I would check the ohms with it still installed in the tank as is, at the 2 dead spots which would confirm an open in the sending unit windings. I'm wondering if when you mock it up & manually move it thru it's travel that you are holding a bit of side pressure onto the arm as you move it across the windings whereas in the tank it is more free floating and hitting 2 places in the windings that are not flat & hence no contact. but for 2 senders to be acting the same (if I read that right), I agree that does not have a logical explanation but it's gotta be something simple (after we find it) and I wanted to throw something out there to (hopefully)contribute.