New Olympic Clocks Go for the Gold
In the 1994 Lillehammer Olympics, German racer Georg Hackl took the gold in the men’s luge by just 13 milliseconds, or thousandths of a second, then the closest finish in Olympic history. Four years later at Nagano, countrywoman Silke Kraushaar won the women’s luge by just two milliseconds-the new closest finish in Olympic history. “Two thousandths is nothing!” the silver medalist-yet another German-said at the time. “It’s one centimeter, I think. Unbelievable!” Which led Olympic officials to ask: could their timing equipment really be believed, right down to the millisecond?
It was a good question. Olympic sled events-bobsled, luge and now skeleton-have for decades been timed by photoelectric sensors; the timer starts when the racer crosses a light beam at the top of the course, and stops when the racer crosses a second beam at the bottom.
Kraushaar’s victory combined the results of four separate trials-a total of eight time measurements. Exact specifications for the Nagano timing system are not available, but two milliseconds, over four trials, likely fell within its margin of error.
Would a better clock have taken the gilt off Kraushaar’s gold? No one can say, but the U.S. Olympic Committee wanted to avoid similar questions at the Salt Lake games. In a saga that begins with angry athletes and allegations of cheating and continues through last Friday’s atomic clock-calibrated luge finish, organizers say the 2002 Winter Games boasts the most accurate timekeeping in the sports world.
A Matter of Time
The story of how it got this way reads more like a detective novel than the sports page.
In the mid-nineties, even before the Nagano Olympics, American luge racers approached the U.S. Olympic Committee’s principal engineer, Tom Westenburg, with a complaint. They suspected that some countries’ teams-Westenburg wouldn’t say which-were using reflective suits to trick the timers on certain courses. On older courses, the timer’s light transmitter and receiver sat next to each other. A beam of light left the transmitter, crossed the track, and bounced back to the receiver.
The U.S. racers charged that some athletes used highly reflective suits that didn’t trigger the starting timer until the racer’s helmet passed through the beam. Since luge is raced feet first, this could shave as much as 300 milliseconds off a single run-enough to move a team from eighth to first in this year’s Olympics.
Westenburg says he was skeptical, but that he trusted the instincts of the racers-among them an astronaut, an aeronautical engineering student and a commercial airline pilot. “They’re a sharp bunch of people, and when they saw something, I took them seriously,” he says. He conducted some tests and concluded that they were right. Today, thanks in large part to his research, no world-class sled competition uses reflective timing systems.
The experience also got him thinking about timekeeping-a less glamorous field than the lasers and aerodynamic sensors that he usually designs to help train world-class athletes. “I’ve seen athletes pour their hearts into this sport, and I want everything to be judged accurately,” he says. “I want everything to be fair.” Westenburg traveled the world, measuring the accuracy of sled course timers. In the summer of 1998, he presented his surprising results to the conference of the International Sports Engineering Association.
The problem went beyond cheating, he told his audience. The International Olympic Committee and the International Luge Federation had until that point left it up to individual courses to guarantee the accuracy of their timekeeping. But many courses employed older photoelectric timers, which use a light beam that flashes on and off at only a few hundred cycles per second (hertz). At this level of precision, the systems produce an uncertainty of more than a millisecond in each sensor-or two milliseconds per sledding run.
The biggest shock: even the Olympic course in Salt Lake City, built only the year before, had a margin of error of three milliseconds per run. As soon as the Salt Lake Organizing Committee learned of the problem, Westenburg says, they agreed immediately to replace the system with a more accurate one.
But none existed, so the Committee requested proposals for a new system from the world’s top timekeeping companies. In the fall of 1998, they accepted a bid from Leipzig, Germany-based Wige Data on the condition that the firm work with Westenburg and his research team: engineer Tim Conrad and technician Robin Korf. In the spring of 1999, Wige and Westenburg’s team installed at the Salt Lake park a triply redundant system of photoelectric timers that cycle at speeds of more than 9,000 cycles per second (nine kilohertz). With faster light beams, and various other improvements, Westenburg’s team calculated that their timers would be accurate to below half a millisecond-three to ten times better than most timers then in use.
To be certain, Westenburg and his colleagues subjected the park’s timers to more rigorous tests than any previous system. Standard procedure would have been to take the timers to a certified laboratory, where technicians would test each part independently. But, he says, no one had tested an entire system on-course. “They weren’t testing end-to-end, from the start through the system to the final timing light.”
So in the summer of 2000, after performing some preliminary tests himself, Westenburg contacted the United States’ ultimate authority on time, the National Institute of Standards and Technology. The Institute, which among other duties decides the official U.S. weight of a kilogram and length of a second, put Westenburg in touch with Marc Weiss, a mathematician in its Time Frequency Division. Although less accustomed to working with bobsleds than with the Institute’s cesium fountain clock-which measures time in increments of 10-10 (100 trillionths) seconds-Weiss agreed to help test the accuracy of the Salt Lake timers. “The millisecond level is pretty coarse for us, but the Olympics is a big thing,” Weiss notes.
Weiss joined Westenburg’s team in calibrating the Salt Lake timers using a shutter system conceived by Conrad. The shutters-similar to ones used to control the scalpel in laser surgery-were controlled by signals from Global Positioning System satellites, which are set to National Institute of Standards and Technology’s cesium clock and accurate to 100 nanoseconds, or one ten-thousandth of a millisecond. During the test, a signal from the satellite closed the shutter at the top of the course, which blocked the photoelectric sensor and started the timer. Precisely 50 seconds later, another satellite signal closed the bottom shutter, stopping the timer. By this method, the team determined that the course’s system was, as predicted, accurate to below half a millisecond, at least three times better than that of any other course in the world.
Since the test, three other sled courses-the Olympic parks at Calgary, Canada, Lake Placid, NY and Innsbruck-Igis, Austria-have announced plans to implement timers based on Salt Lake City’s. Westenburg, who gets little sleep these days as he races from setting up one event to watching another, says he is happily surprised at how smoothly the technology behind the 2002 Games, including the sled races, has held up.
As it turned out, Salt Lake’s system was put to the test on February 15, when the American two-man luge team of Brian Martin and Mark Grimmette edged out teammates Chris Thorpe and Clay Ives for the silver by just four milliseconds.
Weiss says he doesn’t foresee Olympic timekeeping needing to do much better than that. Not because systems can’t get more accurate, he says, but because they don’t need to. “You have a 100 millisecond reaction time in the best athletes. At the one millisecond level, are we really measuring skill?” he asks. “Maybe the next digit after that doesn’t measure anything.”
The inside story of how ChatGPT was built from the people who made it
Exclusive conversations that take us behind the scenes of a cultural phenomenon.
ChatGPT is about to revolutionize the economy. We need to decide what that looks like.
New large language models will transform many jobs. Whether they will lead to widespread prosperity or not is up to us.
Sam Altman invested $180 million into a company trying to delay death
Can anti-aging breakthroughs add 10 healthy years to the human life span? The CEO of OpenAI is paying to find out.
GPT-4 is bigger and better than ChatGPT—but OpenAI won’t say why
We got a first look at the much-anticipated big new language model from OpenAI. But this time how it works is even more deeply under wraps.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.