Building an Impossible Clock
The 18th-century horologist John Harrison claimed that he could make the world's most accurate pendulum clock, but his methods were scorned for hundreds of years—until someone proved him right.
As a rule, Donald Saff doesn’t collect clocks he can’t see inside of. It’s the harmonic entanglement of gears—and the skill needed to craft them—that first lured him to horology, the study of measuring time.
In 2004, browsing through a tiny shop in midtown Manhattan, Saff saw a 55-inch clock sitting on the windowsill. He was immediately riveted—the clock was beautiful, he recalled, in a way that reminded him of Grace Kelly. Looking at its insides, visible through the dusty glass, he could tell that the clock was in poor condition. But he could also see that it was a technical tour de force.
Examining the clock, Saff noticed something particularly remarkable: a grasshopper escapement. All pendulum clocks have an escapement, a swinging mechanism that pushes the pendulum at a steady rate over the seconds, minutes, and hours. The grasshopper escapement is a low-friction version invented by the clockmaker John Harrison in the early 18th century, and rarely used in modern clocks.
Actually, “rarely used” is something of an understatement—both 18th-century and contemporary horological communities have rejected Harrison’s pendulum designs. They were difficult to understand and seemed to contradict accepted knowledge about how to build clocks. Yet here was a fragment of Harrison’s design centuries later, an anachronism—and a seeming impossibility—preserved in steel.
The owner of the store told Saff the clock was too damaged to ever be restored. But Saff was persuasive; he was sure he could fix it. He took home the clock that day—and unknowingly set the wheels in motion for a renaissance of Harrison’s clock-making science, dismissed and ignored for 300 years.
Martin Burgess, who made Saff’s clock, lives 3,500 miles from New York on a rural plot of land in Essex, England. He is a clockmaker trained by blacksmiths, a restorer of Egyptian antiquities, and one of the foremost experts on chainmail in the U.K., if not the world.
In the 1950s, in a bookstore in London, Burgess bought a biography of Harrison, who is best known for inventing the chronometer, or sea clock. In 1714, Queen Anne of England passed the Longitude Act, offering a £20,000 prize (roughly $3 million today) to anyone who could devise a way to calculate longitude at sea; without it, British trade ships were losing lives and precious cargo. As Dava Sobel documented in her book Longitude, Harrison was able accomplish the task with a clock called H4. Besides for allowing sailors to calculate how far east or west they had traveled, the clock was extremely accurate, losing only 39.2 seconds over a 42-day voyage.
But in the years before he created his marine timekeepers, Harrison made precision pendulum clocks, also known as regulators. At the time, George Graham, the period’s most prominent clockmaker, was building clocks that were accurate to a second per day. Harrison, who was trained as a carpenter, not a clockmaker, made clocks almost entirely out of wood; according to Harrison, they were so accurate that they changed by only a second a month, far better than any other clock being made for use on land or sea. But he believed that he could do even better: By using his theories, Harrison claimed, it would be possible to make a clock with an error of one to two seconds per year.
Harrison left behind three manuscripts about his methods. The most important, published a year before his death, is titled: A Description concerning such mechanism as will afford a nice, or True mensuration of time. As the name suggests, the language is dense and convoluted. The contemporary British clockmaker George Daniels has called the text “rubbish”; Rupert Gould, the author of the biography Burgess read, described it as “gibberish.”
“Everybody said, ‘That’s a lot of nonsense,’” Burgess said. “‘Second to the month? No. Those things couldn’t possibly.’ Well, I was pretty sure they did.”
Burgess and another horologist, William Laycock, began to study Harrison’s writing more closely, and then to advocate for his theories within the horological community. In 1976, the two delivered a lecture to the British Horological Institute arguing for the validity of Harrison’s ideas, but they were generally dismissed—except by a few. In the audience was the engineer Mervyn Hobden. He had never met Burgess. But after the lecture, Hobden said, he felt compelled to approach him.
He wanted to do more than offer his support. Hobden proposed that they should form a group called the Harrison Research Group, dedicated to investigating the theories that Harrison had left behind. Andrew King, a clockmaker who had also been in the audience, said he would join too. Burgess and Laycock immediately agreed.
“They came forth and said we’ll use our specialties and we’ll communicate,” Burgess recalled. “I wasn’t out in the cold anymore.”
The group decided the best way to prove that Harrison had been right would be to build a Harrison clock using modern materials.
“No one would accept that Harrison’s ideas were scientifically sound,” Hobden said, but “that there were ideas buried in Harrison’s writing that had not been recognized. This was where the start was.”
The Harrison Research group met once a year in Greenwich, England. The original group soon shrunk to five—Laycock died from cancer in 1976, shortly after the lecture—and then grew to six with the addition of Jonathan Betts, the senior curator at the Royal Observatory, and the horologists Anthony Randall and Beresford Hutchinson. Each time the men gathered, they brought dry madeira, in honor of Harrison’s H4 correctly directing a ship to the port of Madeira in 1764. They toasted their wine to “the Master.”
