Hi, Grant.
Your original was posted to the reflector as well, and I'll let this
go there as well.
3:1 current imbalance, whether scalar or not, is grotesquely large
with 8 evenly spaced elevated equal length radials. I presume you have
already looked for connection issues at the center of the radials and
insulator issues out at the ends of the radials, and have walked the
radials looking for vegetation hanging on the radials. You would have
fixed that and moved on.
I think Kirchoff is overkill for the problem. I'm not sure he applies
at all with radials coupled inductively and capacitively to "infinite"
planetary size current sinks like ground viewed as a geographical
conductor. I think a lot of RF earth current (hence also radial
behavior) is better understood by the way charge moves around in a
lightning strike than by a circular radio circuit. The math is
dreadful for a radial analyzed as a single ended circuit, but that's
our problem, not nature's. We have a history of wanting to promote our
modeling programs to "natural" law, and somewhere in our unconscious
resist any addition of other factors to our calculations which can
create square law increases in problem and solution complexity.
Some practical things: Radial field screw-ups generally have almost no
effect on pattern at all. I modeled your antenna finally increasing to
200 ohm (!!!) loads inserted almost at the center into all the radials
on one side. This is a very large distortion of radial function, the
degree of unbalance required to produce three times the current in the
unloaded radials as loaded. This whale-sized distortion only makes the
azimuth pattern favor the unloaded side by about a dB. A counterpoise
hopefully will return 99% of the current forced into it, with minimal
energy sent to the far fields to distort the pattern of the aerial
wires. RF gone to counterpoise losses will not largely change pattern
any more than a parallel feed of antenna and dummy load will change
the antenna pattern.
So while the discovery of current imbalance in the radials may be
unnerving, it could be a result instead of a cause. And it does not
seem to reach the level of why you're getting thrashed in the
pile-ups.
What's more likely going on is akin to work done at N1LN where THREE
towers needed to be detuned to get clean pattern all around from his
two phased verticals. At the time only one vacuum variable was
available for the detuning. The closest tower was detuned with that.
The next closest tower was adjusted with that variable cap, measured
for value, and then returned to the first tower and reset. Using caps
rated for 100 watt transmitters, if that, the measured value was
constructed with a combination of fixed caps and put in place.
Detuning the towers made massive changes in the verticals' tuning.
A 160 contest was coming up in a few days and no chance to get a
variable in place. (Isn't it always? Do any of us ever do anything
sufficiently in advance?) As the story goes, a multi-op effort,
sometime 10 or 11 pm Friday night, one of the operators stepped out
the back door to smoke a cigarette, and noticed a strong acrid burning
smell. Not seeing a fire he ran in to get Bruce to find out what was
burning. Long story to short, the 1500 watts on the phased verticals
put enough current on the tower detuning to totally
smoke/roast/charcoal the detuning cap made up from fixed caps and
stink up the entire back acreage with the smoke therefrom.
The Acom 2000A had auto-tuned around the changing Z to the verticals
as the caps burned up, leaving the smokey back acreage as the only
tangible evidence of what was going on.
Without specific work to prevent it, RF current induced in a tower
will be driven into the dirt at the tower base, via the tower members
themselves, and via capacitive/inductive coupling of dirt from control
cables and coax shields. The circuit equivalent is a parallel feedline
to a big transformer (verticals inducing the tower) connected to a
huge big resistor made of dirt.
Since you have NEC 4.2, no excuse for not doing a whole property model
of all conductive materials and seeing where the current is flowing.
You can validate this by doing current measurements on every conductor
you can reach.
I did a whole property model for Bruce and it clearly showed a
completely unexpected high current on all the towers and conductors at
the base. It immediately pin-pointed the need for multiple detunings.
It also predicted severe far field pattern distortion that was
verified by measurements on a Potomac Instruments FIM-41 commercial
field strength meter at distances 2-5 miles in various directions.
A tower only 250 feet from a 160 vertical should be considered a
close-coupled transformer "winding".
Getting back to where you are
Can you supply the current reading and compass orientation of each of
the radials? There may be something suggested by that data. Also can
you run the following and report those results. (I'm including
rudimentary EZNEC instructions in case others want to do this. I doubt
you need them.)
1) In your model, open the "Loads" window. Add a load to each radial
in the second segment out from the center point. If the upper left
corner of the "Loads" window does not show "Loads (RLC)", then in the
"Other" drop down menu click "Change Load Type" and then click "RLC"
and then "OK".
2) Start with "Short" in all the RLC boxes. If they are anything else,
enter 0 as a value and hit return. Set all the "Config" and "Ext Conn"
to "Ser"
3) Do an "FF Plot" with a 3D "Plot Type" to get the 3D "Average Gain",
displayed beneath the selection grid on the main control window. Mark
that value down as the theoretical minimal loss.
4) On the main window, "Options" menu > "Power Level", adjust the
power level so that in "Load Dat" the "Current =" value is the same as
the largest of your actual measured currents.
