[TowerTalk] Tuning raised radial verticals

Fri Sep 8 18:19:01 EDT 2017

I was motivated by Jim's post to dig deeper into measuring my 160m T 
antenna performance.  A number of questions herein for experts willing 
to dig thru this.

My 160m T has 8 elevated radials and I measured the currents, they vary 
about 3:1.  That seems significant.  Given the "make radial currents 
equal" advice,  what performance am I losing with unequal currents?   
How do I tweak my EZNEC Pro4 model to generate unequal radial currents 
without introducing loss?  Then, if there is more than a db or 2 azimuth 
asymmetry from the 8x 120' radials how would I equalize the current in them?

The T is PE insulated Davis CCS 13.5ga.  All 8 radials are elevated 10' 
using #12.5 aluminum electric fence wire and are 120'+/- 4' long.  
Pretty much a sweet spot in the N6LF analysis.  It is fed with a 
50:25ohm TLT UNUN.   Plus 3 high stack of 240-31 cores with 8 tight 
turns of RG142 (>4k ohms).  Buried 100' RG8 feedline in conduit.  
Measured resonant at the amp thru the feedline and TLT at 1807KHz 1.07:1 
swr, Sept 2017 at the end of a very dry summer, <0.5" of rain in 3 
months.  In wet ground conditions I recall the resonance being closer to 
1825KHz, but modeling shows that is not likely over a +/-5x variation of 
ground conductivity from "average".  Perhaps some other summer to winter 
change in the antenna or foliage nearby. Winter water table is at/above 
grade. The T top wire runs SE-NW.

The measurements were with an MFJ-854 clamp on RF Current meter (got to 
curse the ferrite clamp design first, then found a big instrument error 
see below)
My ProIII says 25w output, the amp (2K-FA) says 16w.

Measured current and what each radial goes near/into:

East  0.068a  into woods
SE    0.075a into woods
S  0.035a into woods
SW  0.100a goes within 22' of a metal building corner, about a 45d angle 
to the building, then over lawn
W 0.115a goes within 10' of a metal building corner, then lawn at about 
a 15d angle parallel to the building, and ends 1' short of a 89' tower 
(actual 116' long)
NW 0.120a goes within 50' of a metal building corner, then lawn, and 
ends 30' short of a 140' tower it is tied to
N 0.060a into woods
NE 0.090a into woods

Sum of radial currents is 0.66a.    Vertical current after meter 
calibration 0.65a .  Measured coax current ~0.002a (the MFJ ferrite 
won't fit over RG8, so I measured on the RG142), and barely moves the 
needle on the 30ma MFJ scale.    Currents measured at 1860KHz. SWR at 
1860 measured 1.09 to 1.16:1 depending on the day, thru the series 
capacitor tuning.  Note 4 different scales are used on the MFJ, so that 
introduces error and the meter is tiny so guesstimates abound.

The initial as measured currents sum of radial vs vertical was different 
by 0.13a.  This led to a calibration of the MFJ which discovered the 1a 
scale used for the vertical current measurement was reading high by 
20%.  The 100ma and 300ma scales were within 5% except below 30% of full 
scale, which I  think is error from the sense diode forward drop.  
However, each antenna measurement was very repeatable, even standing on 
an 8' aluminum ladder.  The calibration setup used 6, 10, 20db 
attenuators, a 50w Bird slug, 100w dummy load,  and IC706MkII.  In the 
test setup the MFJ was insensitive to the location of the split center 
conductor within the sense ferrite.  (btw some component values in the 
MFJ schematic don't make sense to me).

If I ignore N & S radial currents and compare sum of E vs sum of W than 
the imbalance is 0.233a E vs 0.335a W, maybe that is significant.  OTOH, 
prior modeling of the 160 T with the other 3 towers (89', 89', 142') all 
to the west of the T and at guesstimate electrical lengths (actual 
height plus top loading) yields a couple of db more gain to the east, 
without the metal building in the model.  So the towers may be acting a 
bit as a reflector.   When I put the antenna up, I added radials over 
time and noticed the largest change in resonant SWR with the one to the 
W.  None of the radials run parallel to buried conductors.

I agree that for 2 to 4 radials equal currents are desirable, unless an 
azimuth gain skew is desired as with a CrankIR "on the beach". Somewhat 
related is AC6LA's sample modeling with 8 equal radials shows minimal 
differences in azimuth gain (0.1db) due to the coupling to the T top 
wire for a T similar to mine.
https://ac6la.com/aecollection3.html

It also seems to me that there is a number of radials at which the 
concept of radial resonance makes little sense.  It is like asking "what 
is the radial resonance of a large sheet of perfect conductor or for 
seawater under a vertical".  And then as is seems likely in my case the 
metal structure couplings (towers, building) are the reason the nearby 
radial current is high, how would changing the resonant frequency affect 
the current?  Or would it be better (rhetorical question) to remove high 
current radials to force more "equality"  (the radials are pretty much 
equal angular spacing as installed)?  Or add more radials where the 
currents are low?

To force asymmetric current distribution over the ground, it is easy to 
remove radials in my model (AGT=1.01), so I did that.  With three 
adjacent ones removed, the imbalance at 26 deg elevation (peak gain) is 
1db in azimuth.  Using 5 asymmetric radials as a proxy for radial 
current imbalance, it seems that the affects are small.

Lot's of questions when a model meets a measured real world situation!  
Am I missing something?

Grant KZ1W

  9/3/2017 9:25 AM, Jim Brown wrote:
> Dan and Gary are two fine engineers, and, as usual, their work is 
> excellent. Gary's work was careful to study each element of the array 
> alone, with the other elements open circuit to minimize coupling from 
> them, but also to study each element as part of the array. This is 
> textbook, and exactly right.
>
> Any study of radial systems should also include the extensive and very 
> disciplined work of Rudy Severns, N6LF. He has done multiple studies 
> that attempt to model what he measures, then measure what he models, 
> then model again, and measure again, with each study learning more 
> about the various mechanisms at play.  One of the subjects of Rudy's 
> work was elevated radial systems, and it was published in two parts in 
> QEX.  It is required reading for anyone implementing raised radials.  
> One of Rudy's most important observations was that it is very 
> important to make radial currents equal, that the variability of 
> ground under the radials can affect that a lot, that 4 radials is 
> minimum and 8 is better, and that it helps to tune radials slightly 
> short of a quarter wave.
>
> N6BT has also done a lot of work with verticals and elevated radials. 
> I had installed a couple of 160M verticals with radials elevated only 
> about 6 ft, and was not getting the gain predicted by the model. He 
> advised me that for 160M, radials should be elevated by at least 16 
> ft. I raised mine to an average height of about 20 ft (gull wing 
> style) and gain did improve.
>
> http://www.antennasbyn6lf.com/design_of_radial_ground_systems/
>
> 73, Jim K9YC
>
> On 9/2/2017 12:41 PM, Bryon Paul Veal NØAH wrote:
>> I never have heard of a successful array using raised radials ... so 
>> this
>> has been a good exercise.  Also strongly agree with tuning the 
>> verticals to
>> the intended frequency  ..... but this is all with comtek. From the 
>> power
>> dump port,   I had  300 foot roll of coax coming back to my meter 
>> measuring
>> the power dump onto the dummy load in my shack!
>
>
> _______________________________________________
>
>
>
> _______________________________________________
> TowerTalk mailing list
> TowerTalk at contesting.com
> http://lists.contesting.com/mailman/listinfo/towertalk