5F10 Tweed Harvard‑style · 1958–1961 · 10 W

The tweed line's fixed-bias oddball: a roughly 10-watt student amp built around a 6AT6, a 12AX7, a fixed-biased 6V6GT pair and a 5Y3GT rectifier. Where almost every other tweed Fender cathode-biases its output tubes, the Harvard runs a proper negative-bias supply — the same idea as the bigger Bassman and the later brownface amps, shrunk into a one-knob-tone practice combo. It also opens with a 6AT6, a twin-diode/high-mu triode borrowed from the radio-and-television parts bin (only the triode section is used), rather than the usual 12AY7/12AX7 front end. Produced through the late tweed years.

Circuit walkthrough (short form)

Three inputs (each a 68 kΩ stopper, grid grounded through the switched jacks) → 6AT6 first stage (100 kΩ plate load, 1.5 kΩ cathode with a 25 µF bypass) → 0.02 µF coupler → 1 MΩ volume, with a 1 MΩ tone control whose wiper sits on the volume's input and whose ends blend between a 0.005 µF cut to ground and a 500 pF path to the volume wiper → 12AX7 driver (100 kΩ plate, 1.5 kΩ unbypassed cathode) → 0.02 µF → 12AX7 cathodyne phase inverter (56 kΩ plate load, 1.5 kΩ + 56 kΩ cathode stack, 1 MΩ grid leak returned to the junction) → 0.1 µF couplers → 6V6GT pair, fixed-biased at −27 V through 220 kΩ grid leaks with 1.5 kΩ stoppers → output transformer → speaker. A 56 kΩ negative-feedback resistor runs from the speaker back to the driver's cathode.

Power: HT winding → 5Y3GT → +305 V plates (16 µF) → 470 Ω → +302 V screens → 22 kΩ → +250 V preamp (16 µF). The bias supply takes a tap on the HT winding through a 6.8 kΩ series resistor into a selenium rectifier; a 56 kΩ bleeder and one 25 µF can (printed 25 V) sit on the −27 V grid line.

Fixed bias in a tweed

Two details set the Harvard apart from its cathode-biased tweed siblings. First, the 6V6 grids are held at a fixed −27 V from the selenium bias supply rather than lifted by a shared cathode resistor, trading a little of the tweed "sag and bloom" for a firmer, louder response; at that bias the model idles each 6V6 near 20 mA, about 6 W. Second, the phase inverter is a cathodyne: the 12AX7's cathode runs through 1.5 kΩ to a junction and then 56 kΩ to ground, with the 1 MΩ grid leak returned to that junction so the stage self-biases — simulation puts the cathode pin at 47.7 V and the junction at 46.4 V, a clean −1.3 V grid-to-cathode. The plate (56 kΩ from the +250 V rail) sits at 199 V, balancing the two drive signals to the output pair. The lower output coupler is taken off the cathode pin itself — the +44.3 V node at the top of the 1.5 kΩ — not off the junction: both the schematic and the layout page draw it there.

The tone control is a blend, not a cut

Both F-EF pages draw the same network, lug for lug: the 0.02 µF coupler lands on the volume's input lug, and the 1 MΩ tone pot's wiper is strapped to that same lug. One end of the tone track bleeds to ground through 0.005 µF; the other reaches the volume's wiper, the driver grid, through 500 pF. Turning the knob slides between the cut path and the treble path. Until 2026-10-03 the sheet drew a 500 pF bright cap across the volume plus a separate 0.005 µF and rheostat to ground, and the board ran the coupler into one end of the tone track and grounded the other bare. The two drawings disagreed with each other as well as with the source, which is what the sheet-board gate had been reporting. It is all capacitor-coupled, so no DC operating point moved.

Verification — against the printed factory chart

The F-EF drawing prints a full voltage chart, and every node on it is gated against simulation. The +302 V and +250 V rails land within 2 %, and every tube pin is within the chart's own ±20 % convention. The worst of them is the driver cathode at 15.2 %: it is unbypassed and loaded at DC by the 56 kΩ feedback resistor returning from the speaker, and the anchor-fit 12AX7 model runs a little light there, so simulation reads 1.3 V against the printed 1.5 V. Every other pin lands within 8.1 %. Nothing is disputed and nothing is excluded. The 6AT6 uses a purpose-built, public-domain model fitted to its RCA datasheet (triode section only); the two diode units play no part in the amplifier and are left unmodeled.

The heater layer: what the sheet shows, and a question it answered

The board drawing here used to close its heater chain with a link across the 6AT6's two heater pins. Nothing on the F-EF sheet joins them, and the 6AT6 has two heater pins and no centre tap, so those two pins are the supply's two legs and a link between them is a short across it. The link is gone; a drawn pair already spans both of a socket's heater pins, so the chain's last hop never needed one.

What the layout sheet shows, read at its native 6378 × 4525 with the 12AX7's socket circle fitted to its drawn rim and the pin bearings measured: one green transformer lead meets the yellow-red HT centre tap at a solder point and goes to a ground hatch, and the other green runs to the pilot light's single terminal and on to the first 6V6GT at pin 7; each 6V6GT takes pin 2 to a ground hatch, with the unused pin 1 bowed to it as the tie point; the 6AT6's heater pin 4 carries a short lead to a ground symbol and pin 3 takes the feed; and at the 12AX7 a short bow outside the socket rim ties pins 4 and 5 together while pin 9 carries a lead up to a ground hatch — the 6.3 V parallel grouping, ends strapped as one leg with the centre tap the other. A single-ended supply, every socket returning to chassis on its own, and the board declares it as such and is drawn as single green leads.

Two things about the 12AX7 were open until the socket was measured rather than looked at. A stroke that seemed to join pins 8 and 9 at 9× is the socket rim: with the circle fitted (coverage 1.00), the ink between those pins has the same inner and outer edge as every other arc of the same rim, and nothing lies outside it. And a straight line leaves the pins-4/5 tie and runs right past the cathode at pin 3 — either landing on that terminal or grazing it on its way to the 6AT6. The scan cannot separate a landing from a crossing there, and the circuit's own argument (a heater leg tied to a cathode with the other leg grounded would short the supply) can reject a reading but not establish one. It does not need to: which pins the supply's two legs sit on is settled by the bow and the grounded centre tap, and that is all the heater declaration states. The collision stays recorded here as what it is.