This is great I really appreciate that advice. I've been playing around with similar techniques using sun tracker apps and the like.
I live at roughly 40 degrees north latitude so my max winter sun angle is 26 degrees, 49 degrees on the equinox and 72 degrees at summer solstice.
What I understand you to mean by
Now you can just draw three lines on your graph that all pass through the center of your structure at GROUND level, one at 78degrees, one at 55degrees, and one at 32degrees. Where they intersect the curved wall of the building will be were the sun will shine at noon. You just substitute your latitude for the one I used in the example to get YOUR values.
is something like this first picture where every part of the sphere above grade is convex and that all seems to make sense where the winter sun would fall.
but if I were to raise the whole structure above grade and apply that same process it would look like the second picture, which looks funny. haha
my suspicion is that there would be an angle at which the sun would reach the equator of the structure and then start to track predictably. and that leaves a short portion of the day where the sun is tracking on the "underside" of the sphere and then it would begin tracking over the top and the underside would be shaded
any thoughts?