5) Using the "Load Dat" button to pump the calculation engine, adjust
each load's R value up in resistance just enough to get the current
values you are measuring on the actual radials. This may cause the
current value of max value radial load to vary. Readjust the "Power
Level" setting to return the max radial current value to measured.
6) When you get the "Load Dat" currents to match measured values all
around, then rerun 3) and mark down "Average Gain" as theoretical
problem average gain.
7) Subtract the value from 3) from the value from 6) and mark it down
as theoretical additional loss in dB from the unbalanced radials.
8) Change "Power Level" to 1500 watts. Click "Load Data" and then add
up the eight "Power =" values. That value would be the theoretical QRO
operating losses in watts implied by the scalar current ratio's you
were getting. I suspect that is a large loss. But do the exercise.
Please pass along the results. Also if you create a whole property
model, please send the dot EZ or dot NEC to me as an attachment.
73, Guy K2AV.
On Mon, Jan 7, 2019 at 3:28 PM Grant Saviers <grants2@pacbell.net> wrote:
>
> Guy,
>
> I need some more db's on Tx. For Rx I hear much better than heard into
> EU from Seattle area. It's a hard path and easy to believe EU QRN/QRM is
> the main culprit. Your "loss list" is a great list, but I am thinking
> about a different potential problem with my T with eight 125' long
> elevated 10' radials, pretty much even in directions around the compass.
> It has a good choke and buried feedline.
>
> I have 3:1 measured current unbalance with an MFJ RF current meter I
> calibrated. Perhaps unavoidable given the forest, lawn, towers, and
> buildings near where the radials run. Plugging those values into Eznec
> 4.2 as a source for each radial yields insignificant pattern distortion.
> Kirchoff's law is ok on the actual values, sum of measured radial
> currents = vertical (85' to top of T, 33' each side top loading).
>
> N6LF with his modeling shows that radials longer than 1/4wl can cause
> significant losses. Now it occurs to me that the MFJ gave me scalar
> current values and those are not necessarily the actual i+jx radial
> currents. I think Kirchoff is happy as long as the eight radial plus
> vertical i+jx radial values all add to j0 which was true at resonant
> frequency where I was measuring the currents. So perhaps some of my
> radials are longer than 1/4wl RF and increasing the losses.
>
> So my questions for the wizards of top band verticals are:
>
> 1. Am I correct with my non scalar interpretation of Kirchoff's law for
> radials?
> 2. What are easy ways to measure current phase for each radial? (I have
> a dual channel scope and was thinking of making current probes of some
> sort).
> 3. Was my 4.2 pattern/gain analysis correct for the modeling of the
> unbalanced scalar currents?
> 4. Since the summed measured RF currents were correct, am I overthinking
> this about potential losses?
> 5. Do other Topbanders have experience with measuring radial current phase?
> 6. Most of the literature gives strong admonitions to "equalize" radial
> current "within a few percent" (ON4UN and others). Yet no analysis is
> given for why or how to do that. My modeling seems to disagree. My
> calculation of extra skin depth loss due to higher vs equal currents is
> only a few watts.
>
> W8JI advises to think about antennas as systems that include everything,
> and the more I learn the more the complexity of the system unfolds.
>
> Grant KZ1W
> A db here and a db there and pretty soon its 3db or more.
>
>
> On 1/7/2019 9:07 AM, Guy Olinger K2AV wrote:
> > Apologies to all for delay in response.
> >
> > Losses related to ground and close dielectric materials remain the
> > single monster gorilla in the room for improving TX performance of
> > vertical antennas.
> >
> > Setting aside content on k2av.com relating to the FCP, the other
> > issues in that web page’s "Loss List" section apply to any inverted L,
> > e.g. not having a tree inside the bend, avoiding proximity to
> > dielectric materials, quality of whatever counterpoise, etc
> >
> > Simply having the RF current maximum on a vertical wire at the ground,
> > largely diminished at the top, is a loss issue, unless the in/on
> > ground radial system is EXCELLENT.
> >
> > If you are considering beefing up your radial system in stages, then
> > by definition your radial system is NOT EXCELLENT. You already know
> > the deficiencies, those are what you haven't done yet and plan to do,
> > sometime.
> >
> > The net effect of a collection of radial system deficiencies
> > separating you from EXCELLENT will be at minimum over Midwest USA
> > magic super dirt. Ground losses move in the direction of catastrophic
> > as ground quality deteriorates from super to average, to poor, very
> > poor, junk, awful, to unbelievably bad. The vast majority of the
> > country has ground quality that will make you pay if your radial
> > system is NOT EXCELLENT.
> >
> > If you have that magic, super 30 millisiemens dirt, you can do
> > anything and get away with it. Please do not tell us, just keep it to
> > yourself. We already have enough things to make us jealous or feel
> > bad. Enjoy all your good results with schemes that blow up in ordinary
> > places. I have no problem with your good fortune, I just don't want to
> > be reminded over and over how poor NC dirt is for RF.
