RE: NEC-LIST: CFA modeling results - 2

Len,

This note is in reference to my TechCorrespondence (TC) sent last
Friday, re using voltage sources, which I copy below. I have
highlighted by star the text I want to comment on.

I have sorted out what is happening when one calculates impedances,
inserts a series load needed (so one thinks) to cancel the port
reactances, and repeats and repeats. And, I do say repeats and
repeats. With current sources one does this only once.

For a voltage ratio of one, phases (cylinder/disc) 45/-45 degrees, no
loads

        Zcylinder = -222 - j 554.7
        Zdisc = 98.9 - j 395

Final load values to resonate, by the calculate-and-change method, are
j 505.8 ohms, and j 306.9 ohms (for the loads associated with the
cyclinder port and the disc port respectively, impedances

        Zcylinder = 1.3 - j 1.5
        Zdisc = 1.37 + j 1.5
        E(200m) 0.188 V/m

With a whole range of values in between.

The total antenna system resistance is 2.67 ohms, but the total coil
loss resistance is 2.7 ohms (coil Qs 300), so the antenna's resistance
is very very small.

If I insert loads having the same reactance to resonate as found with
current sources, viz. j 715.2 ohms and j 466.6 ohms, the port
impedances are exactly resonat with voltage feed, and impedances are

        Zcylinder = 177.6 ohms
        Zdisc = - 189.5 ohms
        E(200m) 0.1598 V/m

Compare with current sources.

Why the great difference? I specified voltage sources 1/+45 degrees
and 1/-45 degrees, but the phases on the feed wires when resonated are
180 degrees (first case above); and 90 degrees (second case above).

With current sources the phase on the feed wires is the phase of the
source currents.

So the apparent confusion concerning what happens with voltage feed is
now clear.

Jack
20 July 1999

>Hi,
>
> I have a question about CFA modeling.
>
> The phase relation between the disk and the cylinder is clear, but
>how about the amplitude? I read the NAB99CFA paper (
>http://www.antennex.com/preview/cfa/nab99cfa.htm ), it seems that CFA
>not only require E and H in-phase but also |E|/|H|=377ohm. So it seems
>that the amplitude of the feeding to the cylinder and disk may not be
>arbitrary.
>
> Another page( http://www.antennex.com/preview/Folder02/cfa_5.htm )
>gives an instruction about the feeding network. How to model it with
>NEC?
>
> I do not know whether my understanding is correct.
>
>Yaxun

Yaxun,

For the CFA (according to NEC-4D) the near field ratio |E|/|H| cannot
be made equal to 377 ohm no matter what source currents or voltages
you apply.

I have used voltage sources and current sources, and keeping the phase
difference equal to 90-degrees, I have varied the ratio of
Icylinder/Idisc and Vcylinder/Vdisc over the ratio one to ten (or ten
to one).

This produced a set of data which I have yet to summarize in a compact
way.

All I can say at present is that the port impedances change very
significantly as one changes the current or voltage ratio. But the
CFA is a poor radiator however you try to feed it!

To realize performance claimed one has to have the magic hand of the
inventors.

Regards, Jack

Post Script

A follow up on my message this morning
_______________________________________

In the table below I show computed values(according to NEC-4D) of
impedance, and field strength at 200 m (for a transmitter power of 1000
watts, antenna resonated, coil Q = 300), for the MF CFA over a PEC ground
fed by current sources (phase difference 90 degrees), for various Current
Ratios Icylinder/Idisc.

___________________________________________________________________________
Current Zcylinder Zdisc FS(200m)*
Ratio V/m
_____________________________________________________________________________

0.1 1869 - j 715 -17.9 - j 466.6 0.1132

0.2 934.3 - j 715 -35.8 - j 466.7 0.1088

0.4 467.2 - j 715.2 -71.6 - j 466.7 0.1169

0.6 311.5 - j 715.2 -107.4 - j 466.7 0.1305

0.8 233.7 - j 715.2 -143.2 - j 466.7 0.144

1.0 187 - j 715.3 -179 - j 466.7 0.1561

1.25 149.6 - j 715.3 -223.8 - j 466.7 0.1693

1.667 112.3 - j 715.3 -298.3 - j 466.7 0.187

2.5 74.9 - j 715.3 -447.7 - j 466.7 0.213

5 37.5 - j 715.3 -895.4 - j 466.7 0.2523

10 18.8 - j 715.3 -1791 - j 467.1 0.2817
______________________________________________________________________________

* Recall, previously calculated, that the FS(200m) for a 75 m monopole
over a PEC ground is 1.3555 V/m.

Notice that the reactive component of the port impedances is independent of
the Current Ratio, and that the cylinder is the better radiator.
Resonating the port impedances should be straight forward, inductive
reactances equal to +j 715.2 and + j 466.7 ohms in series with the
current sources feeding the cylinder and the disc resonates the antenna.

***************************

If one employs voltage rather than current sources the impedance
characteristics of the antenna appear at first sight to be very very
different. The first point to note is that I have to reverse the phase
relationships, cylinder + 45 degrees/disc - 45 degrees (we now have the
phase relationship shown in the ICAP'91 paper). A second point is that the
apparent port impedances depends not only on the Voltage Ratio, but on the
value of the inductances inserted in series with the sources (a strong
dependance), the purpose of which is to resonate the port impedances.

As an example for a Voltage Ratio = 1, if one tries to resonate the port
impedances by "change-and-try", the apparent resistive component can have
any value from 200-100 ohms (one value negative), to a few ohms (both
values positive), but it is not possible to exactly resonate the antenna.
However, if one inserts reactances having the same values as computed above
(for current sources), the port impedances for voltage sources are
resonant, and the resistive component and field strengths are almost
identical with the values computed for current sources.

It appears therefore that one can fool oneself (using voltage sources) into
thinking that you can adjust the port impedances to a value close to
50-ohms (for example), but the port impedances are not resonant. Kabbary
et.al. (NAB'99 paper) state "a facinating feature of CFAs is that the input
impedance can always be adjusted to match any desired impedance at the
broadcast frequency." In principle my modelling tells me that by
luck-and-chance this might be possible, since for certain values of series
inductances it might be possible to find port impedances where the
resistive components are nearly equal and equal to 50-ohms (for example),
and the reactive components are also nearly equal but of opposite sign. So
when driven by a single source (transmitter) the antenna system is
resonant.

**********************

But all this is just a question of matching. It appears to me that it
would be better to employ CURRENT SOURCES (current baluns), since the
resistive components do not change as the reactances to resonant are
adjusted.

Recall (practically), a current balun forces equal currents into unequal
impedances (the two arms of a dipole or a loop, or with centre-tap to
ground, the two ports of the CFA). A voltage balun applies equal voltages
to unequal impedances.

John S. Belrose
16 July 1999

_____________________________________________
John S. (Jack) Belrose, PhD Cantab, VE2CV
Senior Radioscientist
Radio Sciences Branch
Communications Research Centre
PO Box 11490 Stn. H
OTTAWA ON K2H 8S2
CANADA
TEL 613-998-2779
FAX 613-998-4077
e-mail <john.belrose_at_crc.ca>
_____________________________________________
Received on Wed Jul 21 1999 - 21:03:54 EDT