Why choose series or parallel model of impendance?

[WA6MJE]

I am trying to get more sophisticated in my stealth antenna experiments.  I have a RigExpert AA-54 antenna analyzer. Typically I have been tuning antenna elements or loading coil taps for lowest SWR.  But that is not necessarily resonance, and I would like to tune for resonance instead and then deal with matching whatever impedance that might present at the feedpoint to see how that works out instead of just lowest SWR as the goal.

The AA-54 will report capacitive or inductive reactance in addition to resistance.  I could use this number to seek out zero reactance which should be the antenna or element resonant point.

But the AA-54 manual says that "you may choose either series or parallel model of impedance of a load through the Settings menu..." but does not explain which of these I would choose or why.  I cannot find anything written on why I would want one or the other or which would apply to what type of antenna.  I am experimenting with HF verticals, or more specifically ground planes with elevated radials.  My best guess is that this would present a series model of impedance as resistance would be in series with reactance. If this is the correct guess, then what type of antenna would use the parallel model?  Maybe phased stacked arrays? 

I would appreciate some guidance on which setting choice (series or parallel model of impedance) to use with the AA-54 analyzer and under what circumstances I would want one over the other when measuring my antennas.

Rene - WA6MJE

[WB6BYU]

Any impedance can be represented either in a series or parallel form.
Then analyzing a matching circuit or using a Smith chart, we often
end up converting back and forth a couple of times.

Either will tell you when the load on the analyzer is resonant.

If you are adjusting a series element (a base loading coil, for example)
then the series reactance tells you how much reactance you need. 
If you are adjusting a shunt element (a beta match, for example) then
the parallel form is easier to use.

HOWEVER:  when the SWR is not 1 : 1 (as it probably isn't if you are
still trying to match the antenna) then you can't count on X=0 as an
indicator of antenna resonance unless the analyzer is connected
directly to the base of the antenna with zero coax length.

This is one of the biggest problems that has been introduced with
SWR analyzers that measure reactance due to a mis-understanding
of transmission lines.

Let's take an example:  assume you have a vertical antenna with
a base loading coil.  The feedpoint impedance is 25 ohms, with adjustable
reactance in the form of a coil.  We'll use 20m, and connect the
SWR analyzer through a 6' length of RG-213, then adjust the antenna
reactance until you get an X = 0 indication on the analyzer.

You can work out the results using VK1OD's transmission line calculator:
http://vk1od.net/calc/tl/tllc.php

In this case, when we adjust the antenna for X = 0, we end up with
about 13 ohms, so the SWR is around 4 : 1.  But if the antenna really
is 25 ohms, shouldn't it be closer to 2 : 1?  Yes, it should:  the actual
feedpoint impedance of the antenna in this case is 25-j45 ohms.  It
is NOT resonant, even though you are measuring X = 0 at the analyzer,
because of the impedance transformation caused by the connecting
coax.  If you tuned for minimum SWR instead (using the same coax length)
the impedance at the antenna feedpoint would be very close to 25+j0
ohms (basically resonant) even though the impedance measured by the
analyzer would be 42+j31 ohms (due to the impedance transformation
along the coax.)  Even one foot of coax in this case is enough to add
about 5 ohms of reactance to the measurement.

[KH2G]

It may help to also recall that in a series resonant circuit, Z is low and in a parallel resonant circuit the Z is high.