Tobes,
Yes, we theoretically should adjust VTF for different record thicknesses. And, we have a more difficult issue with warped records.
Please forgive me if this is ground you have already covered, but there is a real reason for what you have observed. For the Graham 1.5 and for many unipivot arms, the center of gravity is designed to be below the pivot point. This results in a small but real force that acts to restore the arm to the correct positioning with the stylus at the record surface -- any deflection of the arm (such as the stylus lifting above the record surface) causes the VTF to vary. It is analogous to a pendulum wanting to be perfectly vertical, experiencing a force pushing it toward the perfectly vertical position when it is deflected. At one stage of the evolution of unipivot arms, this may have been considered a desirable characteristic because it contributed to the arm's dynamic stability.
Bob Graham has eliminated this effect in his new Phantom arm by designing the center of gravity to be at the pivot point. He instead achieves arm stability with the help of a magnet.
When I set up my 1.5, now upgraded to a 2.2, I needed to be careful how high my weigh scale was when setting my VTF. It needed to be as close as possible to the height of the actual record surface, or else I would have suffered a different VTF than what I wanted. You and Carl have both observed this effect. It sounds to me like your scales are close to accurate, but their heights were different.
