Brian Beezley, K6STI designed a 9-element Yagi for the 88-108 MHz VHF-FM band with a horizontal folded dipole and three reflectors. His antenna optimizer managed to achieve 7-8 dBd gain and F/R up to 30dB over entire bandwidth. The distance between first and last element is 100″ or 254 cm. I built two of them in 2013 for receiving vertically polarized signals.
Currently, a shortened version of this antenna is known as Körner 9.2.
The 100″ Wideband Yagi write-up is not available on Brian’s website anymore, so here are the details. All dimensions are in inches. Use insulated mounts for each ϕ10mm (3/8″) tubing element.
100" 9-Element Yagi Free Space Symmetric 88 92 98 100 105 107 108 MHz 11 6063-T832 wires, inches z = 20.8558 ; reflector height r0 = 38.59764 ; reflector half-lengths r1 = 36.38081 de = 31.14668 ; driven element half-length d1 = 26.07912 ; director half-lengths d2 = 25.77762 d3 = 25.34543 d4 = 24.59307 d5 = 21.915 de1 = 13.70858 ; driven element positions de2 = 18.02412 d1p = 22.33073 ; director positions d2p = 31.9882 d3p = 49.21343 d4p = 71.10764 d5p = 100 1 0 0 z 0 r1 z .375 1 0 0 0 0 r0 0 .375 1 0 0 -z 0 r1 -z .375 1 de1 0 0 de1 de 0 .375 1 de2 0 0 de2 de 0 .375 1 de1 de 0 de2 de 0 .375 1 d1p 0 0 d1p d1 0 .375 1 d2p 0 0 d2p d2 0 .375 1 d3p 0 0 d3p d3 0 .375 1 d4p 0 0 d4p d4 0 .375 1 d5p 0 0 d5p d5 0 .375 1 source Wire 5, end1
The antenna dimensions are also available in MMANA-GAL format: 9el_K6STI.maa.
Horizontal vs vertical polarization
A single dipole antenna has a different pattern depending on its polarization. In vertical it is omni-directional and in horizontal it has two lobes with deep nulls at ±90°. Similarly, Yagi antenna has much better directivity in a horizontal polarization.
This antenna additionally uses a horizontal folded dipole with its size optimized for horizontal polarization. Because of that, the 30dB F/R figure applies only to this polarization. The following plots compare both patterns for vertical and horizontal usage of a single antenna.
The 30dB F/B ratio is preserved in vertical, as both arms of the horizontal folded dipole are inline with the vertically polarized signals coming from the behind.
- aluminium profiles: 20x20x2mm (main boom) and 20x20x1.5mm (reflectors),
- 1.5mm duraluminium sheet for main and reflector boom joint,
- 10x1mm aluminium tubes (use 10×2 for reflectors) and insulated element mounts,
- 40x2mm aluminium tube (horizontal mast) and some u-bolts,
- many screws, nuts and washers (mostly stainless A2),
- plastic boxes, cable glands, ferrite chokes, F-connectors, 75Ω coaxial cable,
- T adapter and 50Ω coaxial for 37.5Ω -> 75Ω matching.
I think that bending a tubular driven element is the most difficult part of this construction. It is not so easy to bend a dipole that perfectly matches all dimensions. Finally, after some corrections, I managed to obtain a precision of each driven element of around 1 mm.
Each tube end should be closed. The water may freeze inside an element and break it during the wintertime. I used some teflon plugs for that.
The main and reflector boom joint is made of two pieces of duraluminium sheets.
For each antenna I made a typical half-wave coaxial balun. The cable has a velocity factor of 0.84, so the matching coax segment length is 300/98/2*0.84 = 128.6 cm. Additionally I put a ferrite choke to reject any unwanted shield current. All connections at the feedpoint are joined with a special bi-metallic Cu/Al washers. Always use very short leads at the feedpoint, as they can introduce much stray reactance.
My coax has an aluminium braid, so I put a piece of F-connector on each cable in order to solder them together. The cable glands should keep the water away. Additionally I used a spray dedicated for PCB coating to protect all connections before closing the dipole box.
Stacking two Yagis next to each other in a horizontal plane makes the main-lobe significantly narrower. However the mutual interaction between elements in vertical polarization degrades the F/B performance, in the worst case to around 20 dB as the model suggest. Placing two antennas above each is unfavorable for DX-ing, because such antenna is not more selective to other signals. Stacking in a vertical plane has also another disadvantage – it is mechanically difficult.
