Actually you might be onto something. Mechanical assemblies have tolerance and tolerance stack-up. If your camera is on the small end of the tolerance range, or rather a smaller flange and larger gap, and if your lens is similarly within specification but at the edge of an acceptable dimensional tolerance range, then everytime you change the lens it might shift or settle into a different position. I don't know if a few thousand'ths or +/-.003mm would make a difference in the camera focusing function, but it does seem like a plausible reason why camera 1 does good and camera 2 does variably good/marginal/bad. I do some mechanical design and have had to chase down issues like tolerance stack-up issues before, trust me that it ain't no fun chasing random problems.
Probably not the issue, but the potential is there for it to be a contributing factor.
that's an absolutely valid point. However, that does not explain the inconsistency between shots taken 2 seconds apart of the same subject, same lens, same settings.
Again, I'm suspecting something mechanical too such as an improperly working mirror mechanism. If, for some reason, the mirror does not return to it's proper position after a shot, the next shot's AF could be compromised and may result in a blurry picture due to inaccurate AF.
Does the 7D have a unique mirror/shutter mechanism? Since it sounds different from the 40D/50D/60D and of course the Rebels, it may be a one of a kind (to help achieve the 8 fps) and perhaps it's not without flaws. So my theory is that the AF electronics may be functioning properly but perhaps there is something mechanically wrong with certain 7D due to subpar QC.
Bottom line, I still believe that the 7D is a great camera and I didn't have many problems with mine. It's just that when I did have OOF pictures, it was unexplainable. And the last couple of days, after a trip to Canon, it became worse.