The lower arm seems to be ok more or less, giving almost a circle and no saw teeth with the known calibration method.
It seems the upper arm needs to be further than 9.5 which i cannot reach with the standard plate, though I'm not sure :). A object further away shifts away a bit, compared to an object nearby and shooting the object left and right in the photo. I have a T20 so I can use it to test if it's really further than 9.5. Will keep you posted.
NN3 D7000 Samyang 8mm LRS = 56 without Nadir Adapter, URS = 93
I do not understand why you can not reach an URS of 95. With CP-2, I reached up to 106.5. The base plate of the CP-2 can be mounted to the top plate in two ways: short side to the end of rail and long side to the end of rail. The white marks divides it by 1/3 to 2/3. So you have to mount the short side of CP-2 to the end of rail.
Last edited by hindenhaag; 09-06-2013 at 10:06 AM.
Try this test: take two nadir shots, rotating the head through 180 degrees for the second. Stitch them in PTGui with control points only on the floor. Generate a rectilinear layered PSD file and view in Photoshop, with the opacity of the top layer set to 50%. This will reveal any problems with the lateral position and also with the mount of the camera to the upper rail. The centre point of the knob should be aligned in the two images. The example below shows a small misalignment. The scale markings on the rail can be used to estimate any correction needed:
(The papers scattered on the floor were there for control point purposes).
Don't spend hours over the setup. Be realistic: a 1mm alignment error will not generally prevent your panoramas stitching perfectly well. Remember that parallax cannot be perfectly eliminated for a fisheye lens because the entrance pupil is not located at a single point. Any minor problems at the nadir itself will in any case be covered over by the nadir patch.