Whenever I used a VNA to measure s-paramerers, I would do a calibration, and that would remove them from having any impact on the end measurement results. Why is it that even after a VNA calibration for 2 port shunt measurement, you still have to use an isolator? Why isnt the calibration good enough? Thanks!