An understated Jose detail: he typically removed the input grid resistor from the new V0 stage and substituted a 250pF ceramic cap soldered directly to the chassis at the tube socket. Per Dave Friedman (Headfirst Amps Panel #11, ~[48:57]): *"In most of the Joses he never had an input grid resistor. He eliminated the input grid resistor and right at the tube socket, soldered to the chassis was a 250 picofarad ceramic cap."*
What this does electrically: - The grid resistor (typically 68kΩ in stock Marshall practice) provides RF immunity and limits any leakage current into the grid. Removing it gives a slightly quieter noise floor — there's no resistor adding Johnson noise at the input. - The 250pF cap to chassis ground forms a high-frequency shunt right at the tube pin, taming RF interference without adding noise.
Tradeoffs: - Quieter noise floor in moderate-RF environments (recording studios, well-shielded venues). - More RF-prone in dirty-RF environments (radio stations nearby, fluorescent dimmers, switching power supplies). Without the grid resistor's series impedance, RF can couple into the grid more easily.
Why Jose did this: the input stage's noise floor is the most-audible source of hiss in a cascaded high-gain amp because everything downstream amplifies it. Trading slight RF susceptibility for a quieter floor was a deliberate stage-build choice. Friedman notes that Jose's amps from this era are generally "Wild West shit show" on grounding practice (see panel [50:51]) — Jose tied the input coax shield to chassis at both ends, causing ground loops — but the specific input-end cap-to-chassis trick is a deliberate, documented part of the recipe.
JMIL status: the lab does not surface this as a toggleable component (modeling the noise-floor difference would require a richer signal-path simulation than the surrogate audio engine provides). Documented here as historical context.