70-200G back from af adjustment with A700

[Chris Malcolm]

It's very difficult to know exactly what the AF seizes on in these kinds of shots. When doing my own focus tests with other lenses I got inconsistent results which varied with focal length until I set up a more exact test. I used a textured surface like your pebbles, but instead of the knife lying flat I stood a tall vertical item with a very high contrast edge on it a few cm up. Then focusing on that above the lower contrast support I could rely on where the AF would settle. A ruler lying on the surface gave me a scale by which to measure DoF and degree of focus error.

Note that a spherical wide aperture lens will have aperture related focus drift at its widest. Asphericals will have less, but may still have some. So test at different apertures as well.

[lock]

Chris, that is more less similar to a box test. I did that test with three boxes, the outside boxes behind and in front of the central one and placed on a board with parallel lines at 1 cm each.
The end distance forward and backward was set when both outsideboxes had sufficient blur.
Here's an example testing the A700 70-200G 70 mm wide open aperture, at 127 cm from the central box (min. distance 120cm) :


Effectively, total blurr was found at:
70 mm +2cm and -4 cm, DOF is about 3.5 cm equally spaced in front and behind.
100 mm +1 and -3 cm , DOF is only 1.7 cm equally spaced.
135 mm +1 and -2 cm , DOF is only .9 cm equally spaced.
200 mm+1 and -1, DOF only .4 cm equally spaced.

Thus at 70 and 100 mm there is 1 cm BF, at 100 mm there is .5 cm BF and at 200 mm it's more or less gone. However, at 3 meter distance and 200mm on the lens, a FF pops up. In fact all misfocus effects increase in absolute size with increasing distances. Ergo the lens is BFing from 70-135 mm, and tends to turn the dice with a slight FF at 200 mm.

I also found out the sensor and matte screen are misaligned. If I put the central box of the matte screen on a black spot on a white piece of paper, the spot is displaced from the cetner of the picture, upward (9%) and to the right (5%) .
Also, if I checked where in the circle the central AF sensor starts to lock. It seems like the sensor does not lock on objects at equal distances from the center.
When trying to get a focus in the top half of the circle, the sensor locks almost on the circle it self. In the bottom half I need to get is close as almost 1/3 distance between circle and central block before a lock appears:


So the fast AF F2.8 sensor shifted upwards. I did not find much sideways displacement.

lock