This site is only a doorway

Everything here is gathered from things other people actually researched, thought through, and built: books, papers, historical records, and the people behind them. If something here catches your interest, go look at the original sources listed at the bottom of the page. They carry more depth than a page like this ever can.

How Does a Clock Keep Time So Steadily? From Shadows and Water to Springs, Crystals and Atoms

Category: Technology

A clock on the wall ticks away all day without anyone watching it. But a clock has no idea how long a second is. What it does is count something that repeats at an even pace. NIST, the US national measurement laboratory, says every clock needs two parts: a regular, repeating process that marks off equal bits of time, and a way to keep count and show the result. [1] The whole history of clocks is a search for a better thing to count, and the rest of this page follows that search.

What Did People Count Before Machines Existed?

In some early societies, people used the Sun's path across the sky to divide the day. Egyptian obelisks, tall four-sided stone monuments, were built as early as 3500 BCE, and their moving shadows split the day into parts, such as morning and afternoon. [1] A shadow clock is simple, but it only works when the Sun is out.

So people looked for a repeating process that did not need sunshine. Water clocks, which were in use in Egypt by around 1500 BCE and taken up by the Greeks around 325 BCE, let water drip out of a vessel while the level showed how much time had passed. [1] Like an hourglass, it uses the movement of a substance, though a water clock is read from the water level.

Why Were Water Clocks Hard to Make Accurate?

The trouble is that the flow of water is hard to control. NIST notes that a clock based on it could never become very accurate, and that later Greek and Roman designs added extra parts to make the flow more constant by regulating the pressure. [1] A simple way to picture it: a full bucket with a hole in the bottom squirts faster than a nearly empty one, so the "even pace" quietly changes as the bucket empties.

This lesson comes back again and again. A good clock needs something that repeats at the same speed whether it has just started or has been running for hours. The next big step was a way to build exactly that from metal.

Why Does a Mechanical Clock Not Just Unwind All at Once?

Large weight-driven mechanical clocks appeared in the towers of several Italian cities in the first half of the 14th century. [2] A falling weight (or a coiled spring, in later and smaller clocks) has plenty of power, but if it were allowed to move freely it would run down in moments. The clever part is a device called an escapement, a mechanism that lets the gears advance only one small step at a time, over and over.

The oldest design was called verge-and-foliot, and versions of it ruled for more than 300 years. [2] Its weakness was that its beat depended heavily on how hard the weight was pushing. [2] That ticking you hear from a mechanical clock or watch comes from this stop-and-go mechanism letting the gears move forward a notch at a time.

A big improvement came when a pendulum took over the job of setting the beat, which is the point of the next section.

How Did a Swinging Pendulum Improve Clocks?

A pendulum swings at a rate set mostly by its length, not by how hard it was pushed. Christiaan Huygens used this idea in his pendulum clock, and NIST says its error was less than one minute a day, the first time such accuracy had been reached. Later refinements cut the error to less than 10 seconds a day. [2]

Look at what changed. In a mechanical clock, one part supplies the power (the weight or spring), and a separate part decides the pace (the pendulum, or in a watch a small balance wheel with a hairspring). Splitting those two jobs is what made clocks steadier. In a pendulum clock, even as the driving force changes a little, the pendulum's rhythm stays relatively steady.

Why Did Watches Start Counting the Vibrations of a Crystal?

If steadier and faster repeating is better, then a vibrating crystal is a good candidate. NIST dates quartz crystal oscillators and clocks to the 1920s and onward. [2] On December 25, 1969, the Seiko Quartz Astron went on sale in Tokyo, a watch that was accurate to within five seconds per month. [3] Its quartz crystal vibrates 8,192 times each second, and the crystal oscillator replaced the hairspring, balance wheel and escapement of a mechanical watch. [3]

The idea is the same as before: count a repeating event. The difference is how fast and how evenly it repeats. A crystal does thousands of vibrations per second, so a tiny electronic circuit can count them and divide them down into a slower signal that drives the display. For more about why quartz is used this way, see the Kininarunote article on how common quartz became part of your watch.

What Does the Most Accurate Clock Count?

The best clocks today count the "vibrations" of atoms. NIST says quartz clocks have been substantially surpassed by atomic clocks. [2] In fact, the length of a second is now defined this way. The international definition says one second is the time taken by 9,192,631,770 cycles of the radiation from a transition between two energy levels of the undisturbed caesium-133 atom, in other words, the signal the atom gives off over and over (spelled "cesium" in the US). [4]

So the answer to the title is the same from sundials to atoms: a clock keeps time steadily because it counts something that repeats at a steady pace, and people kept switching to repeats that were steadier. Even the second itself now comes from such a count, not from the spinning Earth.

How Can You See This Counting Idea at Home?

You can build a rough clock of your own with a bit of string and a small weight, such as a few washers, tied securely to the end. Swing it gently in a clear space, away from people, and count how many swings fit into a minute you measure with a clock. Then try a shorter string. If you change the length, the count per minute should change too, and that is the pendulum's pace-setting at work. Keep the swings small, and do the experiment away from anything breakable.

Sources

  1. NIST, "A Walk Through Time - Early Clocks." https://www.nist.gov/pml/time-and-frequency-division/popular-links/walk-through-time/walk-through-time-early-clocks (Read via a summary tool, not as the full page. Supports the two basic parts of a clock, Egyptian obelisks from about 3500 BCE, water clocks, and the difficulty of controlling water flow. This page is an archive, accurate as of 2004. The pail-with-a-hole picture is our own illustration, not from NIST.)
  2. NIST, "A Walk Through Time" (mechanical and later clocks page). https://www.nist.gov/node/446006 (Read via a summary tool, not as the full page. Supports weight-driven tower clocks in the early 1300s, the verge-and-foliot escapement, its dependence on driving force, Huygens' pendulum clock error of under a minute a day, quartz clocks developed from the 1920s, and atomic clocks surpassing them. The statement that the pendulum's rate depends mainly on its length is general physics, not taken from this page.)
  3. IEEE Engineering and Technology History Wiki, "Milestones: Electronic Quartz Wristwatch, 1969." https://ethw.org/Milestones:Electronic_Quartz_Wristwatch,_1969 (Read via a summary tool. Supports the December 25, 1969 introduction, accuracy within five seconds per month, the 8,192 cycles per second, and the replacement of the hairspring, balance wheel and escapement.)
  4. BIPM, "Second." https://www.bipm.org/en/si-base-units/second (Read via a summary tool. Supports the definition using 9,192,631,770 cycles of the caesium-133 transition.)