The following has been reformatted using HTML
Do not panic about the 501K reading just yet! That might actually be perfectly healthy.
Here is why: The AVC bus in the SX-28 *must* have a high-resistance DC return path to ground so the negative voltage doesn't instantly drain away. Depending on the exact schematic revision of your early SX-28, there is almost certainly a 500K resistor (often R29, R31, or R35 on the factory schematics) intentionally tied between the AVC line and ground. You are likely just measuring that exact resistor, and it is right on spec.
However, your finding that the **+2.0V remains with V9 removed** is the smoking gun. You have isolated the fault to the physical wiring, switches, or transformers.
Here is the math of exactly what you are hunting for, and why your plan to unsolder that node is brilliant.
### The "Micro-Leak"
You have a 501K resistance to ground, and a +2.0V charge on that line. Ohm's Law ($I = V/R$) tells us that $2.0\text{V} / 501,000\text{ ohms} = \mathbf{4\text{ microamps}}$.
You are hunting for an incredibly tiny, 4-microamp B+ leak. If the high-voltage B+ rail in this radio is around 250V, that means the physical "short" causing this leak is roughly **60 Megohms**.
A 60-Megohm leak is so small that it won't show up as a dead short on a continuity tester. It is usually caused by one of three things:
1. **Bandswitch Wafers:** The AVC line and the high-voltage B+ lines run right next to each other on the bandswitch wafers. Decades of dirt, flux, or a microscopic carbon track can easily leak 4 microamps across the phenolic wafer. 2. **Inside an IF or RF Coil:** The primary winding (which carries B+) and the secondary winding (which carries the AVC) are physically close. Moisture in the coil form, or a slightly conductive mica insulator on the trimmer base, can bleed B+ to the AVC line. 3. **The Broken Terminal Strip:** Phenolic terminal strips can become conductive over time, especially if they fracture and absorb humidity or if dirt bridges two adjacent lugs (one carrying B+, one carrying AVC).
### The Unsoldering Plan (The Perfect Next Step)
Your instinct to unsolder the connections at that floating AVC node is exactly the right move. That node is the main intersection where the AVC splits off to the IF strip, the RF deck, and the bandswitches.
Here is how to execute it surgically:
1. **Power off.** 2. Unsolder the wires from that floating node, separating them so they don't touch each other or the chassis. 3. **Power on.** 4. Take your DMM (set to DC Volts) and clip the black lead to the chassis. 5. Probe each of those separated wires one by one.
Most of them will read $0\text{V}$. **One of those wires is going to have the +2.0V (or higher) sitting on it.**
Once you find the wire carrying the positive voltage, you instantly know which branch of the radio (the IF strip, the RF switch, etc.) is hiding the 60-Megohm leak. Trace that specific wire, and you will find your culprit!
Page: /Avc_test_and_repair/Gemini_905.shtml
Last modified: Monday, 20 Jul 2026