Dating & notes
Every valve complement above that is claimed at all is read from an Ampeg title block, because the Portaflex is one of the lines where the factory wrote the complement into the drawing rather than leaving it to the tube chart. That makes one widely repeated claim easy to check and easy to correct: no Portaflex drawing in Ampeg's own published schematic library specifies 7591 output tubes. The B-15N, B-15NC and B-15NF sheets all print 6L6GC, and the B-15S prints 7027A. The 7591 belongs to Ampeg's guitar amplifiers of the same years, not to the Portaflex. Two gaps are left open rather than filled. The first B-15 of 1960 and the solid-state-rectifier B-15NB of 1962 have no factory drawing read here, so they carry years and wattage but no valve complement. And the B-15N sheet that does exist is revision C of January 1974 — it documents the circuit as Ampeg last issued it, more than a decade after the model was introduced, so it dates the circuit's final form rather than its first. One circuit from this family is now drawn in the corpus — the B-15N, from the revision-C sheet above — and the reading that unlocked it is worth recording so the rest of the line does not re-walk it. A first attempt stopped at three connections it could not read: where the R28 470 kΩ / R29 510 kΩ midpoint goes, which supply dropper feeds the channel anode loads, and a driver-cathode node whose printed voltages refuse to balance. Re-read at 400 dpi, zoomed to the pixel, the first two are a drafting convention rather than a circuit question. The third stands, and is published as a disputed node. The sheet runs a long horizontal ground bus across the middle of the drawing, and several vertical lines cross it. At working magnification each crossing looks like a junction; at pixel level each is a **hop-over** — the little S-jog the draughtsman used for "crosses, does not connect". The same jog turns up wherever this draughtsman crosses a line: at the channel mixing resistors' crossings of their own tone-board input lines, and at every crossing the driver rail makes on its way down to the supply. Read that way the drawing resolves into three series chains, each spanning the two 6SL7 driver anodes (the sheet prints 235 V on one and 225 V on the other) with a node in the middle:
* R26 120 kΩ and R27 120 kΩ — the driver's two ANODE LOADS, sharing one supply node between them. That node runs down to the foot of R38, a 22 kΩ · 1 W dropper off the 6L6GC screen node; its twin R37 leaves the same screen node for the preamp rail both channels share, and the two never rejoin. * R28 470 kΩ and R29 510 kΩ — a divider across the same two plates, whose midpoint is coupled by C11 0.022 µF to the second grid. That divider is the phase inverter: a self-balancing (floating) paraphase. * R30 270 kΩ and R31 270 kΩ — meeting at the node the sheet labels −50 V. They are the two output-tube GRID RETURNS, which is what the title-block reading above already said they should be: R30 reaches the 6L6GC grid line beyond C13, R31 the other beyond C14, since the two 0.022 µF capacitors break each driver plate from the grid line it drives.
Two things remain open on the B-15N and are published as such on its own page. The driver cathodes contradict themselves on the sheet's own annotations — the printed 3 V and 2.2 V, taken with the 220 Ω and 1 kΩ drawn under them, imply currents no 6SL7 section passes — so both are disputed nodes carrying their arithmetic. And one crossing in the negative-bias supply is not settled, so the −50 V line is driven at the value the sheet prints rather than derived. Everything else solves: eight preamp nodes and the screen node land inside the era's ±20 %. No new tube model was needed for it. The corpus already carries clean-room fits for all three bottles — the 6SL7 anchored on the RCA 6SL7GT Class A1 point (Va=250 V, Vg=−2 V → Ia=2.3 mA, gm=1600 µmho, µ=70), the 6L6GC on the RCA 6L6-GC sheet, and the 5AR4 under its other name, the GZ34. Dates are as tight as their evidence allows and no tighter. Introduction years come from the published chronology; the closing years of the B-15NF and B-15S rows are bracketed by the drawings read here — the B-15S sheet is dated March 1971, which puts a ceiling on the NF — rather than pinned by a dated source of their own. One row — the B-15N — links a drawn circuit; every other row is history-tier.