AI Analysis of Logs for Antenna Directionality

Where is this dang thing pointing?

Wire antennas generally have directionality broadside to the antenna but multi-band antennas have "lobes" of directionality that vary greatly as you go up in frequency from their fundamental frequency.  I can generate a KML file of my contacts from the Log4OM logging software, but it is a bit overwhelming and I don't have the ability to break it down by band.

I wondered what Google Gemini AI would determine from evaluating the contacts from my LOTW logs and how that differed from the theoretical directionality of the antenna.  I have over 2800 logged QSOs in LOTW that I provided for evaluation.

I have made a lot of contacts when portable but I paper log those and they rarely make it back into my electronic logs, so nearly all of these logs are from my station operation and 95% of those have been on the 80m OCF.

Gemini provided me with this evaluation

• 80m: Broad "figure-8" pattern with maximum radiation broadside to the wire at 55° (NE) and 235° (SW).
• 40m: A "four-leaf clover" pattern. Your strongest performance is at 45° and 315°, perfectly hitting Europe and the US Midwest.
• 20m: Four sharp lobes that pull closer to the wire axis. Your log shows high-gain "spikes" at 35° (Europe) and 330° (West Coast).
• 10m: Eight or more very narrow lobes. Your 10m contacts are extremely directional, primarily clustering at 50° and 130°.

I then asked Gemini to evaluate how my log results differ from the theoretical directionality of the antenna.  In general the results match the theoretical performance of the antenna.  I also asked it to include the results from 30m and 17m in this analysis:

• 80m, Broad Figure-8, Strongly Biased NE/SW: Contacts cluster at 55° and 235°. Coverage is broad but noticeably stronger toward the Northeast US/Canada. High: Matches the broadside radiation of a dipole.
• 40m, 4-Leaf Clover, 4 Strong Peaks: Contacts are highly concentrated at 45°, 135°, 225°, and 315°. You are successfully hitting Europe and the US West Coast via these distinct lobes. Very High: Confirms the 2nd harmonic pattern.
• 30m, Distorted 6-Lobe, NW Dominant: You have a massive cluster toward the Northwest (330°). This band is notoriously asymmetrical on an OCF antenna, and your log shows you've "found" the dominant lobe. Moderate: The theoretical pattern is messy; your data simplifies it.
• 20m, 4 Long, Sharp Lobes, Sharp Spikes: Contacts are extremely localized at 35° (Northern Europe) and 330° (Washington/Oregon). This matches the narrowing of lobes as frequency increases. High: Matches the "tilting" of lobes toward the wire axis.
• 17m, 6 Very Sharp Lobes , Strategic Clusters: Contacts cluster at 13° (New York/New England) and 225° (Mexico). The narrowness of these clusters indicates you are operating within high-gain "fingers" of radiation.  High: Confirms the 5th harmonic pattern.
• 10m, 8+ Needle-Thin Lobes, Pinpoint DX: You have specific, isolated successes at 50° and 130°. Many other directions show "nulls" where no contacts were made. Moderate: High-QSB (fading) makes this band less predictable.

Conclusions

Using AI to provide me with analysis of my antenna directionality was certainly faster than manually breaking down all the data and plotting it myself.  I can also see how much I'm likely missing on 20m and 10m due to the severe directionality of the few lobes.  I would be well served by having separate antennas for those bands, but I am very limited where I can put outdoor antennas given the lack of tall support structures at my house. The OCF works because the tallest thing in my lot is the peak of my roof and I have too short a distance to the tree in front of my house for a fan dipole to work.

I did look at some tools for evaluating directionality based on logs such as https://qsomap.org but the ones I found are visual and require manually evaluating the graphics.  There are likely other tools that would do what the AI is doing but I'm not aware of them yet.  Please leave a comment if there are log analysis tools that you use for this purpose.

The analysis of the 30m being severely clumped showed me that I could likely do better using my attic antenna for that band, but I had pretty much given up using the attic antenna years ago due to high receive noise given it's proximity to all the noise in the house.  Now that I have a Loop on Ground receive antenna that allows for quiet receive I will begin using the attic antenna for transmitting and try to determine how its directionality differs for the for the WARC bands that the 80m OCF is extremely inefficient on.

That's all for now,

Lower your power and raise your expectations, or let AI tell you what you should expect.

DE AA4OO - Rich

Loop on Ground (LoG) Receive Antenna

Dirt Shark Antenna

How a Humble Wire on the Ground Can Transform Your Radio Listening

Ever find yourself locked in a frustrating battle, wrestling with all those fancy knobs and buttons on your shiny radio to bring a signal out of the noise, just to catch a faint whisper from that station that started out copyable? Or that elusive DX station is sending a call that you just can't copy because the SNR is like 3-6 dB.  You're not alone.  Our primary antennas often become indiscriminate collectors – drawing in not only the desired signals but also a relentless barrage of buzzing, static, and digital hash. In our increasingly noisy world, the conventional antenna, while a decent transmitter, can become a veritable noise vacuum.

My Noisy, Old-Faithful 80m OCF

My homebrew 80m OCF has been a very good antenna for me at my QTH.  I replaced my 40m OCF with it over 10 years ago.  Ice and wind have broken its supports a few times over the years but it has continued to be a good performer for me on all bands except 30m.  I still get regular reports where the receiving station is hearing me better than I'm hearing them.

Even with all that praise, it always has been subject to noise.  The balun is supported by a rope connected at the apex of my roof, so it is right up against my home.  My house is full of electrically noisy stuff, the worst of which is a treadmill worthy of an all-band WW2 radio jammer and a HVAC that has a noisy blower and gas furnace igniter that wipes out 5kHz segments on 20m in regularly spaced sections.  My neighbors have something that turns on and off and gives me a nice S8 noise in segments across 40m and 80m, usually happening after I've begun a QSO.

