The following has been reformatted using HTML
You are absolutely right to be confused. The AVC bus should **never** be positive. With no signal, it should rest at roughly 0 V (or slightly negative, around -0.5 V due to normal tube contact potential).
What you are observing is a classic vintage receiver fault, and it is the exact reason your sensitivity is terrible and the audio is fading in and out.
Here is the mechanical breakdown of what is happening, why the S-meter is acting that way, and how to track down the fault.
## 1. The "Fight" on the AVC Bus
Your AVC system is actually working, but it is fighting a massive DC leak.
Right now, a fault in the radio is leaking **$+1.4\text{ VDC}$** onto the AVC line. Because the AVC line connects directly to the control grids of your RF and IF tubes, this positive voltage is driving those grids positive.
* Vacuum tube grids are never supposed to be positive. When they are, they draw grid current. * This grid current acts like a heavy resistor placed directly across your meticulously tuned RF and IF coils, killing their "Q" factor, destroying the receiver's gain, and deafening the radio. This is why a $20\,\mu\text{V}$ signal is barely audible.
When you crank the signal generator up to $530\,\mu\text{V}$, the RF signal is finally strong enough that the AVC rectifier diode creates $-1.4 V. This negative voltage cancels out the $+1.4 V leak, bringing the bus to $0 V.
* At exactly $0 V, the grids stop drawing current. * The receiver suddenly "wakes up," the gain returns, and the S-meter (which measures the plate current of those tubes) finally begins to register.
## 2. The Antenna Ground Issue
You noted: *"No ground connection to A1 or A2."*
If you are injecting a signal directly from your generator/attenuator into the SX-28, **A2 must be strapped to the chassis ground terminal (G)**, and your signal generator's shield/ground must also connect to the chassis.
* The antenna coil primary sits between A1 and A2. If A2 is floating, the primary coil has no reference to the chassis, meaning the signal transfer to the secondary (the tube grid) is severely crippled. * *Note: If you are using your custom isolated PVC matching box, A1/A2 can float, but for direct bench injection with an attenuator, A2 needs to be grounded.*
## 3. Finding the Positive Leak
There are two primary culprits that leak positive DC voltage onto an AVC bus: **a gassy tube (Grid Emission)** or a **leaky coupling capacitor**.
Here is how to isolate the exact component on your bench right now:
1. **Establish the Baseline:** Disconnect the signal generator. Turn the SX-28 on, set RF Gain to maximum, and connect your scope/DMM to the AVC line. Verify the $+1.4\text{ VDC}$ is present.
2. **Test for a Gassy Tube:** With the radio running, pull the tubes connected to the AVC line one at a time (1st RF, 2nd RF, Mixer, 1st IF, 2nd IF).
* If you pull a specific tube and the $+1.4 V instantly vanishes (dropping to $0 V or slightly negative), that tube is gassy. It is emitting electrons from the grid, creating a positive charge. Replace that tube. * Be sure to reinstall each tube after testing if the voltage does not drop.
3. **Isolate the AVC Rectifier:** If all the RF/IF tubes are pulled and the positive voltage is *still* there, pull the AVC Rectifier tube (on the early SX-28, this is usually the 6H6).
* If pulling the 6H6 clears the voltage, the tube itself may have an internal short.
4. **Check the Coupling Capacitor:**
If the voltage remains, or if it points to the AVC circuit, you likely have a leaky capacitor. The SX-28 uses an Amplified AVC system (a 6AB7 AVC amplifier feeding the 6H6 AVC rectifier). There is a coupling capacitor between the plate of the 6AB7 and the diode of the 6H6. If that capacitor has even a microscopic electrical leak, high-voltage plate DC will bleed directly onto the AVC line.
Page: /Avc_test_and_repair/Gemini_902.shtml
Last modified: Tuesday, 21 Jul 2026