When Burgess was asked to make a clock for the Gurney’s Bank of Norwich, the group decided to use this commission to test Harrison’s science. But they only had Harrison’s text to guide them—and it wasn’t the easiest of manuals. Like a person trying to communicate in an unfamiliar language, Harrison repeated the same message multiple ways, often confusing the main point. Some of the group even thought Harrison has intentionally written in a garbled manner, to protect his theories from outside eyes.
The first time he read it, Betts said, he was convinced that Harrison had simply written down a nonsensical collection of thoughts. “But the more I read it, the more I was convinced that, actually, Harrison was desperately struggling to make himself understood,” he said. “Far from trying to cloud the issue, he was trying to make it clear.”
The fact that the group members came from varied backgrounds, Betts said, helped to keep things moving forward as they began work on the clock. Some were engineers, some mathematicians, some horologists; Hobden even began to study 18th-century physics and mathematics, so that they could view Harrison’s work from the proper mathematical perspective. Slowly, Harrison’s approach became clear, and so did all the ways that it contradicted centuries of wisdom about pendulum-clock making.
* * *
Harrison’s science is unique because it looked at a clock as an ecosystem, with each of its organisms balancing out all the others to create a perfect equilibrium. For every reason that the pendulum might move faster that it should, for example, there’s a design element in place to slow it down again. It’s an elaborate dance of synchrony made up of seemingly disjointed pieces.
One of Harrison’s key theories is that the circular weight on the end of pendulum, called the bob, should be light, and the pendulum should have a large swing, or arc. Most horologists have incorrectly interpreted this as Harrison calling for the lightest pendulum possible, which wouldn’t make for a very effective clock—too light, and a pendulum would be disturbed by every breeze or passerby. Using this logic, it makes sense that clock makers of Harrison’s time usually did the opposite, making their pendulums heavy so that nothing could interrupt them. (Big Ben, a pendulum built with conventional methods, has a pendulum that weighs 600 pounds.)
Harrison figured out that the same stability could be achieved without the same heft: By creating a lighter pendulum that traveled a larger arc, he could increase the energy it stored, which made it even better than a heavy pendulum at resisting outside movement.
Other clockmakers didn’t rush to copy this idea, though, because the large swing created another obstacle. In pendulum clocks, larger arcs take a slightly longer amount of time to swing than short ones, a problem called circular deviation; Harrison’s light pendulum varied in the length of its swing, and each arc could take slightly more or slightly less than a second. Unsure how to resolve this issue, clockmakers stuck to their small swings and heavy bobs, which helped keep circular deviation to a minimum.
But Harrison had a fix. To create a light pendulum that circumvented the problem of circular deviation, he used something called circular suspension cheeks—a device that looks like a sideways number three laying on its side, with the curvy part (the “cheeks”) facing the bottom of the clock. The pendulum swings on a steel ribbon from the center, where the two cheeks meet. When the swing of the pendulum is smaller, it remains in the indented space and barely touches the sides of the cheeks. But as the arc gets larger, the steel ribbon holding the pendulum wraps up around the bottom of the curve, hugging its surface.
When the arc gets larger, the cheeks help to shorten the pendulum, because part of it is wrapped up around the curve of a cheek. Harrison knew that when a pendulum is shorter, it moves faster, which made up for the increased length of the arc. To ensure the perfect level of compensation, the cheeks have to be empirically tested over and over; Burgess said that his misunderstanding of Harrison’s directions about the cheeks cost him years in building the Gurney clock.
But after Harrison had perfected his cheeks to remove the circular deviation, he un-perfected them, altering the cheeks to account for another possibility: that external conditions could create a time discrepancy. Clocks don’t just have to compensate for internal fluctuations, like the variable swing of the arc; they also have to compensate for changes in barometric pressure and temperature. When barometric pressure goes up, it indicates that the air has become denser. As the pendulum swings through the dense air, the arc becomes smaller, causing the clock to run a little fast. Though Harrison, with his cheeks, had devised a way to make the perfect pendulum, he saw that a little circular deviation could negate the effect of the air pressure. It turned out that the imperfections were actually what made the clock run so well.
To compensate for the air temperature around the clock, Harrison invented the grid-iron pendulum, made with thin, alternating rods of brass and iron. Most metal expands when hot, so a pendulum can become longer in warmer temperatures, and as a result will take more time to complete its arc. But by using metals that have different expansion rates, Harrison created a pendulum that would remain the same length in all temperatures.
The final touch was Harrison’s grasshopper escapement. A traditional escapement, the mechanism that keeps the clock moving, slides in and out of the teeth of the wheel that push the pendulum; to keep this sliding motion smooth, the escapement needs oil or other lubrication. Running out of oil could stop a clock, while too much oil could make it run too quickly.
After he had carefully constructed his elaborate system of checks and balances, Harrison wasn’t going to let a little oil mess with the accuracy of his clock. His grasshopper escapement eliminates friction—and, as a result, the need for lubrication—by pushing the escapement leg with a short jump, so that it seems to hop out of each spoke rather than sliding.