I enjoy a challenge as much as anyone but I realized I could do better.  I wanted an antenna optimized for receive rather than transmit.

Receive Only Antennas?

While researching, I read about a number of receive only antennas.  

• Magnetic loops- Too fiddly to re-tune when you change bands
• Beverage - Give me land, lots of land
• LNA augmented, phased verticals - Money, money, money
• Loop on Ground - Cheap, but they can't possibly work

Loop on Ground Antenna

Enter an unlikely Receive Only Antenna, known as the "Loop-on-Ground" (LoG). It’s a marvel of minimalist design – nothing more than a simple loop of insulated wire. Its genius, however, lies not in boosting signals, but in not hearing noise.

Think of the LoG as a more approachable, compact cousin to the venerable Beverage antenna. Harold Beverage's experiments in the 1920s, involving long wires hugging the ground, revealed the potential for noise rejection and clear signal reception that the LoG continues to explore.

So why would 60 feet of wire oriented as a square, fed at a corner, pressed into the ground and covered by your lawn make a good receive antenna?  Haven't we always been told that antennas work better, the higher they are?

I'm no expert on this (or much of anything).  The most useful information I found that demystified LoG antennas was on Matt Roberts KK5JY's website. 

A Loop on Ground (LoG) antenna rejects noise by primarily responding to the magnetic field of radio waves, not the electric field, making it less sensitive to common electrical interference from household devices (like TVs, computers, power supplies) that create strong electric fields. Its low-to-the-ground placement also helps it "see" less local electrical noise, effectively acting as a directional antenna with deep nulls, especially when oriented away from noise sources, significantly improving the Signal-to-Noise Ratio (SNR) for weak signals, according to KK5JY.Net. 

How to Make One

The keys to the success of the antenna is making sure it is electrically isolated.  Build a transformer using a binocular core with 5-6 turns (I used 6) of 30 AWG magnet wire connecting the wire and 2 turns going to the coax (instructions on KK5JY's site).

The coax thus becomes electrically isolated from the antenna.  I forgot to take a picture before I sealed up the box with RTV but the photo above is from Matt's site.  Stripping the enamel and soldering 30 AWG magnet wire is a challenge for me, especially in that small enclosure but I sorted it out.  I filled my enclosure with RTV to secure the core and make it extremely difficult to work on in the future :)

I used 60 feet of my surplus, insulated telephone wire, bonded 2 stainless steel terminals to the ends and laid it out in the suggested pattern to align with my TX antenna.  I have it placed about 25 feet away from my house in the back yard.  It is about 50 feet away from my 80m OCF.  It is secured using ground staples to hold the wire down into the grass.  It literally disappears in the yard.  I mean seriously disappears.  I didn't have the transformer with me when I laid out the wire and stapled it, and when I came back out it took me 5 minutes to find the wire.  You want to use a very small transformer enclosure so that it sits low in the ground so that your lawn mower won't destroy your hard work.

I used 75 feet of weather resistant RG6 75 ohm TV coax to get it back to my grounding point by the house where it goes into an arrestor before I use another 75 feet of coax to get it back up to my operating position.

I used coax seal on the connections both to the transformer and the arrestor.

Results

At my station I run it to my SDRPlay, using the SDR Antenna Switch and Front End protector I built 8 years ago.  

I clearly see signals on the SDRUno display that don't even appear on my Yaesu FT-DX10 waterfall connected to the 80m OCF.  The FT-DX10 has one of the best receivers in ham radio at this time, so it can dig those invisible signals out (barely) if I tune to what I see on the SDR, but if I switch to the audio from SDRUno they can be heard clearly.

Signals are being picked up by the LoG that are lost in the noise and are invisible on the waterfall of the FT-DX10 no matter how much I fiddle with the display gain and display peaking filters. But I can work them when I find them because that 80m OCF is a good performer as a TX antenna.

Similar to how I configured the SDRPlay to work with my Ten-Tec Eagle; SDRUno is feeding an IQ signal to CW Skimmer.  CW Skimmer acts as my cluster server and my logging software shows me what I don't see on the Yaesu's waterfall.  I click a station, either in CW Skimmer's display or from the cluster list and the FT-DX10 tunes to the station.  

If I can't hear it on the DX10 it is a bit of a pain to turn off the IQ output from the SDR and send it's audio to a speaker rather than CW Skimmer, but I can work the station receiving on SDRUno and transmitting from the DX10.  

Flipping the IQ on/off and changing the output is annoying so I have found a used DX-Engineering RTR-1A that I plan to put in series with the SDRPlay allowing the DX10 to listen on the LoG and transmit on the OCF, while protecting the SDR.  I'll make a video showing the results when it arrives.

Conclusion

I had read mixed results from other hams on forums discussing using Loop on Ground antennas so my expectations were not super high.  Maybe hams who aren't getting good results could try re-orienting their receive loops or maybe the transformer wiring could use improvements.  Maybe they just don't have as much local noise as I do, but for my station this is a game changer.

While I've had mixed feelings about my DX10 I will admit that it hears and cleans up noise far better than my Ten-Tec Eagle and KX3 ever did, but I didn't know what I was missing.  Since I had the waterfall on the DX10 I had stopped just slowly moving the VFO across the band.  I would just tune to a signal I saw on the waterfall.  I had no idea there was so much hiding in the noise.  

"Now I see" said the blind man

That's all for now,

So lower your power and raise your expectations... Or build a receive only antenna and see what you were missing.

73 AA4OO