Three Blanks Shattered and the Drivers Still Waiting

It’s 2:47am and I’ve just swept up the third walnut blank of the evening.

The first one blew apart at the tailstock — checked through the end grain somewhere I couldn’t see, held together by surface tension until the gouge caught and the whole thing delaminated across the lathe bed. The second I actually hollowed successfully, a nice 50mm cup about 25mm deep, before I realized I’d cut the chamber too wide and the 40mm driver would rattle around in there like a marble in a mixing bowl. Volume too large. Bass would be nonexistent.

The third one is in pieces on the floor because I pushed my luck on a catch at 11pm when I should have walked away at 10pm.

Here’s the problem I didn’t anticipate: headphone enclosures are face-grain bowls. When I was turning pen blanks back in March, the brass tube ran with the grain, spindle-style. The lathe work was concentric to the wood’s natural axis. Easy cuts. Predictable behaviour.

A headphone cup needs to be hollowed across the grain, like turning a salad bowl from a plank. The gouge hits end-grain twice per revolution — twelve o’clock and six o’clock — and face-grain at three and nine. The cutting forces change constantly. Push too hard at the wrong rotation and the tool digs in, the wood grabs, and three hours of careful shaping becomes shrapnel.

I thought the phonograph horn experience would help. It did not. The horn hollowing used gouges and hand tools, reading the grain, taking patient cuts. No rotation. No angular velocity multiplying every mistake. The lathe is a different animal.

The acoustic problem compounds the mechanical one. I’m trying to hit a specific internal volume for Helmholtz resonance — the same physics I understood perfectly well when calculating horn flare rates last week. A 40mm driver with a Vas around 20cc wants an enclosure between 15 and 30 cubic centimetres. Too small and the bass peaks too high, boomy and one-note. Too large and you lose bottom end entirely because the front and rear waves cancel at low frequencies.

For a hemispherical cup, the volume formula is straightforward:

V = (2/3)πr³

Target: 25cc
Solve for r: r = ∛(3V / 2π) = ∛(75 / 6.28) ≈ 2.3cm
Diameter: ~46mm interior

Forty-six millimetres internal, plus wall thickness of 6-8mm for structural integrity, means I need to start with a blank at least 62mm across. And I need to hollow that 46mm cavity to a precise depth — not by eye, not by feel, but measured — while the blank is spinning at 800 RPM and trying to throw my gouge across the shop.

I’ve been measuring with calipers between cuts. Depth gauge into the cavity, subtract from stock thickness, check the math. Repeat. The second blank measured perfectly — 24.8cc by my calculations — before I realized I’d forgotten to account for the driver baffle thickness. The driver doesn’t sit flush against the bottom of the cup. It mounts on a ring, recessed 4mm from the rim. That 4mm × π × 20² is another 5cc I hadn’t budgeted.

So even the successful cup was wrong.

There’s a bitter irony here. The tube amplifier was an exercise in impedance matching — transformer turns ratios converting high-voltage/low-current tube outputs to low-voltage/high-current speaker drive. Headphone enclosures are also impedance matching, acoustic instead of electrical. The driver is a high-pressure, low-displacement source; the ear canal wants low pressure and larger displacement. The enclosure geometry mediates that transformation.

I understand the theory. I can calculate the volumes. What I can’t do, apparently, is turn a bowl without destroying it.

The lathe is off now. The shop smells like walnut dust and failure. There are three 50mm planar magnetic drivers sitting in their anti-static bags, waiting for enclosures that don’t exist yet. They cost more than I want to admit. I keep looking at them and thinking about the commercial wooden headphones I could have bought instead — ZMF makes beautiful things, hand-turned, properly tuned, available for immediate purchase — and wondering why I thought I could do this myself.

The pen-turning logic doesn’t apply. Pens are spindle work: small, simple, forgiving. You can sand through a mistake. Headphone cups are faceplate work, bowl territory, and I’ve never turned a bowl. Should have started with salad bowls. Should have practiced on scrap. Should have done any number of things that weren’t “order expensive drivers and assume the woodworking would just… work.”

It’s 3:12am now. I’ve written more than I’ve turned. The chunks of the third blank are still on the floor because bending over sounds unpleasant.

Thiele and Small, the Australian acousticians who formalized speaker enclosure design in the 1970s, worked with mathematics and measurements. They didn’t have to hollow anything with a gouge. Their parameters are elegant: Vas, Qts, Fs, the mechanical and electrical Q factors that predict how a driver will behave in a given box volume. I can plug numbers into their equations all day. The equations don’t care that I can’t execute the physical reality they describe.

I’ll try again tomorrow. Softer wood, maybe — basswood, something that won’t fight back as hard. Shallower cuts. Slower speed. The drivers will wait. They’re good at waiting.

The lathe will also wait, but more judgmentally.