Keith Dysart wrote:
So sometimes a 600 to 100 ohm discontinuity
produces a 36.6 degree phase shift and sometimes
it produces a 22.7 degree phase shift (and probably
any value in between).

Yes, of course - nobody said the phase shift wasn't
a variable. Why would you expect it to be a constant?
It is a variable that depends upon the phase of the
component forward and reflected waves.

I suggest that "work[ing] up the phasor diagrams of
the component voltages (or currents) at the junction
where rho = (600-100)/(600+100) = 0.7143" will
not be useful for predicting the phase shift.

It will be useful for reporting that particular phase
shift. If other conditions change, that phase shift
will change. What is unexpected about that?
--
73, Cecil http://www.w5dxp.com