Tuning antennas with ChatGPT

 I've messed with a little AI in the past couple of years since LLMs exploded into the mainstream a couple years ago. Results initially were not promising but you just have to ask it to not lie (hallucinate) and not ask it something too obscure (something it's not trained for). 

Short story long, I had an EFHW antenna that ran north-south and was long enough for 160m and went over some trees that are probably pushing 100ft. Unfortunately, I used cheap wire and it eventually snapped before I got around to replacing it with copper-clad steel(CCS). After that I put up an OFCD but that was dedicated to my Winlink station. Other than that I have a butterfly T2FD in the attic which is essentially a dummy load. TLDR, I only have one antenna and its committed full time.

I just finished my master's and it's the holidays so I have time to kill. There's plenty of house projects to do and the weather's great but night owls like me can only replay COD MW3 so many times. I had a remainder spool of the CCS wire of unknown length and a transformer ready for work. I decided to send it so I launched a mason line over the tree in my front yard by tying it to a rock, pulled the wire and got it all hooked up. The end of the wire was close to the top of the tree, held down by some 550 cord so it would be hard to lower it, shorten it, and bring it back up. I decided to pull it in, clamp it so it doesn't slip out, and re-terminate. 

I started by sweeping the antenna with a NanoVNA. To calculate where I was resonant vs where I needed to be, I could have fumbled around until I lost interest but I chose to ask ChatGPT what it thought of the situation.

I gave it the SWR readings I got with the NanoVNA, which were antenna long, and asked it a couple questions with the following prompt:

i currently have an end fed half wave antenna of unknown length that has 1.2 SWR at 6.5mhz and 1.3 SWR at 13.2mhz. approximate the length of the wire and tell me how much to remove to make the antenna resonant on amateur radio frequencies (7mhz and 14mhz)

the response was impressive, informational, and, as it would turn out, pretty accurate.

Good measurements — those two low-SWR points tell us a lot. We can approximate the electrical length of your end-fed half-wave (EFHW) and then calculate how much to shorten it so the resonances land on 7.0 MHz (40 m) and 14.0 MHz (20 m).

---

1️⃣ Estimate the current wire length

For a half-wave wire in free space, a common practical formula is:

[
\text{Length (ft)} \approx \frac{468}{f\text{ (MHz)}}
]

This already assumes a typical velocity factor ≈ 0.95 for bare wire in air.

From your measured resonance at 6.5 MHz

[
\frac{468}{6.5} \approx 72.0\ \text{ft}
]

Cross-check with the harmonic at 13.2 MHz

[
\frac{468}{13.2} \approx 35.5\ \text{ft} \quad (\times 2 \approx 71.0\ \text{ft})
]

✅ These agree very well.

➡️ Approximate current wire length:

~71–72 feet (21.6–22.0 m)

---

2️⃣ How much to remove to shift to amateur bands

Because resonance frequency is inversely proportional to length, we can scale directly.

---

🔹 Shift 6.5 MHz → 7.0 MHz (40 m)

[
\text{New length} = 72.0 \times \frac{6.5}{7.0} \approx 66.9\ \text{ft}
]

Amount to remove:
[
72.0 - 66.9 \approx 5.1\ \text{ft}
]

➡️ Trim ~5 ft (≈ 60 inches)

---

🔹 Check the 20 m harmonic (13.2 → 14.0 MHz)

[
72.0 \times \frac{13.2}{14.0} \approx 67.9\ \text{ft}
]

This is within ~1 ft of the 40 m calculation — exactly what we want.
Both bands will move together.

---

3️⃣ Recommended trimming strategy (important)

⚠️ Do NOT remove 5 feet at once.

EFHWs are sensitive near resonance, especially depending on:

• Height above ground

• Feedpoint impedance transformer (49:1 vs 64:1)

• Nearby objects

Safe trimming plan

1. Trim 12 inches

2. Re-measure SWR

3. Repeat in 6–8 inch increments

4. Stop when:

   • Minimum SWR is near 7.05 MHz

   • 2nd harmonic lands near 14.1 MHz

This usually gives the best on-band coverage.

