Guide
Clock synchronization in race timing: GPS, NTP and why a one-second difference ruins results
· 6 min read
In short
- If two devices show different times, split times and results will differ by the same amount. That is why they must be synchronized before the start.
- Without synchronization, a phone’s quartz crystal can drift by several seconds per day: at a drift rate of 10 ppm, the difference is approximately 0.86 seconds per day.
- GPS provides highly accurate time from atomic clocks aboard satellites when used with a specialized receiver. NTP synchronizes clocks over the internet, usually to within tens of milliseconds.
- For results recorded in whole seconds, clocks must be synchronized so that the difference is well below one second; timing to hundredths requires hardware-based solutions.
Discussions about race timing usually focus on the equipment: readers, antennas and chips. Less attention is paid to something equally important—the clocks. If two devices record the start and finish, they must use the same time.
Why clocks drift apart
Electronic clocks run on quartz crystals, which are not perfect. A drift rate of 10 ppm, or ten parts per million, amounts to approximately 0.86 seconds per day—more than enough to corrupt split times recorded by two devices. If a device has not been synchronized for a long time, the difference may reach several seconds.
GPS
GPS satellites carry atomic clocks, allowing a receiver to derive highly accurate time—on the order of tens of nanoseconds with a specialized receiver. Professional timing systems use this technology to synchronize devices equipped with specialized receivers. Phones obtain their time primarily from the network. The drawback is that the receiver needs a clear view of the sky.
NTP
Network Time Protocol synchronizes clocks with a reference server over the internet, usually to within tens of milliseconds and even more accurately on a local network. Phones use it automatically when they have a signal. In locations without internet access, devices should be synchronized before heading out onto the course.
How much accuracy is enough?
For results recorded in whole seconds, only a small difference between device clocks is acceptable—on the order of tenths of a second or less. The exact value depends on the rules and the expected gaps between competitors. For hundredths and thousandths of a second, as explained in Race timing standards, hardware synchronization and photocells are used. The technologies are compared in Race timing accuracy.
Practical procedure
- Before the start, synchronize every device with the same time source.
- Disable manual clock settings and leave automatic time enabled.
- When using multiple checkpoints, check the difference between devices.
- After the event, compare the times recorded by the primary and backup devices.
Learn more about backup systems in Backup race timing and about typical problems in The most common race-timing mistakes.
How TAQIO handles it
TAQIO records times relative to its own offset from the reference time, so an internet outage does not affect timing. Even so, it is worth setting every device to the same time before the start so that split times from different devices match.
Frequently asked questions
Why is clock synchronization important in race timing?
If two devices record the start and finish, any difference between their clocks directly affects the final time. A one-second difference between the clocks produces a one-second error in the result.
How accurate is a phone’s clock?
When synchronized through a network or GPS, it is highly accurate. Without synchronization, it drifts: at a drift rate of 10 ppm, the difference is about 0.86 seconds per day.
What is NTP?
Network Time Protocol is a protocol that synchronizes device clocks with a reference time source over the internet, usually to within tens of milliseconds.
What should I do if the event operates offline?
Before the start, synchronize every device with the same source and use an app that records times relative to its own offset from the reference time while storing data locally.