[Top] [All Lists]

Re: Topband: Elevated Radials

To: Guy Olinger K2AV <>
Subject: Re: Topband: Elevated Radials
From: Art K6XT <>
Date: Sat, 30 Oct 2010 08:03:40 -0600
List-post: <">>
My 160 ant is a 1/4 wave 80M GP. A 1/8 wave GP. Currently 18 radials, 
all 1/8 wave on 160, which begin at 15 ft, end more or less 10 ft. 
Seeking ways to optimize flexibility I turned to EZNEC, discovering that 
adding a pair of short more or less horizontal wires attached 2/3 
distance up the GP would lower the feed Z by about 100 ohms while having 
small effect on the pattern. One wire skewed the pattern more than I 
wanted. 2/3 of the distance up because that is where tower ends and 
adjustable pipe tip begins.

Modeling the radials showed, as I expected, that more is better. To a 
point. I started with four thinking that would be an absolute minimum. 
It worked OK but as Guy points out plenty of my RF went into keeping 
earthworms comfortable. I have space so began adding radials. In the 
area of 10 radials there were no further, significant, changes in 
measured feed Z so I began to think the antenna was reaching a decent 
balance between gain and the work effort to keep adding radials.

I tried some 1/4 wave 160 radials without apparent effect so made more 
1/4 wave 80 radials of them. This confirms the suggestion in a paper 
many years ago in, IIRC, Ham Radio that short radials work.

Worthy of note is I have a philosophical objection to the work effort 
required to lay in a bed of 160 buried radials (I could), which is why 
the antenna is a GP in the first place. Rephrasing, I'm too lazy.

So at this point I had an antenna that fed fairly well on 160 but those 
two wires EZNEC recommended had negative effect on 80M. I started to 
build a through the coax powered relay setup. Fortunately N6ND 
intervened with the suggestion to use traps to isolate the two 
horizontal wires. This was the magic bullet. With traps there is no 
change in 80M Z, they work fine on 160, and I get two bands from one 80M 
1/4 wave GP. I did end up after a couple seasons' use doubling the 
number of radials - but saw no change in feed Z on either band and no 
subjective change in performance. The antenna continues to work most 
everything I hear on the rx ants.

I'm thinking if only I could hear 5R8X and 5X0CW, I could work them.

73 Art
Allison, Colorado
Don't make old people mad. We don't like being old in the first place, so it 
doesn't take much to set us off !

