Twenty-Five Turns for Every One the Speaker Needs

The crystal radio I’ve been hunting hot spots on uses a transformer to match the high-impedance detector to my earphones. Twelve days ago that transformer was just a component — a small black cube with wires coming out. Tonight, while rewinding a damaged output coil, I started thinking about what’s actually inside. Primary turns. Secondary turns. Core materials. Impedance ratios. I looked up transformer winding tutorials, and four hours later I’d watched someone wind their own output transformer for a single-ended 300B tube amplifier, explaining gap spacing and saturation curves while copper accumulated on a bobbin.

The kit arrives Thursday. But I’m not waiting to understand how the thing works.

The Output Transformer Problem

A vacuum tube is a high-impedance, low-current device. An EL84 power pentode wants to see a load of about 5,000 ohms at its anode. A loudspeaker is a low-impedance, high-current device — 4 or 8 ohms. You cannot connect these directly. The tube would see near-infinite impedance, the speaker would see near-zero drive, and nothing useful would happen.

The output transformer bridges this gap. It converts the high-voltage, low-current swing at the tube’s anode into the low-voltage, high-current swing the speaker needs. The transformation follows the square of the turns ratio:

Impedance ratio = (Primary turns / Secondary turns)²

For 5000Ω to 8Ω:
Ratio = 5000 / 8 = 625
Turns ratio = √625 = 25:1

So a transformer with 2,500 primary turns and 100 secondary turns would match an EL84 to an 8-ohm speaker. In practice, the transformer manufacturer calculates this — you buy one rated for your tube and speaker impedance. But understanding why it works matters when things go wrong.

The DC Problem in Single-Ended Designs

Push-pull amplifiers use two tubes that conduct alternately, each handling half the waveform. The DC currents through their halves of the output transformer’s primary winding flow in opposite directions and cancel out. The transformer only sees the AC signal component.

Single-ended amplifiers have no such luxury. A single output tube conducts continuously, and its full DC bias current — anywhere from 30mA for a small triode to 250mA for a 6C33C — flows through the transformer primary at all times. This creates a constant magnetic flux in the core.

Iron cores saturate. Push enough current through the winding and the core’s magnetic domains all align — it can’t hold any more flux, and inductance collapses. To prevent this, single-ended output transformers include an air gap: a thin slice of non-magnetic material (or literal air) in the core’s magnetic path. The gap increases reluctance, preventing saturation but reducing inductance.

Lower inductance means worse bass response. The transformer becomes a high-pass filter. To restore low-frequency performance, you need a physically larger core with more primary turns. This is why single-ended output transformers are disproportionately heavy and expensive — they’re fighting physics.

Why Class A Runs Hot

The EL84 kit I ordered is push-pull, Class AB. But I’ve been reading about single-ended Class A designs, and the efficiency numbers are startling.

In Class A, the output tube conducts continuously, biased at a fixed current. When there’s no signal, the tube still dissipates full power as heat — the quiescent current times the plate voltage. When a signal arrives, the current swings above and below the bias point, but the average power consumption doesn’t change. Maximum theoretical efficiency is 50% for a transformer-coupled Class A stage. In practice, it’s often closer to 15%.

A 5-watt Class A single-ended amp might draw 30 watts from the wall. The other 25 watts becomes heat — in the tube, in the transformer, in the chassis. This is why SET amplifiers glow. They’re space heaters that incidentally make music.

The Harmonic Argument

The preference for tube sound has a measurable basis. When a tube amplifier clips — when the signal exceeds what the circuit can reproduce cleanly — it generates distortion products. Single-ended stages produce predominantly even-order harmonics: second, fourth, sixth.

The second harmonic is one octave above the fundamental. If you’re playing an A at 440Hz, the second harmonic is 880Hz — also an A. The ear perceives this as fullness rather than harshness. It’s the same phenomenon that makes a violin sound richer than a sine wave.

Push-pull stages cancel even harmonics and produce predominantly odd-order distortion: third, fifth, seventh. The third harmonic of 440Hz is 1320Hz — an E, a note outside the octave. Odd harmonics sound dissonant. This is why solid-state clipping sounds aggressive and tube clipping sounds warm. It’s not mysticism. It’s Fourier analysis.

When I wound eight thousand turns of 42-gauge wire for that guitar pickup, I was making a transducer — converting string vibration into electrical signal. The output transformer is the reverse transducer at the end of the chain, converting amplified signal back into mechanical motion in the speaker cone. Both involve copper wound around magnetic cores. Both reward attention to the same variables: turns, spacing, core material, saturation limits.

The kit arrives Thursday. I’ve already ordered a variac for slow power-up — the tubes need to form, and hitting them with full voltage on first power risks damaging the cathodes. The turret board is pre-drilled but unwired. Point-to-point construction: each component soldered by hand, routed by me, positioned with attention to heat and voltage gradients.

There will be 300 volts on that board. I’ve worked high voltage before, but not often, and not lately. The kit includes a bleeder resistor to drain the filter capacitors after power-down. I’ll be checking it with a meter before I touch anything.