5F8-A High‑power Tweed Twin‑style · 1958–1960 · 80 W

The last tweed Twin, and the circuit that sets the shape of every Twin after it: the 5F6-A's front end and long-tailed-pair inverter driving four 5881s in two parallel pairs instead of two, on a supply stiff enough to hold them. Two channels of two jacks feed a 12AY7; a 12AX7 stage direct-coupled to a cathode follower drives the treble/middle/bass stack; a second 12AX7 splits the phase; and the output quartet runs fixed-bias at −41 V with 470 Ω 1 W screen resistors and a 1.5 kΩ stopper bridging each pair's grids. A GZ34 rectifies, a choke follows the reservoir, and a standby switch stands ahead of both.

Circuit walkthrough (short form)

Bright and normal channels, two jacks each (68 kΩ stoppers, 1 MΩ leaks) → V1 12AY7 (100 kΩ plates, shared 820 Ω cathode with 250 µF bypass) → 0.02 µF couplers → 1 MΩ volume pots (100 pF bright cap) → 270 kΩ mixers → V2A 12AX7 (100 kΩ plate, 820 Ω cathode, unbypassed) → V2B cathode follower, DC-coupled (100 kΩ cathode load) → TMB tone stack (56 kΩ slope, 250 pF treble, two 0.02 µF; 250 kΩ/1 MΩ/25 kΩ pots) → 0.02 µF → long-tailed-pair PI: 82 kΩ and 100 kΩ 5% plates, 470 Ω + 10 kΩ tail, both 1 MΩ grid leaks returned to the tail junction, 47 pF across the plates, 0.1 µF holding the second grid at the tail foot → two 0.1 µF couplers → four 5881s in two parallel pairs, fixed-biased at −41 V, into the 45268 output transformer and two speakers, with 56 kΩ of negative feedback returning to the tail foot alongside the 5 kΩ presence pot and its 0.1 µF wiper cap.

Power: 300-0-300 (PT 7993) → GZ34 → standby → +397 V plates (two 20 µF cans) → choke (14684) → +395 V screens → 4.7 kΩ 1 W → +355 V phase inverter (20 µF) → 10 kΩ 1 W → +295 V preamp (8 µF). Bias supply: selenium rectifier, 15 kΩ/56 kΩ, two 8 µF/150 V cans → −41 V.

The output stage, as the drawings wire it

The four bottles are two parallel pairs, not two push-pull pairs, and the distinction is the whole circuit:

• Each phase of the inverter drives one pair through one 0.1 µF coupler and one 220 kΩ grid leak. The leak hangs on the coupler node, which is the first bottle's grid.
• The 1.5 kΩ stopper stands between the pair's two grids, not in series with the coupler. On the factory board it is a single resistor bridging one socket's grid pin to the next socket's — which is exactly what the schematic draws, and what a builder wiring from the layout sheet will find.
• Both bottles of a pair share a plate connection to one end of the output transformer's primary, and both cathodes go straight to ground. Only the screens are individual: each has its own 470 Ω 1 W resistor to the +395 V node, and all four are socket-mounted rather than on the eyelet board.

The tone network, as the drawings wire it

The published 5F8-A schematic and its factory layout sheet draw the same ladder the 5F6 and 5F6-A print — not the textbook redrawing of these parts — and the schematic and layout here follow the sheets:

• The 250 pF treble capacitor and the 56 kΩ slope resistor both leave the cathode-follower output.
• One 0.02 µF capacitor runs from the slope resistor's foot to the node shared by the treble pot's lower lug and the bass pot.
• The bass pot is a rheostat in series down the ladder, its wiper strapped into the treble-lug node.
• The other 0.02 µF runs from the slope foot to the middle pot's wiper, so the Middle control slides the capacitor's injection point along a 25 kΩ leg that never leaves the circuit.
• The stack's output is the treble pot's wiper alone, and the middle pot's foot runs straight to ground; the presence control sits at the phase-inverter tail instead.

What this page does not claim

• The secondary impedance. The drawing prints the output transformer's part number (45268) and nothing about its windings, so no impedance is stated here and the amp carries no iron rating. Two speakers, no figure.
• The power rating and the production years. The sheet prints neither. The 80 W figure and the 1958–1960 span come from a separate published source, cited below, rather than from the drawing.
• The bias supply's own voltages. Only the −41 V line it delivers is printed, so the simulation drives that node as an ideal source rather than modelling the selenium rectifier and its 15 kΩ/56 kΩ divider.
• The 100 pF bright cap. It mounts on the bright volume pot's own lugs, inside the volume network the DC model abstracts, so it is annotated on the schematic rather than drawn as a two-terminal part.

Where the model simplifies, and why

The DC deck drives +397 V at the reservoir and derives everything below it. One simplification is worth stating plainly, because it is visible on the sheet:

• The choke is modelled as 130 Ω of winding resistance, an estimate. The sheet prints 397 V on one side of it and 395 V on the other — the same 2 V the 5F6-A's sheet prints across the same part number, 14684.

The tail foot, and a chart that does not fit it

The drawing runs the 10 kΩ tail's foot to the presence pot's hot lug and grounds the pot's cold lug, with the 56 kΩ feedback resistor and the second grid's 0.1 µF landing on the same foot: the schematic draws the four on one vertical, and the layout on one eyelet. The model carries that foot as drawn. The 56 kΩ (back to the speaker, DC ground through the output transformer) and the pot's 5 kΩ track are both DC returns, and simulation puts the foot at +12.5 V.

The printed chart does not fit it. Its plate figures (355 → 230 V through 82 kΩ, 355 → 225 V through 100 kΩ) imply 2.82 mA of tail current, and 2.82 mA through the 10 kΩ alone is 28.2 V — the printed +27 V junction, with +28.5 V at the cathodes. Through the drawn foot the same current would put the junction at 41.2 V. The two tail figures are therefore shown as disputed: the page prints them beside the simulated +39.8 V junction and +41.1 V cathodes rather than bending the circuit to match.

Verification — against the printed factory chart

The I-EG drawing prints a full voltage chart, and simulation matches twelve of its fourteen compared nodes (the 5881 screen carries no chart value and is informational only): rails within 0.4 %, every gated tube pin within 7.5 %, against the sheet's own stated convention of ±20 % read with an electronic voltmeter. The other two are the tail figures above, disputed.

Reading the chart off a scan

The I-EG sheet circulates in copies of very different quality, and its chart is lettered small. Every voltage on this page was read from a 693 ppi capture (6368 × 3218 for the schematic page); the widely mirrored copy carries the same sheet at 1506 × 841, roughly 100 ppi, and its chart digits do not resolve. Two figures come out wrong if you read them from the coarse copy anyway:

The cathode is the one to check against your own copy, because the circuit settles it without the drawing: it has to sit above the tail junction's +27 V, since its current reaches that junction through the 470 Ω. A figure below +27 V cannot be right, whatever the scan appears to say.

Component values and topology read cleanly off either copy. It is only the chart that needs the resolution.