---

4️⃣ Final numbers (summary)

Parameter	Value
Current length	~71–72 ft
Target length	~67 ft
Total to remove	~5 ft
Practical trim	4–5 ft total, done gradually

---

If you want, tell me:

• Feedpoint transformer ratio (49:1 or 64:1)

• Mounting height and orientation (inverted-L, sloper, horizontal)

• Whether this is bare wire or insulated

I can tighten this to within a few inches and predict your final SWR curve very closely.

I didn't take its advice and cut off 29.5in followed by 39.5in for a total of 69(nice) then rechecked. I asked it to recalculate with new measurements using the below prompt and received the following answer:

recalculate. we are now reading 1.168 swr at 7.03mhz and 1.407swr at 14.320mhz

Perfect — these updated readings are actually very clean and let us dial this in tightly.

You now have measured resonances, not estimates, so we’ll use proportional scaling rather than rules of thumb.

---

1️⃣ What the new readings tell us

• 40 m fundamental:
  7.03 MHz @ SWR 1.168 → antenna is slightly long for the center of 40 m

• 20 m harmonic:
  14.320 MHz @ SWR 1.407 → harmonic is too high, which is common on EFHWs and confirms end-effects and transformer influence

We will tune primarily off the fundamental (7 MHz); the harmonic will follow.

---

2️⃣ Calculate current electrical length

We’ll back-calculate length from the actual resonance, not the old one.

[
L = \frac{468}{f}
]

[
L_{current} = \frac{468}{7.03} \approx 66.57\ \text{ft}
]

➡️ Current wire length ≈ 66.6 ft

(This matches where you should be after earlier trimming.)

---

3️⃣ Target length for ideal resonance

Most operators aim for 7.10–7.15 MHz so the antenna is usable across the entire 40 m phone segment without an ATU.

Let’s target 7.12 MHz.

[
L_{target} = \frac{468}{7.12} \approx 65.73\ \text{ft}
]

---

4️⃣ How much to remove (precise)

[
66.57 - 65.73 = 0.84\ \text{ft}
]

➡️ 0.84 ft = 10.1 inches

✅ Recommended trim:

Remove 8–10 inches total, not all at once.

---

5️⃣ Expected result after trimming

If you remove ~9 inches, you should see approximately:

Band	Expected resonance
40 m	~7.12–7.15 MHz
20 m	~14.20–14.25 MHz
SWR	≤1.3:1 on both bands

Your current 14.320 MHz harmonic should drop by roughly:

[
14.32 \times \frac{7.03}{7.12} \approx 14.14\ \text{MHz}
]

Which is excellent placement for 20 m.

---

6️⃣ Trimming plan (best practice)

1. Trim 5 inches

2. Measure SWR

3. Trim 2–3 inches

4. Re-measure

5. Stop when minimum SWR is just above 7.10 MHz

⚠️ Stop early if:

• SWR drops below 1.15 at your desired frequency

• Harmonic falls below 14.1 MHz (means you went too far)

---

7️⃣ Bottom line

• Current length: ~66.6 ft

• Trim needed: ~8–10 inches

• Do not exceed: 12 inches total

• You are already very close — this is final-tuning territory

If you want, tell me:

• Transformer ratio (49:1 or 64:1)

• Height above ground

• Whether the wire is horizontal, sloper, or inverted-L

I can predict the final SWR curve shape and bandwidth almost exactly. 

I told it I was fine with the  readings thus far and was going to use the built in tuner on my radio instead of additional trimming. ChatGPT was supportive of this decision and gave me some fluff about it being an engineering choice not a compromise

So all in all, it told me to cut 6 feet (to go from 72ft to 66ft), I cut 5 3/4 ft and it then told me to cut 6 in. Overall, a very helpful tool and now I'm working FT8 and some WAS nets in my spare time.

Tell me, have you used AI to do anything ham-related?