# Time synchronization of circuits in different places

How can i synchronize times of circuits placed quite far(a few hundreds of meters) from each other?

Are these PPS signals relative in time (as in the figure)? Or is it generated at the beginning of every second?
What could be the alternative options?

• Use very high accuracy RTC clocks and some kind of synchronization method. This is pretty common in industrial control (think dozens of robotic axis controllers that need to coordinate movement). How close do the times need to be? Jan 9, 2022 at 19:43
• 40-50ms would be enough. RTC clocks need to set curret time, right? How can i start the time counting precisely of all RTC clocks? Jan 9, 2022 at 19:46
• Sound travels about 1m in 3ms. Electricity travels about 0.6 times the speed of light in wires. Light travels at the speed of light. Can any of these help in syncronizing to 40-50 ms? Jan 9, 2022 at 19:52
• 50ms means about 17meters. What happens if the distance between the modules more than 17meters? Jan 9, 2022 at 20:01
• Because you are using sound waves, I assume that your "quite far" is different from an astronomer's "quite far" or even a pilot's. Could you edit your question to put engineering units onto your "quite far"? I.e., do you mean meters, tens of meters, hundreds of meters, etc? For the purpose of discerning the time of arrival of sound waves, yes, the PPS service is synchronized. Jan 9, 2022 at 22:03

Yes, the PPS pulses from all GPS receivers — no matter where they are positioned — occur at the same time, with a tolerance that's typically on the order of ±50 ns.

• Way before GPS existed I used WWVB on a board to synchronize 16 channels 100 m apart for seismic wave recorders in 1974 at U of M. 50 ns is kind of meaningless for sound and delays over 300 km for seismic research from blast source are far greater than 1 second for 1 pps. so that results in ambiguity Jan 9, 2022 at 20:20

A much cheaper solution could be to broadcast a 1 pps pulse at RF in a license-free band and then measure the audio pulse to measure distance relative to the synchronized audio burst and the 1 pps RF burst.

Rather than have a GPS Rx at each endpoint just a low-cost RF Rx can give the desired resolution.

• Thanks. What would be other solutions? I will be glad if you share. Precise RTC can be a solution, there was like a above comment? Jan 9, 2022 at 20:55
• What outputs do you need? Absolute , relative , resolution? , power? cost? Jan 9, 2022 at 22:08

Are you sure sync to within 40-50ms is good enough (this means a distance resolution of tens of metres)? Is "a few hundreds of meters" less than 1000m?

If you have a local area network connection available, clocks can be synced (eg NTP) to a good enough level. If you need millisecond-level sync that might or might not be good enough (but with wired gigabit ethernet there is a hardware protocol that can get you to that level). If you need microsecond level sync there are non-GPS (if that is helpful) techniques that can be used (optical, wired, RF) .

GPS PPS is good for <<1us, and since it is something you mentioned I assume that it is otherwise applicable for you (cost, development time, power consumption, view of the sky, etc), in which case you may as well use it. In addition to the 1pps signal, GPS modules have a UART output from which you can parse text messages to determine what the time is corresponding to the pulse it produces (because your signals might occur across a 1s boundary)

• thank you. PPS pulses don't have be occur at the start of every second, is this right? I can learn over UART it's time pulsed. Jan 10, 2022 at 18:55