Build your own K9AY

George Maroti
antennas@qth.net list, February 12, 2000

I've had my homebrew K9AY loop up now for 2 weeks, and decided it was time to pry myself away from the radio to share my results and impressions thus far.
All listening was conducted in Mount Kisco, NY, 40 miles north of New York City on a Drake R8B.

Medium Wave
In short, performance of the K9AY is miraculous.
I can easily and consistently separate two co-channel stations on a given frequency.
Some of the best examples are 530, 780, 860, 1000, and 1020, where I can switch between a stateside station (usually to the north) and a station in Central or South America.
As far as Transatlantic DX goes, I've only detected carriers; Spain 684 and Morocco 1044 appear regularly.
I have no other MW antenna to compare it to, but judging from the K9AY loop's performance, I don't really need any other antenna for MW.
One observation here is that sometimes I can obtain a better null when switching to the East-West loop for stations to the North and South and vice-versa.
It really pays to try all 4 directions to see which works the best.

160 Meters
I don't know code, so can't give any specifics on which CW stations I was hearing, however the loop can null most of them.
I didn't observe any co-channel activity, so I can't say if the loop would be effective at separating two co-channel signals.

120 Meters
I'm able to null WWV on 2500 kHz.

90 Meters
Performance on 90 meters and up is not as dramatic as that on MW.
My main goal was to try to separate R. Maya de Barillas in Guatemala on 3324.8 and R. Bougainville in Papua New Guinea on 3325, but I'm not able to null the Guatemalan.
On other occasions however, I have noticed strong directional characteristics of the loop; as an example I'm able to null Sierra Leone on 3316.
On 90 meters, my best performing antenna is a half wave dipole 45 meters in length. Compared to the K9AY loop, the 45 meter dipole is about 2 S-units stronger. However, the loop's strength is its ability to reduce or eliminate sources of noise.
Most recently, there was a nice opening to India between 0015 and 0145 UTC, and the K9AY loop was my preferred antenna as the signal-to-noise ratio was much better. I was able to hear Lucknow 3245 and Bhopal 3315.
Again, the optimum loop and direction for the most readable signal didn't necessarily follow the great circle path.

75 Meters
On this band, my T2FD cut for the 90 meter band outperforms my 30 and 45 meter dipoles. RRI Merauke 3905 is usually audible during my local sunrise.
Although the T2FD's signal strength is stronger than the loop, the loop wins out at times, as it can reduce the ARO QRM on LSB, and noise.

60 Meters
As expected, the 30 meter dipole's signal strength is 2 S-units higher than the K9AY loop.
Again, depending on the noise, the loop can usually reduce it so that the signal is better on the loop.
Trying to null the noise in the tropical bands is not as easy or pronounced as finding the null on MW, you really have to try all 4 directions, and slowly adjust the terminating resistance.

Overall Shortwave Performance
I was able to observe directionality with the loop on 5950 kHz, where on the south loop I can receive the WYFR transmitter in Florida at S9/+20 dB, while reversing the loop direction drops the signal to S9.
Constructed with 85 foot loops as in the original QST article by Gary Breed (K9AY), the K9AY loop is a very good "compromise" or broadband antenna for all shortwave frequencies.
It should definitely be the antenna of choice for those with small lot sizes.
In my case, the loop is not always the best performer, so I'm keeping my dipoles and T2FD.

Compared to a flag antenna, I definitely prefer the K9AY loop. For me, the flag was "deaf" on the tropical bands. I can usually live with the 2 S-unit reduction of the K9AY.
On MW, the flag exhibited a nice front-to-back ratio, similar to the K9AY. The K9AY is simpler to build than an equivalent four direction flag or pennant array, and requires less space.
Personally, I'd like to see more work in the area of optimizing loop size and dimensions for even better shortwave performance. Once you are able to cleanly separate two co-channel stations on MW, it is easy to get greedy and hope for similar performance on 90 or 60 meters.
I hope others will share their experiences in this area.

Another area for discussion might be the different techniques for grounding the antenna.
I'm wondering if anyone has tried using elevated radials with their loop, essentially suspending everything in the air, and not connecting anything to a ground rod.
I may be forced to do this during the warm weather months as how I have my loop mounted now will hinder the use of my backyard.

Construction Details
My design borrowed elements from the original K9AY loop, Al Merriman's loop and the Wellbrook Communications design.
My loops are 85 feet in length, and are roughly equal on all sides.
The horizontal portion of the loops are only 1 foot of the ground.
My ground rod is a half inch four foot copper pipe driven 3 feet below ground. In addition I have one ground radial directly under each loop to supplement the ground rod connection.

For the controller box, I used a 9 volt Radio Shack AC/DC adapter, two toggle switches for selecting the loop and direction, a 1k ohm potentiometer that controls the Vactrol terminating resistance, and a four conductor telephone cable was wired to the relay box to control the two relays and the Vactrol.

The main feature of the relay box is that I used a Vactrol (Type VTL5C2) in place of the terminating resistor. This is an absolute must for getting the best nulls in any given situation.
I've observed that different values of terminating resistance is required when going from band to band, and even day to day the resistance varies for the same frequency. Thanks go to Al Merriman for coming up with this essential design modification.
Also mounted in the relay box is a 9:1 impedance matching transformer, two 12 VDC relays for switching the feed point and termination of the loops, and an SO-239 connector for the 50 ohm coaxial cable connection to the radio.

For the 9:1 matching transformer I used an FT140-43 toroid core from Amidon, and used solid #24 telephone wire for winding the primary and secondary.
As suggested in the Topband reflector for matching flag antennas, I didn't overlap the windings.
Rather, I wound the primary on one side of the toroid, and the secondary on the opposite side of the toroid.
This type of transformer is a great asset in reducing local noise into the antenna. With this I'm actually able to DX even while working on the computer, even with my radio directly next to the monitor.

Acknowledgements
Thanks go to Gary Breed, K9AY for his original design, and Al Merriman and Andy Ilkin for their modifications.
Thanks also to Guy Atkins, John Bryant, and Tony Ward for sharing their successful experience with their K9AY loops, which served as motivation for me to build one and try it for myself.