> >
> > But please do not offer shortcut low-band advice, because you have no
> > idea what awful results your scheme might get in 2 millisiemens North
> > Carolina dirt, or even worse in the barely covered-over rubble that
> > sometimes passes for dirt in urban and some suburban building lots, or
> > in historic areas that have been built over previous ruins for
> > thirteen centuries, places where the skeletons of ancient kings have
> > been discovered buried in medieval church ruins found under
> > current-day parking lots.
> >
> > The enemy of a vertical radiator is loss. Specifically, 1) dielectric
> > loss, RF exciting increased electron energy levels without electron
> > movement between atoms and 2) resistive loss, RF inducing current in
> > conductive but resistive materials. Dirt usually has both. Further and
> > worse, UNLESS there is net field cancellation at the radial wire or
> > below (outcome of an EXCELLENT radial system), dielectric material
> > within a few millimeters of radial field conductors INCREASES
> > dielectric loss.
> >
> > RF fields at the feedpoint are huge, especially where a self-resonant
> > vertical radiator meets in/on ground radials. Radial deficiencies will
> > be costly close to vertical wire meeting counterpoise, further
> > multiplied by the "poorness factor" of the dirt underneath. That is
> > one of the gains of an elevated counterpoise, getting high fields away
> > from the damnable dirt where an EXCELLENT on/in ground counterpoise
> > just ain't possible.
> >
> > Another way to reduce the fields at the ground is to quit using the
> > length of the vertical aerial wire as a cheap way to provide tuning
> > for matching to coax. If (on 160) you use an 88' foot wire as a start
> > for the L's horizontal, you will move the current max to 1/16 wave
> > down from the bend in the L. You will have done two good things, 1)
> > reduced the feedpoint current, hence also the RF field at the base,
> > reducing the power loss by the square of the field reduction, 2) put a
> > fairly uniform current on the *entire* vertical wire, further reducing
> > takeoff losses in trees, buildings, etc, by having a much larger
> > percentage of the total takeoff energy high enough get to sky without
> > encountering ANY dielectric losses.
> >
> > How long to make the vertical wire? As long as you can while still
> > sufficiently avoiding dielectric and conductive materials. Losses from
> > higher current on a shorter vertical wire are outweighed significantly
> > by gains moving the feedpoint up in the air away from dreadful ground
> > losses only otherwise mitigated by EXCELLENT ground radial systems.
> >
> > There are also questions about the efficiency of ground radials on top
> > of typical root content of "woodsy" locations. Roots are high loss
> > dielectric materials. Controversial, but an EXCELLENT *elevated*
> > radial system in the woods is going to outperform radials on ground.
> > Buried radials in the woods outright dismissed by many as impractical
> > at best.
> >
> > The usual killer reason for abandoning or not adopting non-resonant
> > improvements to move current maximum well up on the vertical wire?
> > Desiring low SWR without tuning apparatus at the base of the vertical
> > wire. From the "Taming the exasperating Inverted L" section on
> > k2av.com:
> >
> > **An Efficient Self-resonant Inverted L
> > is NOT a natural 50 ohm antenna**
> >
> > Not even close. Think 20 to 35 ohms.
> > Varies with dimensions and environment.
> >
> > **Lower SWR does NOT
> > predict improved performance**
> >
> > A dummy load has perfect SWR, and at 100% loss
> > is a worse antenna than a light bulb.
> >
> > 73, Guy K2AV
> >
> > PS: How much wire does it take to be an EXCELLENT radial field? Has to
> > be dense, full size all around, uniform in all directions. There are a
> > lot of hams who have had the space and actually done that, put down an
> > EXCELLENT radial field, and are definitely getting EXCELLENT results.
> > Just remember that "merely decent" radials over typical poor ground
> > can evaluate to "disappointing."
> >
> >
> > On Tue, Dec 18, 2018 at 11:07 AM Todd Goins <tgoins@gmail.com> wrote:
> >>
> >> Hi Gary,
> >>
> >> Thanks for the help. I got the new wire in place roughly as you described.
> >> It is a few feet lower to give good clearance from the anchoring branch.
> >>
> >> It doesn't work very well. Last night it was much poorer on receive and
> >> transmit than my existing 43' vertical setup. I'm not sure what to think.
> >> Maybe the light of day will reveal a clue, but it was fairly quick and easy
> >> to construct and kinda tough to mess up but who knows...
> >>
> >> I'll put an analyzer on the feedpoint this morning and see if it looks out
> >> of whack. I was expecting it to be a least a little better but it clearly
> >> wasn't.
> >>
> >> 73,
> >> Todd - NR7RR
> >> _________________
> >> Searchable Archives: http://www.contesting.com/_topband - Topband Reflector
> > _________________
> > Searchable Archives: http://www.contesting.com/_topband - Topband Reflector
> >
_________________
Searchable Archives: http://www.contesting.com/_topband - Topband Reflector
|