> I do not know for sure but I presume all of todays antenna analyzers
> measure SWR and absolute impedance value (mag Z). From these numbers
> they calculate the vector data (real and imaginary part of Z)
The MFJ analyzer measures real and imaginary parts "directly", and
also measures SWR directly. It does use SWR and Z to calculate
Impedance at very low SWR values. The two different values of R and X
are factored together with a certain "weight" or priority.
As SWR increases above 2:1 the MFJ relies more and more on the
"direct" R and X reading, and less on SWR because of the slope of
accuracy when using each one at various SWR readings.
As you know, SWR is less and less accurate as the SWR is higher. By
the same token, direct R and X is not easy to calculate when X is
nearly zero and R is centered on 50 ohms (half voltage in the
detector).
The sign of reactance is not given, because there is no phase angle
detection. Some other analyzers give the sign but only "guess". What
they do is perturb the frequency and watch the change in absolute
value of reactance. The problem with that is an antenna and
transmission line can decrease reactance as the frequency is raised
when the sign is inductive, or can increase reactance with increasing
frequency when the sign is capacitive. Just the opposite of a lumped
component! Of course they can also behave like a lumped component.
Since that happens randomly, the reactance sign (+ or ) is really
just a guess using that common method.
I'd always check the sign by using a lumped component in series with
the output port to confirm it is correct if I was doing anything that
was "+ or  sign critical", unless I had an analyzer that directly
measured phase.
The reason sign is not measured is the cost to do it correctly and
accurately.73, Tom W8JI
W8JI@contesting.com
