Primarily in reply to KJ4FDV's post...
The way to model ground loss invoked by elevated radials is fairly
straightforward if a couple specific gotchas are avoided, and this can also
be factually measured in place. MODELING of BURIED radials requires NEC4
based methods and the high priced programs. A method similar to measuring
elevated radials can be used for MEASURING BURIED radials.
The measurement is made using four elevated resonant radials. Use bare solid
#12 or #14 wire for the radials. This eliminates other confusing factors
including wild variation in dielectric loss among insulating materials, and
the variable dielectric loss added to the ground induction loss. It also
eliminates the issue of accurately measuring the resistance component of a
complex impedance of short radials in the presence of a large reactance. If
you are planning a short radial system you will need to use an antenna
analyst capable of ACCURATELY measuring resistance in the single digits AND
REACTANCE for measurements. AIM 4170, AEA VIA, and similar are
satisfactory. An SWR meter, or the 25x MFJ analyzers, etc will not measure
this accurately across the likely range of results and will most likely
deliver false results.
The 25x analyzers WILL work if one is measuring resonant radials and have
ground conditions that deliver an insystem ground loss resistance per
radial 50 ohms +/ 25 or so.
If you do have an advanced analyst measuring short radials, ignore the
velocity factor, and use ONLY the resistive component of the complex
impedance for calculation below.
To convert convert cheap wire (PVC insulated THHN, etc) to more expensive
and harder to find bare wire, simply strip the insulation. (THHN is a volume
item, bare is a specialty item.) Using a dull pocket knife it is possible to
strip 500 feet of #12 THHN in no more time than it takes to walk the
length. Sharp knives dig into the copper when the angle of the blade to the
wire is increased enough to keep the blade underneath the insulation while
stripping. You can straighten the wire sufficiently for stripping by
pulling the wire off the reel, through a pulley under tension for 180
degrees, and back the other way. This works, I have done it many times.
This is an ideal test to do at the beginning of constructing a raised radial
field to determine how many radials are needed above one's particular
dirt.The radials must be EQUAL LENGTH, EQUAL HEIGHT ALONG THEIR LENGTH, BE
STRAIGHT, be at measured 90 degree angles and use a common center post. The
radials must use the same wire size for their entire lengths. A measured
and accomplished 90 degrees between radials is essential to place
interaction between radials at a minimum to prevent obscuring loss
measurements with effects induced from less than right angle crosswise
wires.
First construct two radials directly opposite and inline with one another.
Connect their center ends. Solder a four inch wire to the center
connection.
Construct the other two radials direct opposite in line one another and at
90 degrees to the first two. Do not connect these two to the center
connection of the first two. Connect these two center ends and solder a
second short wire to the center connection, not touching the other pair's
connection.
We now have two short wires, e.g. one wire to the NS radial pair and the
other to the EW radial pair. Shortening the two connection wires as much as
possible, attach a connector appropriate to your antenna analyst. One to the
shield and the other to the center conductor if your device does not have a
balanced binding post. DO NOT USE A FEEDLINE to the center point. Measure
it with the analyst AT the feedpoint. Presence of a feedline will queer the
readings to a surprising degree. A useful tactic is to tape the analyst to a
nonmetallic support post and make your measurements by reading the display
with a pair of birding binoculars.
For resonant radials, use the graphic display (if available) or vary the
frequency of the analyst to determine the ZERO REACTANCE frequency of the
radials. If the ZERO REACTANCE point is not the target frequency, then use
a percentage add/remove method to make an EQUAL adjustment to all radials to
get it to the target frequency. Repeat if necessary until the radials ARE
zero reactance at the target frequency.
Compute the velocity factor by dividing the actual radial length by the
freespace quarter wavelength. Velocity factors for raised radials are
common in the 9095% range for bare wires at 6 to 8 feet.
At resonance we will be measuring a quarterwave ground induction intercept
plus the effect of the resistance of the wire. If the wire is in new
condition, wire resistance should be negligible. Since the radials are
opposite and equal, radiation loss to the far field should be zero. For new
wire over soil/dirt/sand/rocks, we could see a value between 10 and 100
ohms. If measuring over salt water, or over a salt water marsh, this value
could be quite low.
Supposing that we measure 60 ohms. We are seeing one radial pair in series
with the other radial pair. That means that a single radial pair is
measuring 30 ohms. Since there are two radials in parallel each radial must
have an INRADIALSYSTEM ground loss of 60 ohms.
So it turns out that the insystem loss resistance per radial is the value
measured by the analyzer using this radial setup and connection. For a
radial installation at that location and at that height, the effective
ground loss resistance of N equally distributed radials is the insystem
radial loss resistance per radial divided by N.
So if we measured insystem radial loss at 60 ohms when we put up the
radials, stayed with four radials, and after we put up our inverted L got a
feed Z of 30 ohms, the arithmetic tells you the following: 60 ohms loss per
radial divided by 4 radials in parallel equals 15 ohms loss in the radials.
30 ohms feed Z minus the 15 ohms radial loss means 15 ohms radiation
resistance for the radiator. Ground loss FROM RADIAL GROUND INDUCTION ALONE
is 3 dB.
