2024-09-09

Last modified: 2024-09-10 15:17:17

Balance wheel

I assembled this and tested it hanging from the ceiling in the garage:

And here's the plot of tick rate over time for this version:

Curiously, this appears to be much worse than the last version from 2024-08-26 that was just a pair of mole grips clamped to the bottom of the wire?

That one was:

So why could this be worse? Surely today's version is better balanced, it certainly looks like it wobbles around less. Maybe at first it was twisted up too far, sort of past the elastic limit of the wire? Maybe because it is bigger the air resistance is more significant?

Or maybe because the period is so much higher, a smaller percentage variation is easier to pick out of the data? 10.45 seconds down to 10.20 seconds is a change of 0.25 seconds, or about 1/40 of the duration. 1/40 of the duration last time would be about 0.05 seconds, which is small enough to almost fit within the error bounds of the measurement. That's less than 2 frames of video.

The period here is 10 seconds, which I think is longer than I've achieved with any mechanism before.

It is possible that it would settle down if left for longer, but I can't be bothered filming it for longer.

Reasonable next steps include:

• hook it up to an escapement and see if it works at all
• hook it up to a light gate to measure the period more accurately and for longer
• acquire a flywheel

Since I can check whether or not it will work for free, I should hold off on buying a flywheel. And since I'll need some sort of toggling finger whether I'm connecting it to a light gate or an escapement, I should make a toggling finger for it, and some way of suspending the wire near some other apparatus for measuring it, either with an escapement or a light gate.