You are completely right. I slipped back into thinking about your bench alignment setup rather than your real-world listening configuration. Thank you for pulling me back to the actual wiring of your antenna system.
Let's look at the actual physics of what you have built based on this exact, isolated loop: a 35-foot long wire, a metal roof, an ungrounded SX-28 balanced input ($A_1/A_2$), and absolutely no connection to earth ground.
## 1. You Have Built a "Balanced-to-Balanced" Antenna System
By connecting the long wire and the metal roof across the 28-turn primary, you aren't running an "unbalanced" wire-and-ground setup. **The metal roof *is* your counterpoise/ground plane.**
Because the metal roof is floating (not connected to earth) and your radio chassis is floating (isolated via the transformer and lack of a ground rod), your 28-turn coil is acting as a differential sensor. It is strictly measuring the RF voltage *difference* between the sky (the wire) and the house (the roof).
This explains why you have such clean reception on the AM band: you have isolated the antenna system from the houseΓÇÖs electrical wiring ground, which is usually the source of worst-in-class EMI noise.
## 2. Re-Evaluating the Series Capacitors (Why They Do Nothing)
Now that we know the OWON generator is out of the picture and your antenna source is a real-world wire-and-roof combo, we can look at the math behind those $0.01\,\mu\text{F}$, $0.022\,\mu\text{F}$, and $0.047\,\mu\text{F}$ capacitors.
At AM broadcast frequencies ($550 KHz to $1600 KHz), letΓÇÖs calculate the **capacitive reactance ($X_C$)** of that smallest $0.01\,\mu\text{F}$ capacitor to see what it actually does to the signal:
$$X_C = \frac{1}{2\pi f C}$$
* At 600 KHz, a 0.01 µf capacitor has a reactance of only **$26.5\,\Omega$**. * At 1400 KHz, its reactance drops to a tiny **$11.3\,\Omega$**.
Because $11$ to $26\,\Omega$ is an incredibly small impedance, it acts as a **dead short** to an RF signal on the AM band. The capacitor is essentially invisible. It isn't large enough to choke the signal, and it isn't small enough to introduce a noticeable phase shift or resonance change compared to the much larger inductive reactance of your 28-turn primary coil.
### When *Would* That Switch Do Something?
For that switch to actually affect your AM reception, those capacitors would need to be in the **picofarad (pF)** range (like $100\text{ pF}$ to $500\text{ pF}$), which would introduce hundreds or thousands of ohms of reactance to intentionally block or tune the AM band. Because they are microfarad-value caps, they pass AM and shortwave signals completely unhindered, which is why turning that knob feels like turning a switch connected to nothing.
## 3. How the 302-Turn Coil Feeds the Balanced Input (A_1 / A_2)
Because the SX-28's $A_1$ and $A_2$ terminals connect directly to the primary wrapper of the internal antenna coilsΓÇöand $A_2$ is jumpered to the chassis only if you use a standard unbalanced wireΓÇöleaving $A_2$ floating means you are running the radio in a true **balanced mode**.
Your 302-turn secondary is feeding a balanced, floating differential signal into the radio's front end.
* When you select a lower tap (like 197 turns), you reduce the voltage stepped up from the roof/wire loop. * When you dial it up to the full 302 turns, you are maximizing the magnetic linkage, dumping the maximum differential voltage across $A_1$ and $A_2$.
This high-voltage balanced injection is exactly what is saving your AM reception. Even though your early-type oscillator circuit is slightly mistuned internally due to the $462 KHz crystal shift, your homebrew transformer is providing a completely floating, noise-isolated, high-voltage RF punch directly into the grid of the 2nd RF tube.
Page: /1_need_help/Gemini_22/Gemini_22.shtml
Last modified: Tuesday, 21 Jul 2026