Hanging Pendulums

A device called an "escarpment", which was used with both a pendulum and a balance wheel clock, was the basis for accurate time keeping until the 1930's when quartz came into use.

The most accurate pendulum clock, the Shortt-Synchronome free pendulum clock, was the most accurate pendulum clocks ever commercially produced. In 1984 a Shortt clock preserved at the US Naval Observatory was found to be 200 microseconds per day, equivalent to an error rate of one second in 12 years.

You are thinking of an escapement. An escarpment is what you toss your old mechanical clocks over, if you no longer want to keep them.

And yes, I know of the Shortt clock -- I'm interested in a comparison of the virtues of flexures vs. knife edges for suspending pendulums. (It looks like the Shortt clock used a flexure, which ought to tell me something).

That's them danged computer things. They'll do it to you every time.:-)

Well the Shortt-Synchronome had the pendulum suspended in a vacuum chamber, but I did come across this description of pendulums

" Instead of hanging from a pivot, clock pendulums are usually supported by a short straight spring of flexible metal ribbon. This avoids the friction and 'play' caused by a pivot, and the slight bending force of the spring merely adds to the pendulum's restoring force."

See

which discusses pendulums in considerable detail.

Why not a magnetic bearing in a vacuum - four pairs of concentric rings two pairs polarized radially for the pivot two pairs polarized axially to constrain the pivot add optical detectors and electromagnetic impulse, no contact at all. remaining loss depends on vacuum and possibly hysteresis loss in the pendulum structure.

Tim Wescott fired this volley in news:Ba- snipped-for-privacy@giganews.com:

Tim, MANY clocks use flexures, but the most-accurate ones utilize the knife- edge bearing.

It's simple... The first issue always brought up is a "red herring":

1) although the amount is small, a flexure slightly alters the length of the pendulum over its range of center-to-outside. This slightly shortens the swing time over that of a simple bearing. But... That ALONE could be compensated-for with a longer pendulum. Then, the real reason; 2) The flexing element fatiques with time. It changes its 'spring force', thus altering the period over time.

It's not much. Ordinary humans would never notice.

LLoyd

Easy for you to say, Tarzan.

I think the answer may depend on whether you're asking for the absolutely best pivot or are also considering the real-world survivability of your device. The best knife edge pivots use materials with a high modulus of elasticity and have very sharp edges. Those materials are typically brittle -- sapphire, diamond, tungsten carbide, etc. The drawbacks of a knife edge made of these materials is obvious.

Regarding the material for you flexure, if I understood, you were suggesting a piece of flex circuit cable (or whatever you call that stuff.) I imagine that spring steel has a much lower elastic hysteresis than the flex, and would make a flexure with less loss, and be more robust than a knife edge.

Man, you're really getting into it there. Yes, mechanical hysteresis of different materials is an issue. And confounding it even further is viscoelastic hysteresis. In steel and most other structural metals, at low rates of strain and at room temperature, the viscoelasticity is vanishingly small. But it's enough, in very critical applications, to slightly delay recovery from elastic strain. It can extend the cyclic period of a pendelum -- although finding the means to measure it might be a challenge.

I can imagine measuring hysteresis by means of a pendulum .

Googling - pendulum hysteresis - shows that to be far from an original idea. Lots of studies of effects more obscure than elastic or viscoelastic hysteresis.

True, but the overall design of the thing puts the coil that both actuates the pendulum and detects its velocity on the pendulum itself -- so it needs wires anyway. The flex cable was just a handy way to make a flexure and do the wiring at the same time.

Before I decided to use the ball bearing I had been thinking of a pair of conductive flexures -- one could do that if one were sufficiently motivated, I suppose.