this a non-working or a working TC?
Does the 6 pin version sometimes fail to confirm focus even if it has reached it?