That is BEFORE subtracting REFLECTION LOSS. Unless you are over salt water,
just about all of everything heading away from the radiator also headed
below the horizon will be absorbed. Throw in a little extra loss for
feedlines, and maybe 20 of every 100 watts put to this particular antenna
will actually convert to useful radiation. Sound like a familiar
performance paradigm?
If you have EZNEC (any level) you can monkey with real ground "high
accuracy" ground constants in the model to get the same R and resonance as
you actually measured. Then model the full number of radials with the
vertical radiator using EZNEC 3D plot. The loss number displayed at the
bottom of the main control window minus your computed radial ground
induction loss is the ESTIMATED additional reflection loss due to vertical
polarization.
Reflection loss and ground induction loss is to some degree offset by out of
phase induced current by the main radiator (a VERY complex voodoo subject),
NEVER close to completely offsetting though. Since some currents are
RESULTANT of two separate induction phenomena, it is important for purposes
of discussion to not characterize ground resistance as the inverse of a
resultant current. Ground resistance is a CONSTANT in a given setting, the
same as an ordinary resistor. IT DOES NOT VARY depending on induced current
short of loss heat converting ground moisture to steam or enhancing
evaporation, thus changing a physical characteristic. The induction of
current usually varies positionally as antenna elements are changed. The
degree of LOSS attributed to ground is a complex sum of the instant current
times resistance at points in the ground. This is a power value, which,
given the current at the radial feed, can be expressed as a SUMMARY
resistance. This is the useful term when combined with radiation resistance,
allows us to express the apportionment of our hardbought power to
wormwarming and chasing DX.
The positional movement of induced current, and changes in the resultant
current minima and maxima as antenna elements are varied, are NOT a movement
of ground resistance. If you do the above measurement at six feet, and then
do it at eight feet, the inradialsystem loss per radial will change.
GROUND RESISTANCE WILL NOT CHANGE. How much you induced current in that
ground resistance has changed.
Off reflector, I can supply at request a model that will work with any EZNEC
for above ground only, and below ground (buried radials) with EZNEC Pro and
NEC4 license. If anyone is REALLY serious about radial and counterpoise work
on 160, the Pro version of EZNEC with the NEC4 license from Lawrence
Livermore National Labs, and AIM 4170 or equivalent, really are the minimum
workable tools with enough accuracy to estimate and measure.
At some point, one becomes familiar with the ways in which real measurements
stray from what I term ACCURATELY executed estimates. Accurately executed
estimates mean that the calculation would have been the same as the
measurement, IF all the ground actually HAD the ground characteristics
specified in the model, and was of that characteristic everywhere near the
wire, and there were not nearby miscellaneous conductors, the ground was
flat and level in the measurement area, had uniform moisture content, and
measurements taken over time also conformed to these specs, etc, etc, blah,
blah, blah. Long list of variabilities.
But on the other hand, if we aren't out there doing pro level work, 160
meter ground is just too much a PITA to get real results and have a basis to
specify what to expect if you do X or Y.
I don't know about anyone else, but for me putting up 160 antennas is a real
effort and investment of time and money, with a lot of planning ahead of
time. Basing all that time and expense on urban legends is guaranteeing
frustration.
These days we (hammibus commonibus) have tools that the guys that did the
last real research on this (Bell Labs et al, in the 1920's and 1930's) could
only dream of. There are no paid researchers out there working on this NOW
because there is no MONEY riding on this pursuit. There WAS money riding on
it in the 1930's. We will have to be our own professional researchers with
our own strict academic level peer review standards. We will have to be our
own arbiters of precise language and standards of excellence.
Sweeping statements based on urban legends that SOUND good only serve to
confuse later readers using Google, that won't know that an unanswered false
statement from a reflector post is unanswered not because it is correct, but
because readers who DID know just ran out of time, or worse, got tired of
trying to convince someone who was not going to be convinced.
73 everybody, and may YOUR 160m antenna smoke 'em.
Guy.
On Mon, Sep 27, 2010 at 1:43 PM, K4SAV <RadioIR@charter.net> wrote:
> KJ4FDV: "Sounds great if the ground resistance stays the same for a
> given ground system with different antenna lengths. But that is an
> assumption." ....
>
> KJ4FDV: "I will lengthen my inverted L to 3/8 wl and measure the
> feedpoint resistance, subtracting the modeled 52 ohm radiation
> resistance to get ground resistance. " ....
>
> That may get you a ground resistance as seen from the feedpoint, but it
> won't get you the ground loss because, as you noted, the current maximum
> moves further out the radials. Approximating the ground loss as a
> resistor at the feedpoint seems to work fairly well for short antennas
> but not for antennas over 1/4 wavelength. It does work OK for
> calculating the feedpoint impedance but not for calculating power loss.
>
> Remember you can't get the radiation resistance (or the gain or the
> impedance) of an inverted L using a Mininec ground because there will be
> a low horizontal wire above this ground which will contribute error.
>
> To compound those problems, many people with a lot of modeling
> experience believe that NEC (both NEC2 and NEC4) underestimates near
> field ground loss for very low wires.
>
> Jerry, K4SAV
>
>
> _______________________________________________
> UR RST IS ... ... ..9 QSB QSB  hw? BK
>
_______________________________________________
UR RST IS ... ... ..9 QSB QSB  hw? BK
