Originally posted by blende8 Your findings sound very strange to me. As the AF is measured with lens wide opened TTL-ly, I cannot see why the Av setting in the camera can affect the final achieved focus unless this Av value is *also* used as one of the correction factors (as a parameter) in the calculation / adjustment of the final AF point. But this looks rather illogical, technically.
Even if this is true, the outcome proves that this algorithm is just a failure, as the lens focus wrongly at wider apertures..