On 2010-10-29 21:01, Guy Olinger K2AV wrote:
> Short version:  Anyone who has four elevated radials can easily check
> this out.  How to in Long Version.
> 1/4 wave elevated radials can have 18-60 ohms ground induction loss
> per radial.  Divide by number of radials for series loss.
> 1/8 wave elevated radials can have 5-12 ohms mostly ground induction
> loss per radial.  Divide by number of radials for series loss.
> A single elevated 5/16 wave folded counterpoise will have 2-3 ohms
> mixed wire and ground induction loss. +/- 32 feet.
> Long version:
> Nobody is telling you to take down your radials or modify it if you're
> happy with it.  Far from it.  If you like it then of course you keep
> it.  Go work 'em and have a great time.  God bless you and have a
> great life and all that!   But people are asking for advice here...
> 1) What comes up over and over and over and over in correspondence
> (way, way more times than you guys that are happy with yours) is that
> people don't have the space for traditional radials and what they have
> tried just plain sucks (their appraisal, not mine).  In my case I do
> have the acreage for dense elevated radials other than the highly
> inconvenient fact that a drive way goes right through the middle of
> the available space and either elevated and buried radials are out of
> the question.  What then?  What FACTS to design an alternative to all
> of the traditional radial answers?  What would be the tradeoffs?  How
> to measure?  Is there anything possible besides 5 buried 15 foot
> radials I already know suck big time, that I can try?  Exactly what is
> it that makes that makes 160m antennas so unbelievably bad?
> 2) What does someone DO who is trying honestly to compare alternatives
> and the FIRST TIME AROUND squeeze the last dB out of what is going to
> be a considerable project, because they don't have the time to try
> five different things, one at a time, each one occupying all available
> space?  Even if they did it, what beyond QSO count affected by highly
> variable conditions would they have to compare sky-wave effectiveness?
> To your considerable credit, 227 countries on 160, even using an
> amplifier, probably proves the worth of the operator more than the
> antenna.  I've always said that there's anywhere between zero and 27
> dB gain possible between the ears.  If you've worked 227 countries I'd
> suspect you're one of those 27 dB gain between the ears guys.  That's
> the inverse of the old saw "What do you call a lid with an amp and a
> great antenna?  LOUD lid."  "What do you call a top op with 100 watts
> and a piece of wire?  DXCC"
> Doesn't it bother you just a little bit that there's nothing out there
> carefully verified that would allow you to calculate these kinds of
> 160m sky-wave things?  There hasn't been any serious money into
> research of these issues in our lifetimes.  The commercial MF
> interests are 99% concerned with ground wave which doesn't do us any
> good.  The buried radial aspects of MF vertical performance were well
> settled in the 1930's by Bell Laboratories and legendary figures of
> the time.
> FCC regs put the old research in concrete.  A station builder buys the
> property, constructs the antenna with dense buried radials per FCC
> rules, marks it all up as a capital investment in their tax records
> and moves on to worrying about programming and making a profit.  They
> are not funding research projects to see if there is a way to improve
> sky wave, or see how to do improve sky wave with elevated radials, or
> see how to accomplish efficient sky-wave with something other than
> radials.
> Since you already have four radials up and ready, here's a
> straightforward experiment you can do with them.  Temporarily separate
> the radials from each other and any connection to
> ground/shield/matching network connection.  Now just four 131' wires.
> Call them North, South, East and West for illustration purposes,
> whether they are or not, just so N and S are opposite wires in line
> and E and W are opposite wires in line.
> Connect the N and S wires together.  Connect the E and W wires
> together. Connect the N/S junction to the center conductor of a PL259,
> and the E/W junction to the PL259 shield using the SHORTEST
> connections possible, as in couple inches, preferably less.
> Put an analyst on that directly, or with no more than an inch or two
> of coax/coax connectors (very, very important). Support the analyzer
> in some non-conductive manner away from misc conductors or capacitive
> coupling to the analyzer case, or anything metal and then orient it so
> you can read it without touching it.   Read the resistive component of
> the impedance.  What is your R at that point?  If you look at the
> series/parallel nature of how the radials are fed, you will see that
> the feed R you are measuring is also the average individual radial R
> over your dirt.
> Try feeding the N/S pair as a dipole. Dividing this by two will give
> you a worse number, even though 5 to 10 percent of that power is
> actually radiated. So multiply half the dipole R by 90% and it's still
> probably a worse number.
> A NEC4 model of of the cross-fed 1/4 wave radials in FREE SPACE (just
> the radials) shows a feed R of about 5 ohms, radiation from this
> configuration at 1.98 dBi peak, and an average gain of -1.91 dB.  This
> loss would be wire loss and high circulating currents as that's all
> there is left to make loss.  The 2:1 tuned bandwidth SWR span of only
> 8 kHz shows how very high Q it is.
> On the other hand, model the same exact wires 8 feet above one of the
> Norton-Sommerfield grounds in NEC4, and it's behavior completely
> changes.
> Over one particular ground simulation, the radials' feed R is now 21
> instead of 5, peak radiation is now -17.2 dBi instead of +1.98,
> average gain is now -22.79 dB (yes, MINUS twenty two point seventy
> nine), and the 2:1 matched SWR is now roughly 60 kHz instead of 8 kHz.
>   This last is the classic broadening of a high Q circuit's bandwidth
> Models of other earth characteristics will fetch up to more than 60
> ohms feed R per radial.  Just depends on the dirt.
> Given all the above, how can you have the change if there is no
> current induced in the dirt?.  Why do we NOT put dipoles "up" at eight
> feet over ground?  Just exactly WHY does a dipole at eight feet
> totally suck?  It sucks for the same reason 4 radials in a cross-fed
> configuration at 8 feet somehow come up with 21 or more ohms of series
> resistance that is almost entirely loss in the ground.  8' high
> cross-fed radials, 100 watts, -22.79 dBi, is an ERP of one half watt.
> 1500 watts to the wires results in an ERP of 7.5 watts, or QRP
> radiated.  This is REALLY leaky dummy load grade absorption.  There is
> current in a resistor somewhere and all it can be made of is dirt.
> I can't tell you why you never heard about ground induction (induced
> ground currents) in school.  The polite answer would be by some other
> term?  The cynical answer would be that they didn't know that current
> was induced in dirt, and so they couldn't put it in the book.  Except
> Hal Beverage (and those other legendaries back when this WAS a hot
> subject for commercial money) documented current in the ground.
> You can do the experiment above.  Then you have a quantifiable
> characteristic of an elevated radial that shows 20 plus ohms ground
> induction loss per radial, where it is only mitigated by adding more
> identical radials in parallel.
> If one is so unlucky with THEIR dirt to hit one of those 60 ohm jobs,
> then their four elevated radials in parallel will have a 15 ohm loss
> to go in series with their 15 ohm radiation resistance antenna, and
> measuring a total of 30 ohms and a nice wide comfortable SWR curve,
> think they are in 160 meter heaven, never knowing that of that 1500
> watts, that 750 watts has gone to ground induction loss in the
> radials, and of the 750 left, 325 went below the horizon and was lost
> to reflection loss we can't do anything about, leaving only 325 watts
> spread around the horizon.
> If you were lucky enough to have a 20 ohm per radial figure, you COULD
> do as little 5 ohms, which would put 562 watts around the horizon.
> The comparable four times 1/8 wave elevated radials, shows less than 5
> ohms R per radial, for a series loss of 1 or 2 ohms.  That means you
> put 700 watts around the horizon.  You have to tune the radial left
> over reactance out with an inductor, but why isn't everybody saying to
> ditch 4 full length elevated radials for 1/8 wave.  (Secret: SWR is
> broader with 1/4 wave radials.  Good SWR equals good performance,
> right?  Except WHY is the SWR broader....)
> And for those poor souls who don't have the 90 foot square on their
> property to even do 1/8 radials?
> There is a single elevated 5/16 wave folded counterpoise that comes in
> at 2-3 ohms series loss and only needs + and - 32 feet from the feed,
> and has no more or no less reflection loss than either four 1/4 or 1/8
> wave radials. You only need one 5/16 folded counterpoise.  That will
> be up on a web page soon.
> If you can do the experiment with your 4 1/4 wave radials, the results
> would be interesting.
> 73, Guy.
> On Mon, Oct 25, 2010 at 11:04 PM, Hardy Landskov<>  wrote:
>> Guy,
>> All I know is I worked everything I heard with 4 elevated radials. I even
>> went to 8 elevated radials and no difference on transmit.
>> What is induction loss? Never heard that in all the college antenna courses
>> I took. And yes 99% of 160 antennas are not optimum by any stretch of the
>> imagination. But we do what we can and try and get some fun out of it. So
>> don't grouse about it.
>> I have 227 countries on that band from AZ so something is working.
>> Hardy N7RT
> _______________________________________________
> UR RST IS ... ... ..9 QSB QSB - hw? BK
UR RST IS ... ... ..9 QSB QSB - hw? BK

<Prev in Thread] Current Thread [Next in Thread>