"... Comparisons Between Ground Surface and Elevated Radials"

<< < (2/6) > >>

K6JEK:
Quote from: K5UJ on June 12, 2011, 07:23:25 AM

Quote from: aa5wg on June 11, 2011, 08:52:44 PM

I came across this information today while searching Google.  Here is the link:   http://rudys.typepad.com/files/qex-ground-systems-part-3.pdf
73,
Chuck


That's the guy who has gotten a lot of mileage from coming up with stuff about radials from modeling. 

...


Severns ran many, many careful experiments not just models.  All caveats apply and he included them in his series of articles -- forty meters, his ground conductivity, etc.  Even with the caveats, I find his results more interesting than endless EZ-NEC hypotheticals.

k4kyv:
Another difference between ground-mounted radials and elevated are that elevated radials become resonant to the operating frequency, while ground-mounted become non-resonant accumulators of ground current, effectively lowering ground resistance.  Therefore, a vertical with elevated radials will have narrower bandwidth than one with ground mounted. But a small number of radials still allows a  lot of ground resistance from the lossy earth figure into the equation.  120 ground-mounted radials, each 0.4λ, is said to be a near-perfect ground plane, but most AM BC stations don't even use that much. Ground-mounted radials are a diminishing-return phenomenon; the more wire you have already put into the earth, the less improvement is realised by putting in more.  120 radials is probably overkill; the consensus seems to be that you don't see much improvement past about 60.

Naturally, with fewer radials which add more ground resistance, you get wider bandwidth, but more rf is wasted heating up lossy earth.  The ground plane can be thought of as a shield to isolate the lossy earth from the radiating part of the antenna.  The more complete that shield, the better the isolation, the  less the loss, but the narrower the bandwidth until a point of negligible ground resistance occurs.

Another way to think of it is to compare a typical broadcast installation with 120 quarter-wave ground mounted  radials, to a highly elevated quarter-wave ground plane, such as a CB or 2m ground plane with 3 or 4 radials, mounted a half wavelength or more above ground. Each antenna shows good efficiency and low ground loss, but it makes logical sense that between these two extremes, the closer the radial system is to the surface of the earth, the more radials are needed to maintain complete isolation.

The ground plane may or may not serve as a missing portion of the self-resonant antenna.  A ground mounted half-wave vertical will resonate perfectly with no ground system at all, but a  lot of the rf is still wasted as heat due to the proximity of the lossy earth. With a 1/4λ vertical, the ground plane becomes an integral part of the resonant length of the antenna.  These two functions of the radials are distinct from each other. One elevated radial, usually called a "counterpoise", was  long used for Hammy Hambone 160m inverted-L antennas, dating back to the 20's and earlier. Its function was primarily to achieve resonance in the L-shaped wire, but it did little to shield the radiating antenna from the lossy earth. But raising the counterpoise several feet above ground improves overall efficiency, since the vertical part of the antenna is better isolated from ground by distance, and the counterpoise itself does some radiating as a low-to-the-ground horizontal antenna, along with the horizontal part of the inverted-L.

In addition, a 1:1 match many not necessarily indicate the best efficiency in the antenna.  Take, for example, the quarter-wave vertical.  Its base impedance is about 36Ω. With negligible ground loss, connecting a 50Ω coax transmission line directly between the base of the antenna and the ground system will result in a SWR of 1.39:1. Removing enough radials to increase the the ground  resistance to 14Ω, brings the total base resistance to 50Ω, a perfect 1:1 SWR. But now, 28% of the transmitter power is dissipated in ground loss. The additional loss in the coax due to the 1.39:1 SWR will be almost negligible. So you are much better off  running the SWR at 1.39:1, than disconnecting radials to bring the SWR to 1:1.  You could still bring the SWR to 1:1 by inserting a matching network between the feed line and the base of the vertical, but if the transmitter can be loaded up to full power at 1.39:1 SWR, you probably have less loss in the coax than you would in the matching network that brings it to 1:1.

This problem becomes even more acute with a shortened vertical, which inevitably has lower radiation resistance, for example an 1/8λ vertical, whose resistance is about 12Ω. Now, disconnecting radials until a perfect match is achieved, means that 38Ω is effectively inserted in series with the antenna, yielding an efficiency of 24%. Even with the 4:1 SWR, the antenna will still be more efficient with negligible ground loss.  In this case, a matching network should substantially improve efficiency regardless of the number of radials used.

I downloaded  the whole series of QEX articles.  The writer claims that his modelling figures closely matched experimental data with actual test antennas and radial systems he erected for the purpose.

KA0HCP:
http://www.antennasbyn6lf.com/design_of_radial_ground_systems

Here is Rudy's actual website, which is easier to read and has interactive posts from readers.

K1JJ:
Yep, tuning/optimizing  a vertical's SWR by adding or reducing radials is a very common mistake, as heard on the air.  As the radials are added and the vertical's input swings from >50 ohms down into the ~30 ohm area, the swr will get perfect and then move away again - until it is later matched at the feedpoint using some type of technique.

The following applies to coax-fed antennas, though openwire is basically the same if the tuner is set in the ballpark:

When it comes to general antenna tuning, I have an old rule.  First design the antenna using a Nec modeling program. Build the antenna to the model's specs. Then adjust it for maximum gain and/or maximum front-back.    The LAST thing to do is match the feedpoint to the coax. This may mean adjusting the length of the driven element and juggling the matching network to get a 1:1  50 ohm J0 match.  ie, SWR matching comes LAST because everything else effects it. SWR has very little effect on front-back and gain tuning, as long as it's in the rough ballpark when tuned for these other parameters.

Some guys have been known to adjust reflectors/directors, height above ground or lower the dipole legs to an inv vee, etc., to adjust their swr.  NG.

Once an antenna is positioned and adjusted for performance optimally, (if required)  THEN the swr fine-tuning matching gets done - last. The other parameters are not touched after the match is optimized except for possibly tiny fine-tuning adjustments for maximum f-b, if desired.

T

K5UJ:
Quote from: K6JEK on June 12, 2011, 01:09:20 PM

Quote from: K5UJ on June 12, 2011, 07:23:25 AM

Quote from: aa5wg on June 11, 2011, 08:52:44 PM

I came across this information today while searching Google.  Here is the link:   http://rudys.typepad.com/files/qex-ground-systems-part-3.pdf
73,
Chuck


That's the guy who has gotten a lot of mileage from coming up with stuff about radials from modeling. 

...


Severns ran many, many careful experiments not just models.  All caveats apply and he included them in his series of articles -- forty meters, his ground conductivity, etc.  Even with the caveats, I find his results more interesting than endless EZ-NEC hypotheticals.


I agree with all of this and  my comments may have been on the harsh side and thanks to Chuck for posting the link to the article by the way, and I saw the caveats etc.   I guess I don't understand his point because unless I misunderstood him, the radials are all only a few inches off the ground, less than two feet and his plots tell me things don't start getting good until you have a lot of them, not just four or six, but ~ 30.  Well, to me, he's modeling lots of radials close to ground and probably from a RF standpoint, the same as insulated wire radials lying on the ground, which is what Doty tested and wrote about in his paper I have a link to in the Printed Material section.   Doty did his work on a range; no modeling, because it was done around 30 years ago, before powerful computing was commonly available.  (Not saying anything bad about modeling!)  From a practical point, who's going to put up radials a foot or two high.  I must have gotten something wrong.

Navigation

[0] Message Index

[#] Next page

[*] Previous page