Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: [Phys-L] Allan deviation



On 07/13/2015 07:22 AM, Bill Norwood wrote:

So, why are the rods of more than one metal? Fine "tuning?"

Short answer: Temperature compensation.

Longer answer....... Metals change length
with temperature. The CoTE (coefficient
of thermal expansion) is not huge, but if
you're not careful the effect big enough
to be noticeable in a precision clock,
when the length of the pendulum changes.

The simplest thing is to make the pendulum
rod out of an alloy with super-low CoTE.
However, back in the olden days, such alloys
were not available. Instead, they used a
gridiron. It's similar to the principle
of an achromatic compound lens: Put in
a small backwards component that opposes
the nonideality of the main component.

In the gridiron, there are
a) some members with a relatively small
CoTE. These are under tension. When
they expand, they let the bob go down,
in the obvious way.
b) Some members with a relatively large
CoTE. These are under compression. When
they expand, they *lift* the bob upwards.

The total length of (a) is longer than (b),
so the overall gridiron has a nonzero length,
which is the length of the pendulum. However
the *weighted* sum of the CoTEs is zero to
first order.

In the olden days, it was easier to find
two /different/ CoTEs than to find one that
was zero.

This is a nifty trick, but it stopped being
necessary many many decades ago, when good
low-CoTE alloys became available. In a
modern clock, such a thing is almost certainly
fake, just window dressing.

The hat doesn't make somebody a cowboy.
The cape doesn't make somebody Superman.
The gridiron doesn't make it a precision clock.