Importance of ground for K9AY antenna
I tested a temporarily homebrew K9AY-antenna and compared its behaviour with and without ground. The result is:
The K9AY antenna doesn't work without ground.
It loses the f/b directivity.
It needs good ground; just like the EWE.

This has been proved and documented and lead to the development of the pennant antenna, which is better in this regard.
Pennant and flag antennas are ground free antennas. I can recommend these despite of their larger dimensions.
They are easy to home brew. Description of the pennant antenna: In English see QST 7/2000, in German see funk 7/2000. Or here:
http://www.angelfire.com/md/k3ky/page37.html
Roland.Burkhard, Switzerland, hcdx list, July 26, 2000



In general, the K9AY loop does need the direct connection to ground. However, some have experimented with disconnecting it, and at times it gives better performance.
As far as "good" ground goes, if you look at Figures 7 & 8 in the original QST article:
http://www.arrl.org/tis/info/pdf/9709043.pdf

The plots show that the F/B actually is better over "poor" ground compared to "very good" ground.
The F/B ratio of this antenna is nowhere near as good as a pennant or flag. And as you stated, the flag and pennant are much less dependent on the surrounding ground.
The strength of the K9AY in my opinion lies in the ease with which you can change directions, and if you have a Vactrol (in place of the terminating resistor), the ease with which you can optimize your nulls.
I invite any HCDX readers interested in the K9AY loop to join a separate discussion list on this antenna at:
http://www.egroups.com/group/K9AYloop
George Maroti, USA, hcdx list, July 29, 2000

I have found the same thing here along the Oregon coast with my EWE antenna. During the summer when the sandy soil gets dry, the antenna does not work as well. But let there be a day of even drizzel, then it improves. July/August are two of the worst months for that with little rain. As soon as I get the time I am going to put up a K9AY antenna in the backyard to see how it operates.
Patrick Martin, USA, hcdx list, July 29, 2000



John Devoldere in his excellent book Low-Band DX-ing describes two alternative ground systems for poor to very poor ground conductivity:
1. Several, instead of one, ground rods, separated by at least their length. So, instead of say one 1-meter long rod, use five 30-50 cm rods interconnected in a dice-of-five pattern. Use the middle rod for connection to the antenna system (balun, transformer).
2. A large ground mat made of large strips of chicken wire, e.g. in the shape of a cross measuring approx. 6x6 meters.
3. A large number of short, interconnected radials laid on the ground.
I suspect for points 1 and 2 one way to improve conductivity is to use coarse salt in the holes where the ground rods are put down (or where the chicken wire is laid down), and to water the area (preferably with salted water or seawater) to keep it moist.
I have used the multiple ground rod systems with reasonable success. My ground consists of rock/pellet below a 10-cm soil.
Bjarne Melde, Norway, hcdx list, July 31, 2000



As for ground, the group of K9AY users out here on the west coast of Washington have been putting down a ground plane of wires under the loops with a ground stake. You could duplicate this in your garden and bury the wires. Don Nelson, Oregon, USA, k9ay list, July 30, 2000

Tracy Gardner: In the article "Pennant Antenna with Remote Termination Control" Mark Connelly state that the Pennant antenna may not require a ground to function, I'm not at all sure that the K9AY needs anything like as good a ground as Mark seems to believe. Perhaps he should have tried some radials under his K9AY?

Mark Connelly must be commended for providing a very detailed study into the behaviour of a remote controlled terminated antenna. In-fact most of Mark's observations are also pertinent to the remote controlled termination of the K9AY.
My real points of concern are that the alleged degrading of the K9AY F/B due "fussy Earthing" are negated by the use of a 4 radial counterpoise.
The proposed Pennant design is seriously limited by only offering uni-directional reception over 90 degrees. The gain of the Pennant is significantly lower than a K9AY of the same size.
Also a single ended amplifier is likely to cause in-band second-order intermod compared to push-pull RF amps (a push -pull RF amp typically has 30-40dB lower second-order intermod compared to a single ended design).
Also placing a 30dB amplifier in-front of a typical Comms Rx will considerably raise the noise floor under weak signal conditions. In contrast the 13dB amplifier supplied in the Wellbrook K9AY [please note: no longer sold] provides very low noise performance and significantly improves the transcontinental DX.
Furthermore, the Wellbrook amplifier has an IP2 of about +74dBm. I have now decided to make this amplifier available to all K9AY users.
Andy Ikin, Wellbrook, k9ay list, August 3, 2000

Having used EWEs for several years, I noted the need for constant surface watering during the dry months to maintain an adequate ground. I was hopeful that the new K9AY, with the closed loop construction would be advantageous (I'd already tried the flag; only recently the pennant). My experiences with ground plane requirements differ from your statements, but I am will point out one difference in the sites at the end of this message...
The units owned by John Bryant, Guy Atkins and myself are the commercially available units from Andy Ikin.

