stickmaker | JOHT: Measuring Time

The Joy of High Tech

Rodford Edmiston

Being the occasionally interesting ramblings of a major-league technophile.

Please note that while I am an engineer (BSCE) and do my research, I am not a professional in this field. Do not take anything here as gospel; check the facts I give. If you find a mistake, please let me know about it.

Time on Our Minds

Which came first, the day planner or the calendar? Silly question... or is it? An awareness of time led to planning events for the future, which led to a need for more accurate measurement of time, which led to more detailed planning, which led to... Well, you get the idea.

There is a long history of people gradually creating ways of more accurately and precisely measuring time to meet a perceived need. Once purely mechanical means of doing so - using devices independent of natural cycles - were available the need for long-term consistency - and uniformity - also became more and more important. (Those of you who have not read Dava Sobel's book Longitude on how John Harrison solved the longitude problem - or seen the miniseries based on it - should definitely do so soonest.)

Alexander Pope stated that men's judgements were like their watches, in that "none go just alike, yet each believes his own." Note that he wrote this just before the period when the first truly accurate mechanical timepieces were developed. Mass synchronization of accurate clocks and watches was still over a century away. The need for keeping clocks on the same time over wide areas was felt for centuries before doing so was actually feasible (at first by using difficult astronomical observations and calculations; later through telegraph lines). Many clocks in the early era of mechanical timekeepers only had an hour hand.

Even the synchronization between one's home clock and the time one carried in one's pocket could be troublesome, and drove invention. There are many examples of "father-son" arrangements, where placing a pocket watch in a "dock" (to use the modern term) on a matching mantle clock would set the watch to the time of the clock, and usually rewind it in the process. Of course, that only put those two together, and neighbors' clocks - or even those others in the same home - could be wildly different. Setting other clocks to a master standard was still problematic. Small timepieces were also notoriously inaccurate until well into the Eighteenth Century, and could only be trusted to carry accurate time for short intervals. This was due in part to the difficulty of making a precision piece of equipment so small, but also to the conditions a pocket watch was exposed to.

Portable clocks intended to provide reference to a standard were transported in special cases, to reduce environmental effects such as motion, and temperature and humidity changes. This helped, but even with such measures and great care in the transport, carrying accurate time from place to place was problematic. Which meant that people in the period of the first accurate timekeepers who wanted to set their clocks to an exact time did so with astronomical observations. This, of course, introduced the variations in skill of the observer. Long after Greenwich was accepted as the zero meridian, clocks elsewhere might give different times than those in that observatory.

I am privileged to have several issues of The Mentor, an early Twentieth Century publication of The Mentor Association. These unusual, small magazines had articles on technical, scientific, cultural and historical topics, with excellent photographs or illustrations. One of these issues is on then-modern high-tech timekeeping. By the time this January 1, 1917 issue was printed, the practice of sending standard time signals over wire had long been a common practice, and work on radio time signals was underway. Among other things discussed in this issue was radiotelegraphy time signals, which was then new, but gaining popularity.

These issues have loose cards in the front, each with a photo or art print on one side and an accompanying description on the other, plus the main pamphlet on the general topic. This particular issue I'm writing about has photo cards for some of the clocks at the US Naval Observatory, as well as other timepieces. The difference between Observatory clocks intended for different functions is interesting. The actual astronomical clock was in a sealed case, presumably with humidity, pressure and temperature controls. It rested on a solid pedestal, which was on a separate foundation vibrationally isolated from the rest of the building.

Something which struck me about these images and articles is that except for the use of electricity, the clocks shown and described would have posed little mystery for John Harrison, the inventor of the chronometer. The mechanisms were still driven by springs - though these were wound by electric motors - and they still used swinging pendulums to meter out intervals of time. They also acted through gears to rotate hands showing the time on clock faces. Even the device which sent the telegraphic time synchronization signal was clock-like, with a cam mechanically actuating an electrical contact. The other telegraph signals were generally held in abeyance while the time signal was transmitted.

For sending signals from the observatory to other clocks there was a pair of identical long-case clocks (one being a backup). When this article was written, time signals were only sent twice a day - by wire and radio - at Noon and 10 PM EST. (Mare Island had a similar station which sent the signals at the equivalent hours of Mountain Time.) In many cities, a time ball automatically dropped on receiving the Noon signal. This tradition harkened back to observatories dropping a ball to likewise mark the hour. This was especially useful in harbors, allowing ships' navigators to set their chronometers before leaving port.

