[time-nuts] 1903 Railroad self-Winding / Self-setting Clock

brucekareen at aol.com brucekareen at aol.com
Sat Nov 1 13:11:48 EDT 2014

The Western Union clocks used in broadcasting up through the middle 70's were designed to be corrected through one-second current pulses over a standard 60 mA teletype loop.  The clocks were wired in series like the old series Christmas-tree bulbs.  

Internally, the clocks employed two 1-1/2 V dry cells to supply 3 VDC to the winding mechanism.  This was activated whenever the clock spring ran down and was independent of the setting (correcting) signal.  When winding, the clocks emitted a soft purr.

The setting mechanism was arranged so that the second-hand was seized and forced to the 00 position when the clock pulse arrived at xx:59.  It was released at xx:00 when the one-second pulse ended. I believe a clutch mechanism prevented this setting action from affecting the pendulum. 

As another writer mentioned, the clocks were ideally adjusted  to operate a fraction of a second slow in an hour's time; thus the second-hand was pulled forward a bit rather then retarded.  The lead pendulum was short and the mechanisms not very elegant; thus the clocks were not very accurate when operated without the setting signal.

I believe there was also a mechanical gate that allowed the clock to be reset only when the second hand was close to the 00 position.

My understanding was that master clocks were installed in Western Union offices in the major cities where time service was provided.  These master clocks actually generated the one-second pulses from xx:59 to xx:00 on each hour.  Once a day a correcting pulse was routed across the country from the Naval Observatory.  This was sent over telegraph circuits that were used for other purposes, too.  At the appropriate time a technician had to enable the path to the office master clock to allow it to be corrected.  This was apparently a manual operation.
In originating network broadcast programs, timing accuracy needs to be well within one-second as one second typically represents about three words of the broadcast opening.  At one point the Western Union technicians went on strike forcing management personnel to take over technical operations.  Soon after we began noticing the clock error steadily increase.  Calling Western Union did not seem to help as it is possible the management personnel were not familiar with how to route the daily timing pulse to correct their office master clock.  After a week or so, we were forced to construct a driver for the clocks based upon a crystal-controlled clock that was set from WWV.  When Western Union indicated they were dropping clock service, we elaborated on this to a precision time-code generator with provision for arming so as to introduce leap-seconds when needed and to change from standard to daylight time on schedule.

In broadcasting we used both digital and analog clocks.  The digital clocks were used for starting network feeds, but were not very acceptable to announcers reading copy or talking up to a switching point.  They strongly preferred the proportional indication of a clock hand as it is difficult to translate  digital indications into time left when trying to concentrate on other matters. 

Bruce, KG6OJI

More information about the time-nuts mailing list