The results we've seen out here on the West Coast, with the K9AYs installed on former Sand Dunes a thousand feet back from the shoreline, gave variable results over the past year using just the 4-wire counterpoise suggested by Andy Ikin (Tests were made using Andy's units that three of us have, including one 4-loop model) in December, March, May, June, July, September, and most recently, November.
While some installation sites varied by a few yards over this timeframe, most were sited identically within a foot. The K9AY worked well when the sandy soil underfoot was extremely saturated (including one installation in a foot of standing water December 1999, and performing better than the onsite beverages) in the wettest of months(Dec,March,May) but in the drier months of June, July and September the reception was marginal, never beating the beverages onsite. At the site this past weekend, the K9AY performed poorly so that John Bryant abandoned using it, and Guy Atkins and I did not even bother to put each of ours up. (John noted only one reception where the K9AY made a difference, and here it was not F/B but the apparent lower noise floor in spite of lower signal strength, when compared to one of the beverages ).

Since the K9AY antennas have the same mast height and installation wires, identical heads and controls, constant lengths of cable , and identical sites, this leaves only the ground as a variable. Since the material underfoot is packed sand (former dunes, about 350 meters from the high tide line) over rock, the only variable is ground water saturation at the surface (notably lacking in the past 6 months as we've had a dry summer and no rains to speak of this winter).
In May, I installed my K9AY at the edge of the surf with approximately 900 feet of control wire/900 feet of RG8X, and compared it to a K9AY sited within 100 feet of the building using 75ohm hardline. There were no discernible differences in signal strength or F/B... we don't bother putting out that much cable anymore. Proximity to the Ocean is a key ingredient to successful TP MW dxing...we had hoped that signal strengths would be better the closer you were but this proved to make no difference.

Down the coast in Oregon, Patrick Martin permanently installed his K9AY about a mile from the Pacific Ocean. He's been doing some extensive testing using TP signals to determine how well his F/B and side lobes are controlled. He's noted that an extensive set of driven ground rods (he uses copper pipe as it does not rust away in the salt environment) rather than the simple counterpoise makes a substantial difference. His record of TP DX requiring nulling of US MW stations off the backsite(and down the coast such as the station at Coquelle, Oregon) is evidence of his ability. His initial results with the K9AY, as shared with the rest of us at a recent DXpedition, were said to no better than his existing EWE. Further he was not getting F/B nulling that was controllable. Improvement in grounds made a difference. I trust Patrick's results and controlled experiments of adding a ground rod at a time and making comparisions ("just how many ground rods are required? Wish I knew when to stop" (he mentioned he was at 11 at this time), Patrick speculated)
So, why do you get different results? I'll hazard to make the following speculation...

Its been years since I drove down Chincoteague Islands. Sure was obvious you're on a coastal salt marsh with a lense of saltwater floating on the freshwater water under the hard spots(plenty of wells)...this is different than our Pacific Coast which is sand over rock.
I am betting that you'd have an exceptionally stable, excellent ground even if you failed to put out the counterpoise..... Wish we had this at the site in Washington state .. but then we'd have too many mosquitos.
Don Nelson, Oregon, USA, k9ay web group, November 13, 2000



George Maroti: What Patrick Martin describes doing was the very reason for using a Vactrol for the remote termination resistor. In the original K9AY article, Gary mentions experimenting with different values of termination resistance, depending on one's ground conditions and particular frequency of interest. By adding so many ground rods, Patrick has completely stabilized that parameter, and can now use a fixed resistor. (Read more here)
For those in "average" soil conditions, a single ground rod with counterpoise wires should be sufficient, provided a variable termination is used.

Regarding the comparisons between the K9AY and beverages out on the West Coast; beverages work better over poor ground, so I think when those comparisons are being made, it may be that the beverages were poorer performers over the wet ground conditions, not necessarily that the K9AY was a better performer. It seems that if they were to set up both types of antennas on the next DXpedition, they'll always have an optimum antenna for whatever ground conditions they encounter.
If you read Joe Buch's original article, he describes that the K9AY loop is effectively working as both a loop and ground plane vertical whip, and the key to obtaining a null is to be able to have the "vertical whip" voltage cancel the "loop" voltage. The ground conditions will effect the "whip voltage", so depending on your ground conditions and termination resistor, you may not be able to cancel the "loop" voltage.

Another source for differences in performance may be due to the fact that we all are not using the same model Vactrol. I recall that some are using a model that is less than a hundred ohms on the low range. I use a model that is about 200 ohms on the low range, but is fairly linear between 200 and 800 ohms. Again, depending on your ground conditions, you may be in a situation where you need the lower range Vactrol. I'm not sure how linear that particular model is, it might be better for some, and not for others. It would be ideal if there were a Vactrol from 0 to 5 kohms, completely linear in response.

One way to quantify some of these variations would be for those who are on "good" ground to measure their Vactrol resistance for deep nulls, and then compare these resistance values with those who are on "poor" ground. Since I think most of us would rather be DXing than conducting these science experiments, it would be interesting to at least hear from various users how much variation in "tuning" of their Vactrol is required, either depending on changes in ground saturation, or when changing bands. In Pat Martin's case, he's pretty much at the point where he can used a fixed resistor. But, since he's primarily a MW DXer and concentrates on that band, he might have to change that resistance when he ventures into the tropical bands.
Still, even with some observed inconsistencies in performance, it's an impressive antenna for its size.
George Maroti, k9ay web group, November 14, 2000