Re: Fused quartz pendulum - temperature coef

Harvey Moseley

Hi All,

Just a brief comment on temperature compensation of the pendulum. Since we know accurately the coefficients of expansion of the suspension, pendulum rod, and pendulum bob, you can determine accurately where along the length of the bob to provide first order compensation for the expansion or contraction of the materials in the pendulum system.  If we provide accurate first order compensation, the variation of the pendulum should be parabolic in temperature around the design temp.  The breadth of the parabola should be proportional to the second order changes, which for most materials, are proportional to their coefficient of expansion.  Bottom line, we can design the system to have no first order changes in period at the reference temp.  The breath of the residual quadratic variation will be determined by the coefficients of the suspension and bob.  Keep them all small, and the residual will be small.   This ignores any thermal effects from the air, which I think is of higher order, and thus of lower importance.  Hope this is helpful to those thinking about these system.

Best Regards,

On Sat, Apr 11, 2020 at 10:25 AM James Meaton via <> wrote:
Hi Neil and others
From what I remember from about 50 years ago the bob weight was supported by a brass knurled, threaded nut at the bottom of the Invar pendulum.  Although this facility obviously provided some adjustment for timing, customers could also order a set of precision weights that could be placed on to of the bob weight for more precise adjustments.  From what I remember the majority of the master clocks were hidden away in cellars, lofts and old cupboards; in those days the main priority of a Synchronome system was that all the clocks in the building(s) were at the same time, rather than being spot on for GMT.  I always regret getting rid of the master clock case I had for many years, but to be honest this was well before the times when they were considered of monitory or historical value.
Some of you may be aware of the Watchman’s Tell Tale System that could be linked to a master clock system. It consisted of a 6” dia aluminium tube that rotated once over twenty four hours.  A paper chart was fixed to the drum and up to 24 PO type 3000 relays with a “pricking” angled pin were fixed to the base of the machine.  As the night watchmen (women) carried out their patrols they would operate key-switches at important locations indicating by the pin-prick in the paper chart that they had made this check and at a given time.  (I converted a couple of short lengths of reject tubing into ceiling lights and bongo shells?!) 

Amongst my Synchronome bits and pieces I have a piece of ancient equipment that was fitted in one of the old battery charging cases.  I am afraid it’s horrendous condition is due to a leaking roof from many years ago.
I thought it was some sort of device for comparing mains 50 cycle synchronous time to that outputted by a master clock, but looking at the connections I am not so sure. 
If anyone is interested in this old piece of history please contact me. Please see attachments below.

James Meaton

On 11 Apr 2020, at 14:40, Bepi <pepicima@...> wrote:

Neil, yes Matthys says in his book (pag 3) that it's better ( and easier) to hang a bob from the bottom edge.  Where di you get the description of his experimets though?
What I was saying is that while designing the new bob for my pendulum I realized that the Synchronome original one seems to have been designed for exact compensation, within my ability of determining the gray cast iron and the invar TC.

Two reasons for pointing this out: historical on the Synchronome genesis, it would be nice to know if my guess is right, James Meaton who worked for them might know.

Second, to figure out if it's worth doing so in general.  The only way to understand it would be to measure the effects of putting efforts at this kind of temperature compensation, it's not obvious to me that exact temp comp of the rod and bob thermal expansion leads to a proportional reduction in temperature sensitivity of the period at the level of precision we are interested in. A couple of months ago while I was modifying the bob I exploited the occasion to install 15W of LED strip lights inside the case which, with the inside 4 thermometers, should provide a reasonable chance of controlling the clock temperature boundary conditions. My intuition tells me that the ability to modulate the temperature would be more useful, and easier to implement, than to heavily insulate the case both for sensitivity and to separate time constants. No time to use the system yet though.


Join to automatically receive all group messages.