Is there any way to center the camera on an object and have the camera's looking direction be perpendicular to that object's orbital plane at the same time? In the Galaxy Constellations simulation, if I center the camera on the sun, and I want to see what is directly 'above' or 'below' the sun, the camera rotation stops and becomes unresponsive to dragging when I get somewhat close to 90 degrees from the object's orbital plane.
So I can see that the normal of the sun's orbit points to the constellation Coma Berenices above, and Sculptor below, but I can't see exactly where in those constellations, because I can't move the camera into the correct place.
There's something of a 'camera forbidden zone' in a cone of constant solid angle growing above/below the object's orbital plane. This seems intensely weird to me because most of the time the view can be dragged in any direction from any place, except when it approaches that cone.
I feel like I ought to be able to pick my own coordinate system in my mind, and have the viewing controls act the same everywhere. Instead, when I'm visualizing an arbitrary coordinate system, that forbidden zone occurs at a seemingly arbitrary places. Instead of doing what it normally does, rotating the view past that point, the program will unexpectedly spin the entire face I'm looking at clockwise or counter-clockwise, around the viewing direction, in an attempt to rotate around that point.
So is there a way to work around this? If there's some technical limitation or legacy code that makes it difficult to change, or if this is a 'feature and not a bug', I think a nice consolation would be reducing the solid angle of that cone to a less significant portion of the sphere, hopefully small enough that users will be statistically unlikely to ever find it.