Olympic Timing - How Do They Do It?

[HLA-27b]

The info out there seems quite vague and superficial, certainly not detailed enough to satisfy this forum.
So here are a few questions to keep us busy for a while:


What are the reference standards used? 10MHz rubidium?
How is the clock distribution being done, especially when you have multiple start and finish lines like in swimming?
How do they filter out false readings?
Do they have a human in the loop or is it completely automatic?
And most important of all, is it possible to build something similar at amateur level?

[EEVblog]

Do they have a human in the loop or is it completely automatic?

In swimming at least there is a human there to ensure each swimmer actually touches the sensor plate (and may have a manual timer as backup?). IIRC happened at the last olympics/commonwealth games/word champs or something. Someone swum a clear world record time, but the sensor plate didn't work. He/she officially won because of the manual timing/official present to verify, but no official WR time could be recorded.

Dave.

[EEVblog]

Given that all official times are recorded at 0.01s resolution(?), you'd need a system at least an order better, or 1ms.
Using fibre optic (or even coax) connections between start and finish gets you a skew of only a few µs/km, so well within the required error if using one master clock and one trigger signal.
But it would be interesting to know the exact system used.

Dave.

[Sionyn]

that be a good video for the blog dave should give you a lot of exposer

[HLA-27b]

Here is how they did distance measurements in Sydney.

http://www.leica-geosystems.com/common/shared/downloads/inc/downloader.asp?id=16565

Typical accuracies +- 1mm with standard Leica total stations. These are interferometers working in infrared with lots of other bells and whistles. I am sure they will use the same gear in London as well. So you should see them if you watch out.

edit: Here is what is inside one


Typical usage - The total station sends a laser beam to the prism (a retroreflective prism, the black bulbous thing at the end of the staff held by the red dude) which is reflected back. Time of travel is measured and corrected for temperature and in some cases for other air density factors. The corrected time is used for rough distance calculation. Precise distance is determined interferometrically. Some total stations can track a prism without human assistance.


 

that be a good video for the blog dave should give you a lot of exposer

Certainly

[G7PSK]

Quite likely that they will use GPS signals for the clock, as it is cheap and accurate.

[HLA-27b]

[HLA-27b]

[HLA-27b]

[NiHaoMike]

At work, I routinely use an automated system to deskew probe systems to well within a few ps. Deskewing something to a fraction of a ms would be very easy in comparison.

[Sionyn]

A conversation with sprinter Michael Johnson at the Sydney Olympics caused Peter Hürzeler of OMEGA Timing to realize that even with speakers, the speed of sound was still slowing down the farthest athletes. Johnson's reaction time, Hurzeler said, 'was 440 thousandths of a second. Normally athletes leave between 130 and 140 thousandths of a second. ... I asked him, why did you have such a bad starting time?' Turned out, Johnson was in the ninth position, and the sound of the gun was reaching him too slowly.

"In addition after a four year developmental process, a new false start detection system is being introduced this year that will abandon movement in exchange for 'measurement' of pound-force against the back block to determine sprinters reaction times. 'We are measuring the time between the starting gun and when the athlete is moving because to leave the starting block they had to push against and this power is very high' says Hurzeler. 'We did a test last year with Asafa Powell and he was pushing 240 kilograms (529 lbs.) [so] as soon as he gives the time to push against the starting block, it means he will like to leave and we are measuring this in thousandths of seconds and if somebody is leaving before one hundredth thousandth of second, it's automatically a recall, it's a false start.' In track every event is timed to 1/10,000th of a second, and Omega takes 2,000 pictures per second from right before the start of a race to its finish, as backup.

http://www.theatlantic.com/technology/archive/2012/07/the-speed-of-sound-is-too-slow-for-olympic-athletes/260413/

[HardBoot]

Olympic timings systems are problematic at times, but high speed cameras ensure any results can be contested by hand without error.

[Dawn]

This hasn't been a question that I've pondered and maybe technology has surpassed the last time that I read anything about it. I recall reading an article sometime during the early to mid half of the 90's that they were using video and SMPTE time code and that technology had become the final word in legal cases with regard to time. Technology moves on.