It seemed that for every reaction the clock had, its equal and opposite reaction would emerge to balance it. And with the grasshopper escapement, there would be no interruption to this ebb and flow.
* * *
Harrison always built two clocks, to compare the results against each other as he made minute adjustments. For his assignment in Norwich, Burgess also built two clocks, so that he could make separate modifications on each one. Even though Burgess’s contract was for four years, 12 passed before he was able to complete just one of the clocks. And even then, the group didn’t feel they had fully grasped Harrison’s work. The clock they created was good, but not good enough.
“By the time we’d finished it, we were just beginning to learn what it’s all about,” King said. “It runs about a second a month. But it’s capable of something much better.”
After the first clock was delivered and installed, its twin, called Clock B, remained unfinished on a wall in Burgess’s workshop. Will Andrewes, a clockmaker who had studied under Burgess, used to visit his former mentor’s shop every weekend; when the two would have lunch in the garden, they would leave the shop door open, leaving Clock B vulnerable to invading breezes and temperature changes.
“By the time we came back you could hardly measure the difference of the clock, even with the wind blowing through the workshop,” Andrewes said. “So I had a sort of inward belief that the thing could work.”
In 1993, Andrewes organized a Longitude Symposium at Harvard University to discuss the history of longitude and Harrison’s science. Burgess delivered a lecture at the symposium called “The Scandalous Neglect of John Harrison,” described the Harrison group and its work on the Norwich clock. Nevertheless, the horological community remained skeptical, and Clock B remained at Burgess’ home, quietly keeping time as it passed by.
* * *
By the time Saff bought his clock in 2004, Burgess and his efforts had been mostly forgotten in the horological world. There had been no news about him, or Harrison, in years.
When Saff brought his purchase home from the tiny Manhattan shop, he found Burgess’ signature on its bottom. This particular clock, he would later learn, had been commissioned in the 1960s by Simon Napier-Bell, a British music manager who had requested a clock so mesmerizing that it would distract his clients as they signed contracts. The clock ended up in New York, discarded in a basement and finally donated to the clock shop as a gift. Burgess had, of course, used Harrison’s grasshopper escapement to make it. As Saff restored the clock, he began to research Harrison’s work; eventually, he was able to get it working again.
Then, by chance, he met Andrewes during a visit to Harvard. Amazed to find a man who knew Burgess personally, Saff told him about the Napier-Bell clock he’d fixed up. On Andrewes’ advice, Saff found Burgess’s lecture printed in a book, which is how he learned about the Norwich clock and its twin.
“The second clock wasn’t complete!” Saff said. “And I thought, my goodness, this is the scandalous neglect of Martin Burgess.” Through Andrewes, Saff contacted Burgess and offered to buy Clock B. Burgess was reluctant, but the fact that Saff had fixed the Napier-Bell convinced him the clock would be going to a worthy home. In 2009, Saff traveled to England and took Clock B off the wall of Burgess’s workshop.
Saff hired Frodsham and Co., a leading horological firm in London, to complete the clock. He also reunited Hobden and King from the original research group to help with the elaborate testing and adjustments that would need to take place.
A finished clock began to emerge. The group built dials, and then a hand, and when Frodsham and Co. began to test it, Burgess Clock B performed extremely well. For 80 days, it kept time within a second. But before any claims could be made about its accuracy, the clock needed to be tested in a more official capacity. In April 2014, Clock B was sent to the Royal Observatory in Greenwich, where Rory McEvoy, the curator of horology, welcomed it with skepticism.
At the end of a hundred days the clock had lost only five-eights of a second, within the range Harrison predicted in Concerning Such Mechanism. The Guinness Book of World Records awarded the clock the distinction of “most accurate mechanical clock with a pendulum swinging in free air.” Now, more than 20 months later, it is still running at the same accuracy.
“At the moment it’s absolutely in step with Coordinated Universal Time. There is zero error,” McEvoy said.
This level of accuracy is “mind-bogglingly good,” Betts said. “It’s absolutely rewriting the horological textbooks. It’s proving that what Harrison said over two centuries ago.”
Harrison was, finally, validated. And so was Burgess.
But Burgess, now 84, didn’t gloat for too long. Instead, he harped on a minor temperature compensation that needed to be adjusted in the pendulum. “As Harrison said, the pendulum should be shorter when warmer,” Burgess told me over the phone recently, trailing off into the details of Harrison’s rules about heat and cold.
Other members of the group aren’t slowing down, either. King is building two replicas of the regulators, made of wood rather than modern materials. He hopes they’ll prove that Harrison’s theories even with the rudimentary materials the 18th-century clockmaker would have used. They will be ready to test next year.
Meanwhile, Saff and Hobden will keep modifying Clock B. Once the pendulum is altered, McEvoy plans to put the Burgess Clock in the main gallery of the Royal Observatory, where it will keep time beside the original Harrison clocks.
“The attitude towards Harrison’s work was that it was a glorious dead end,” Andrewes said. “We’ve been able to prove that in fact it was a glorious new beginning. And here we are, still learning.” The work of the Master, it seems, is never done.