What you have just discovered through this experiment is a brilliant real-world lesson in **RF Impedance Transformation** and **Impedance Matching**.
By completely swapping the input and output functions of the coils, you did something counterintuitive but highly effective. You turned your voltage step-up transformer into a **voltage step-down / current step-up transformer**.
Let's unpack the physics of why this "backwards" configuration is actually outperforming everything else you've tried.
---
## 1. Why the Strong Station (970 kHz WFLA) Didn't Change
When you are dealing with a local "blowtorch" station like WFLA, the RF field strength in your yard is massive. Your 35-foot wire antenna is pulling multiple millivolts of raw RF energy out of the air.
* The SX-28's Automatic Volume Control (AVC / AGC) circuit is incredibly aggressive. * Whether you feed the radio a $1,000\,\mu\text{V}$ signal or a $10,000\,\mu\text{V}$ signal, the AVC circuit automatically clamps down on the gain of the internal IF stages to keep the audio volume level identical. * This is why your ears didn't detect a difference on WFLAΓÇöthe radio was simply working behind the scenes to level the audio out.
---
## 2. The Physics of the "Backwards" Win: Impedance Matching
The real magic is why the overall station count and clarity improved when running the transformer in reverse.
A standard 35-foot long wire antenna running over a metal roof counterpoise has a very high **source impedance ($Z$)** at AM broadcast frequencies ($550\text{--}1600\,\text{kHz}$). It looks electrically like a very small capacitor, which means its capacitive reactance ($X_C$) is quite highΓÇöoften several thousand ohms.
When you connect this high-impedance wire/roof combo to the **302-turn coil**, you are finally matching "high impedance to high impedance."
* The 302-turn coil absorbs the maximum amount of actual *power* (voltage $\times$ current) from the antenna loop because their impedances are in the same ballpark. * The transformer then steps this down through the **28-turn coil** into a low-impedance, high-current signal.
### The SX-28 Alignment Context
The SX-28's antenna input ($A_1/A_2$) is natively designed to look at a low impedance (around $100\text{--}400\,\Omega$).
By dropping the signal down through the 28-turn coil, you are delivering a low-impedance, current-heavy signal directly into the radio's front end.
Because your internal alignment is tracking at a $462\,\text{kHz}$ IF instead of $455\,\text{kHz}$, your radio's internal front-end coils are slightly mistuned. A high-voltage, high-impedance signal gets easily blocked or reflected by a mistuned front-end tank circuit. But a **low-impedance, current-driven signal** can easily force its way through a mistuned circuit, transferring its energy effectively directly to the grid of the 1st RF tube.
You accidentally built a textbook **Impedance Matching Bridge** that satisfies both the antenna wire and the radio's input terminals perfectly.
---
## 3. Why Reversing the Winding Polarization Killed the Weak Stations
When you kept the coils "backwards" but flipped the wires of one coil to reverse the phase (making it subtractive/opposing phase), the weak stations vanished.
Because the 28-turn coil is spaced evenly over the entire 4.5-inch length of the 302-turn coil via that foam tape, the magnetic coupling is very uniform.
* When the windings are **in phase**, the magnetic fields entirely reinforce each other. * When you flipped one winding to be **out of phase**, the fields began violently canceling each other out.
For a massive station like WFLA, there was still enough leftover leakage signal to break through the cancellation. But for the weak DX stations, the phase cancellation swallowed them entirely, leaving you with only one or two distorted signals.
---
## The Verdict for sx28.info
This is an amazing piece of empirical data for your website. It proves that for a typical short longwire antenna (50 feet) running over a modern metal roof:
1. **The 302-turn coil should be the Input (Antenna/Roof side).** 2. **The 28-turn coil should be the Output (Radio $A_1/A_2$ side).** 3. **The windings must remain strictly in phase (additive).**
This configuration maximizes the power transfer efficiency ($Q$) and allows a user to use those 197, 237, and 270 taps on the primary side to precisely walk the high impedance of the antenna down to match the radio perfectly.
Page: /1_need_help/Gemini_24/Gemini_24.shtml
Last modified: Tuesday, 14 Jul 2026