Personally, I think that stacking distance 200 cm or a little less is the best for this configuration. The shorter distance, the wider main lobe is and the deeper nulls between main and side-lobes are. The longer distance, the bigger side-lobes are, especially at the top of the band. The changes in gain are negligible. My 2x 9-el setup has currently a stacking distance of 190 cm. The following plots compare the stack with a single antenna.
The signals coming from two antennas can be easily combined with a T adapter. The cable length between each antenna and the T adapter must be the same.
Combining two 75Ω signals gives 37.5Ω output. Matching 37.5Ω back to 75Ω is done with a 1/4λ 50Ω coax. I used Belden H155 cable with 80% velocity ratio, so the length is 300/98/4*0.8 = 61.2 cm.
Brian Beezley also designed a dedicated Stacked Vertically Polarized Log-Yagis. They have much better F/R performance, but also a lower gain and noticeably wider main lobe.
Model vs reality
The antenna patterns were measured with a modified Sony XDR-F1HD tuner and XDR-GTK software. The multipath influence is visible as a choppy signal on the following plots. Signal attenuators were required because of a very strong local signals (90.4 86 dBf, 95.2 82 dBf, 101.2 88 dBf, 104.3 104 dBf!!) that cause AGC clipping.
- 90.4 MHz: VOX FM Płock, 0.5kW ERP @ 6 km, with 16dB attenuator, peak 86dBf (att. to 56dBf):
- 95.2 MHz: Eska Płock 0.1kW ERP @ 6 km, with 26dB attenuator, peak 82dBf (att. to 56dBf):
- 99.2 MHz: PR Łódź, 30kW ERP @ 95 km, with 16dB attenuator, peak 62dBf (att. to 46dBf):
- 101.2 MHz: TOK FM Płock, 1kW ERP @ 9 km, with 30dB attenuator, peak 88dBf (att. to 58dBf):
- 104.3 MHz: KRDP Płock, 1kW ERP @ 5 km, with 46dB attenuator, peak 104dBf (att. to 58dBf):
- 107.3 MHz: PR24 Łodź, 1.5kW ERP @ 95 km, with 12dB attenuator, peak 52dBf (att. to 40dBf):
Reflectors made with 10x1mm tubes are the weakest point of this construction. Frequent vibrations weaken the material near element mounts. While it it safe to use 1mm tubes for the dipole and directors, the reflectors will eventually break after a year or two (and of course it will happen during winter, when you can’t easily replace it – photo below was taken in January 2015).
The 10x2mm tubes were unavailable in my nearby store, so I reinforced all reflectors with a short fiberglass rods (20cm x 8mm) at the element centers.
Furthermore, I filled the remaining space inside tubes with a spray foam.
Well… this was still not enough, because the longest reflector broke again in April 2016, but this time at ~¼ of its height. Perhaps the tube was already damaged when I attached it to the antenna, as I don’t see any other possibility. Nevertheless, I replaced the longest reflector in both antennas with an 10mm alumiunim rod and so far the problem is finally gone.
The following plot compares the pattern before and after the replacement. What a difference! The broken reflector was acting as a director, especially for the bottom of 88-108 MHz band.
I really like this antenna. If I could I would build another two 9-els for a horizontal polarization, instead of my Körner 19.3. I can’t utilize the excellent F/B performance because of a strong multipath interference and narrower main-lobe could give interesting results. Unfortunately, the horizontally polarized Yagis in a horizontal stack are mechanically difficult.
There are some highlights which I have received in central Poland via tropospheric ducting propagation using this antenna stack in last years:
- The Netherlands @ 900-1000 km (reports: 11.02.2015, 02.10.2015, 31.10.2015),
- The UK @ 1300-1500 km, incl. BBC R4 from Holme Moss @ 1443 km with RDS (read full report),
- Norway: 98.6 Radio Sør from Kristiansand @ 968 km, listed as only 0.5 kW ERP (read full report),
- 15 stations from Italy @ 1200+ km and other from BIH, HRV, SVN (read full report),
- …and many, many other stations which are listed in my tropo log.
The following video presents a full rotation of my initial 2×9-el setup in July 2013. An extended stacking distance over 220 cm provided a sharp main-lobe with deep nulls, but also an awful side-lobes, reaching -10dB at the top of 88-108 MHz band.
Last update: March 2017