Fender's brown-Tolex remake of the Super, the two-10-inch step above the
Deluxe. Production ran 1960–1963 on the 6G4 chassis documented here, a
40-watt combo running a fixed-biased 6L6GC pair off a GZ34 rectifier,
before the 6G4-A revision swapped in a 5881 output pair and before the name
passed to the blackface Super Reverb. Where the tweed 5F4 it
replaces cathode-biased its 6L6G pair through a selenium-assisted bias
network and a cathode-follower-fed tone stack, the 6G4 moves to a proper
fixed-bias supply, a long-tailed-pair inverter, and adds vibrato — the same
recipe the brown 6G3 Deluxe carries at smaller scale. The
preamp/PI bottles are marked 7025 on the drawing, the low-noise selected
version of the 12AX7.
Two mirrored channels
Each channel runs the same recipe, wired independently rather than sharing a
tone stack: two inputs (68 kΩ stoppers merged into a 1 MΩ leak, no coupling
cap — the jack-merge node is the first stage's grid directly) → a 7025 input
stage (100 kΩ plate, 1.5 kΩ cathode) off the shared +170 V input-stage rail →
a Bass/Treble two-knob tone ladder (250 pF treble cap, 100 kΩ slope, a 10 kΩ
bleed) → a Volume pot → a second 7025 recovery stage (100 kΩ plate, 820 Ω
cathode) → a mixing resistor into the phase inverter. The two channels'
recovery stages sit on different rails, though — channel 1's plate prints
+160 V and channel 2's +120 V — so this circuit, unlike its siblings, is not
simulated as perfectly symmetric between channels; each recovery rail is
modelled at its own printed value.
The drawing does not print channel names (unlike the following 6G4-A
revision, whose otherwise near-identical drawing labels its two channels
"VIBRATO" and "NORMAL"); this entry follows the 6G4's own drawing and does
not assert a name for either channel.
Phase inverter and output
A 7025 long-tailed pair (82 kΩ hot / 100 kΩ cold 5% plate loads off a
supply derived through the drawing's own 10 kΩ dropper from the 6L6GC plate
rail, an 820 Ω shared cathode resistor to a 6.8 kΩ tail, both 1 MΩ grid
leaks returned to the tail junction) drives the 6L6GC pair, fixed-biased
at −55 V through 220 kΩ · 5% leaks, with 470 Ω · 1 W screen resistors.
Presence is a 5 kΩ-linear pot and 1.5 kΩ foot resistor carrying a 56 kΩ
negative-feedback return from the speaker line — the identical recipe 6G3
uses, DC-neutral at the tail.
Tremolo
A phase-shift oscillator (Speed on a 4 MΩ reverse-audio pot feeding a
0.01 µF/0.01 µF ladder into a 7025 grid, with a 4.7 MΩ/0.005 µF feedback
path) drives an Intensity control (10 MΩ reverse-audio) that injects
into the same −55 V fixed-bias line the 6L6GC grids share — bias-vary
tremolo, the same mechanism 6G3 uses, rather than the optocoupler-shunt
tremolo AB763 carries. The Intensity feed carries no DC (the output grids
draw no grid current), so it moves no operating point.
Power
The GZ34 delivers +456 V at the 6L6GC plates and screens (470 Ω · 1 W
stoppers). A 10 kΩ dropper feeds the phase-inverter supply, landing at the
printed +315 V (hot plate) / +310 V (cold plate) through the 82k/100k plate
loads. A separate rectified/filtered bias tap (56 kΩ + 10 kΩ divider, 8 µF ·
150 V filter) delivers the fixed −55 V bias line.
Reading against the printed chart — and what the scan does not resolve
The drawing prints a full voltage chart directly on the schematic at every
stage (no separate tabulated table), read to ground with an electronic
voltmeter, values shown ± (the companion layout page states the drawing's
usual ±20%). Several of those printed values are gated in the operating-point table —
the phase-inverter plates (+315 V/+310 V, both within a few percent) and the
fixed-bias/screen nodes (exact by construction, driven or one resistor
away). The channel input and recovery rails are not gated: at this
scan's resolution, the drawing's own dropping-resistor chain from the main
6L6GC-plate rail into those three preamp rails was not legibly readable, so
each is instead calibrated — driven at whatever value reproduces the
printed plate reading through the drawing's own (legibly-read) 100 kΩ
plate loads and the tube's own self-bias. That reproduces the plate voltages
by construction, which is not a verification of them, so the operating-point
table reports those plate nodes as informational and gates the cathode nodes the
calibration was not tuned against instead. The joined phase-inverter
cathode/tail-junction nodes (KPI/JPI) are reported the same way: not legibly
read on this drawing, but landing close to 6G3's own verified chart at
the analogous node (which shares the identical 820 Ω/6.8 kΩ tail pair) —
+20.9 V simulated against 6G3's printed +20 V, +18.6 V against +18 V — an
independent cross-check this circuit's own chart could not supply directly.
The tremolo oscillator is excluded from the netlist entirely, the same
documented-exclusion pattern 6G3 and AB763 use: it is a running phase-shift
oscillator with no static DC operating point, and unlike those two circuits'
own charts, the 6G4 drawing prints no dynamic-average pin voltage for it at
all, so there is nothing to report even informationally.
Nine functions, eight bottles. The published layout page draws exactly five
noval sockets alongside the GZ34 and the two 6L6GCs, so the circuit's nine
preamp/driver/inverter/oscillator functions have to share. They share the way
the 6G3 shares: the tremolo oscillator is the second section of the same bottle
as a channel's recovery stage. Which of the two channels it pairs with is not
legible at this scan's resolution — channel 1 is an editorial choice here, not
a read fact, and it has no bearing on the DC model, which excludes the
oscillator entirely.
What stands between this circuit and verified. The drawings have landed:
the schematic passes its grammar gate, both board styles render with zero
collision-lint findings, and the drawn wiring is proved electrically equivalent
to the netlist. What remains is the un-gated preamp rails above — a
higher-resolution read of the three dropping resistors would turn three
calibrated rails into checked ones — and the maintainer's own review, which is
the only thing that grants verified here. A gated node's worst deviation
today (the phase-inverter cold plate, 3.2% against a printed value carrying a
±20% convention) is not what is holding it back.
No sharper copy of the 6G4 sheet is publicly archived. Every mirror of
the two-page schematic-and-layout scan that could be found — el34world,
Schematic Heaven, Prowess Amplifiers — serves the same single 2002 scan:
byte-identical files, or re-saves of the identical embedded page images with
no new information in them. Several archives carry only the following
production revision, the 6G4-A (a 2× 5881 output pair and a different preamp
complement — a different drawing, and the one this entry's own sources
distinguish it from). On both pages of that one scan, no continuous, legible
wire run carrying a resistor value connects the main B+ rail to the
+170 V/+160 V/+120 V preamp-rail takeoffs: the run itself does not resolve
at the scan's resolution, on the schematic or on the companion layout page,
and magnifying the embedded page images adds nothing a magnified copy of the
same pixels cannot. The three preamp rails (BD, BE1, BE2) therefore remain
calibrated rather than checked, and verified requires either a sharper
original scan than any archive currently mirrors, or the maintainer's own
inspection of a physical unit or a better print.