The telegraph and radio signals sent by the Naval Observatory were rarely more than two tenths of a second off as received. The average error during this period was only five hundredths of a second. This was plenty good enough for just about any use conceivable. Then.

The most accurate clock in the world at that time was probably in an observatory in Berlin. It bragged an error of only fifteen thousandths of a second in a day. To attain this accuracy with mechanical clocks required starting from the ground up. Or, rather, with a solid foundation well below the ground, often rooted in a granite or concrete block, reaching perhaps to bedrock. The entire room for such a clock must be custom designed and built, with the support for the clock actually separate from it. That way, vibrations from people moving and working in the room would not be conducted to the clock. These rooms were often underground, not only because of the nearly constant temperature - very important in those days before commonplace air conditioning - but to help isolate it from other environmental disturbances. The electric winding motor might run as often as every minute, to maintain a constant driving force in the mainspring.

Well before this publication saw print, clocks were accurate enough to raise questions. Among them was: What time, specifically, should be measured?

Natural events by which time had been traditionally measured are not divided up into nice, even intervals. For example, there is not a whole number of days in a year. Neither is there an even number of days in a lunar month, nor lunar months in a year.

Astronomers use what is known as sidereal - or star - time. Because of the Earth's movement in its orbit around the Sun the time between when a particular star rises and sets on subsequent nights takes a little less than twenty-four hours. By using sidereal time astronomers greatly simplify observation schedules. However, to go from astronomical observations (used to set solar clocks) to civilian time requires a conversion factor. This is because civilian times are set to traditionally observed terrestrial events, such as local Noon for days and the Solstices for multi-day periods.

Many watches were fairly accurate during this period, at least enough for everyday use by most professions. With clock towers in every city and many towns, usually set by the Naval Observatory time signal, watches could be checked frequently during the day. Most good quality watches didn't need to be checked this often, if kept in good condition and wound correctly. There were few professions which actually required the dependability of good quality watches, but there were a few workplaces which needed accurate, reliable, consistent measurement of time. Among those were the railroads.

It's no coincidence that a traditional retirement reward for a railroad worker was a watch. Before radios were small enough and cheap enough to put on trains the workers' timepieces had to be right, and the trains had to run closely to schedule. Otherwise the fast passenger train might find out the hard way that the slow freight ahead of them hadn't reached the siding yet. Because the pocket watch of one engineer was running a bit slow or the other a bit fast.

Railroads were actually the primary driving force behind the US adopting strict time zones. This practice of standardization was necessary for the same reason that accurate timepieces on trains and in stations were. (As I was writing this, a PBS program on filmmakers who left Hitler's Germany and the countries it conquered showed the segment from High Noon where the station clock strikes Noon.) Local Noon depends on your longitude, but Noon for a time zone is set by telegraph and therefore is the same throughout the zone. Theoretically, everyone there should have the same time.

These days we require far more accuracy, precision and dependability in our timekeeping than even the best mechanical systems can provide. This requirement is met by using oscillators far more rapid than those of any pendulum or balance wheel. A quality quartz wristwatch is a far better timekeeper than any clock from a century ago, because the tiny crystal inside drives a timing circuit at a frequency of around 32 kHz. Larger - and more expensive - quartz chronometers may have timing frequencies in the multiple megahertz. We humans don't personally need time this finely divided and exquisitely rationed for ourselves, of course. The nerve impulses in our bodies travel at a relative snail's pace, not even reaching the speed of sound. However, many of our endeavors do need this close measure of time. Trains, planes and spacecraft - Especially spacecraft! - today need literally split-second timing.

Fortunately, the technology has advanced sufficiently that the handling of this is normally transparent. Our computers automatically synchronize through the Internet. For the continental US, many of our clocks and even watches detect the WWV shortwave signal from the Naval Observatory; worldwide there are many other such radio references, including one from the Greenwich observatory. Our mobile phones constantly check their electronic clocks against the time signal of the closest cell tower. Our technology of time is so advanced that satellites with clocks on board automatically compensate for the slight changes in the passage of time their high velocity and high altitude cause due to the effects of relativity. An effect only recently predicted by Albert Einstein at the time the issue of The Mentor mentioned above was published.

We may not have mastered the use of time, but today we humans are pretty good at measuring it.