A Latitude Within a Degree, and I Still Don't Trust It

The pendulum was still swinging when I came down for coffee, and that alone was new. On Day 1 it gave up inside an hour, a lump of lead losing heart in a stairwell. This time a coil the size of a shot glass sat on the floor beneath it, pulsing for a few milliseconds each pass, and the thing had held a steady twenty-centimetre arc all night without me laying a finger on it.

I should be pleased. Mostly I’m suspicious.

A thin pendulum wire threaded through a small brass Charron ring mounted just below the suspension point in a dim stairwell

The Charron ring did exactly what the old texts promised: a small brass ring a few centimetres below the wire, sized so the wire just grazes its inner edge at the far end of each swing. It shaves the sideways wander off every stroke and leaves the main swing alone, and last week’s ugly pirouetting ellipse ironed itself flat within minutes. What the texts are quieter about is that the ring has to be dead-centred on the rest position. Off by a millimetre and it rubs one side harder than the other — no longer flattening the ellipse but steering the plane, inventing a drift the planet had nothing to do with. I traded one lie for a subtler one.

Same worry with the drive. Tweezing seventeen jewels out of a pocket watch back in February taught me that an escapement’s real job isn’t adding energy — anything can add energy — it’s adding it at the single instant where it doesn’t disturb what you’re measuring. Fire the coil a hair before or after the bob crosses dead-centre and you hand it a tiny lateral shove, every swing, forever. Foucault himself only saw about 2.3 mm of drift per swing; a millimetre of off-centre nudge is not a rounding error against that. So the coil pulls symmetrically toward centre from both sides, triggered off an infrared beam-break at the bottom, and I spent longer trimming that symmetry than building everything else combined.

The number, then. I logged the plane angle every fifteen minutes for a day and fit a line. The one detail I’d have botched without checking: you divide by the sidereal day — 23 hours 56 minutes, not 24 — because the swing tracks the stars, not the sun, and those four minutes move the answer by more than you’d guess. Backed out, I land within a degree or so of where the map puts my stairwell.

I want to believe the planet drew that line. But I built a machine with two mechanisms, the ring and the coil, either of which can forge precisely the signal I went hunting for, and I can’t fully swear I didn’t just measure my own workmanship. The astrolabe never had this trouble — it borrows a real star, something I neither made nor can fake. This one I made end to end, which is the pride of it and the doubt of it in the same breath.

Foucault, apparently, got tired enough of exactly this that within a year he’d built a gyroscope instead. I’m beginning to understand the man.