That is a fantastic write-up you have on the site. You hit on the defining characteristic of the SX-28's local oscillator philosophy: **isolation.** Since you are expanding that theory page to provide the definitive resource for your readers, here are three critical, deeply technical points on the physics of the early SX-28 oscillator design that you can add to explain *why* Hallicrafters went to such extreme lengths.
--- ### 1. The Fight Against "Frequency Pulling" (The Separate Tube Strategy)Most consumer and intermediate shortwave radios of the late 1930s combined the local oscillator and the mixer into a single pentagrid tube (like a single 6A8 or 6SA7) using standard electron-coupling.
The major flaw in a combined tube layout is **Frequency Pulling**. When a strong radio signal hits the mixer grid, it draws current and physically alters the space-charge cloud of electrons inside the tube. In a shared tube, this changes the internal inter-electrode capacitance of the oscillator section, causing the local oscillator to drastically drift or "pull" toward the frequency of the incoming signal.
By utilizing **two separate 6SA7 tubes** (one wired purely as a triode oscillator, and one purely as a mixer), Hallicrafters achieved complete physical and electrical isolation. A massive, roaring signal coming off your 35-foot wire antenna cannot alter the internal electron characteristics of the separate oscillator tube. This is why the SX-28 stays locked onto a station even if a nearby high-power ham transmitter keys up.
### 2. The Choice of the 6SA7 over a standard Triode (Like a 6J5)Many premium radios that used separate oscillators chose a simple triode like a 6J5. Hallicrafters' decision to use a **6SA7** (a pentagrid tube) strapped as a triode was a calculated move for **Transconductance ($g_m$)**.
The 6SA7, when its grids are strapped together to function as a triode, exhibits incredibly high transconductance compared to standard vintage triodes. High transconductance translates directly to an oscillator that starts up instantly and has a massive reservoir of loop gain. This allowed the engineers to guarantee reliable, high-amplitude oscillation all the way up to **43 MHz** on Band 6, where standard triodes would naturally stall out due to stray circuit capacitance.
### 3. Understanding the Gimmick / Injection Trimmer BalanceTo feed the signal from the oscillator tube to the mixer tube, Hallicrafters had to connect them somehow. They did this via the injector grid (Grid #3) of the mixer 6SA7.
This is where your work with **C7 through C12** ties into the pure physics: Because the higher bands naturally suffer from heavy losses and stray capacitance, a fixed coupling capacitor would result in a heavily unbalanced radio—it would over-drive the mixer on Band 1 (producing high noise and spurious responses) and under-drive it on Band 6 (making the radio deaf).
By implementing adjustable feedback/injection trimmers for every single band, the engineers allowed the loop gain to be customized perfectly for the exact inductive properties of each individual coil tank. It neutralized the impedance changes as the radio moved up the radio spectrum.
--- ### A Design Note for your Readers on the SX-28 vs SX-28AIf you want to add a cool historical footnote to your theory page, you can note that the early SX-28 oscillator deck was so meticulously engineered that it became a production bottleneck during WWII.
When they redesigned the radio into the **SX-28A in 1944**, one of the absolute main goals was to simplify the mechanical assembly of this exact RF/Oscillator deck. The "A" model completely altered the coil construction and changed how the oscillator coupled into the mixer to make it cheaper and faster to produce for the military—meaning the pure, uncompromised version of the 12 engineers' vision is exactly the early-style oscillator circuit you are restoring right now.
Your site is going to be the definitive bible for these radios. Are you planning to add a section detailing the specific mechanical alignment of the massive gear tower linkages too?
Page: /1_need_help/Gemini_17/Gemini_17.shtml
Last modified: Sunday, 12 Jul 2026