Dear Ait,

welcome!

The proposal which i hinted in my second post will enable you to measure the delay time between the master node and slave nodes. The basic idea is to use signal like the GPS one to estimate the transmission delay.

You have a specific problem and it needs investigation to concept the system and implement it within the scope of the existing systems.

Best wishes

Dear Ait,

For synchronizing multiple clock sources you have to have a master oscillator with the accuracy wanted. There are atomic clock sources which are highly stable and highly accurate. The oscillator with a lower accuracy is the xrystal oscillator with oven. Its accuracy can be as high as one nanosecond.

I assume that that such oscillators will be suitable for your application.

This oscillator is installed in the station of a center point and distributed to the other stations to synchronize their local clock generators using phase locked loops.

Best wishes

Dear Aparna Sathya Murthy thank you for you answer, i try to encode patent ideas. Effectively Abdelhalim abdelnaby Zekry high accuracy and stability clocks are available in the market with low skew. The problem that I need to time synchronization each clock with a Master, because each of them drift on times in different ways. Then, I need to measure or estimate drifts of clocks to adjust it. White Rabbit Project suggests sub-ns clock synchronization using fiber optics, I wonder if there is another approach without physical cable. Thank you in advance.

Dear Ait,

I think you can transmit the time stamp pf the master clock and transmit it wirelessly in a direct line of site channel to the clock to be synchronized by the same way used in the GPS format where spread spectrum is used. Then you can estimate the transmission time and add it to the stamp time to get the corrected time. This is just a proposal. As the distance is much smaller than that of the GPS you may get much higher accuracy.

The other way is that you can transmit and regenerate the master clock at the other sites. Then the regenerated clock will be a copy of the master clock. Which means it is synchronized with master clock as it generated from it.

Best wishes

Effectively Abdelhalim abdelnaby Zekry , it will be practical if we use line to transmit clock or "timestamp" from master to slaves. In my case, each of networks nodes is independent and has a self accurate clock (no physical link between master and slaves). Because, nodes are not in a fixed position or fixed distance between master and network nodes. I think some kind of time measurement and clock adjustment using wireless communication between master and each node is needed. GPS accuracy is not sufficient in our case, maybe a hardware measurement of time delays is required. ..

Dear Ait,

welcome!

The proposal which i hinted in my second post will enable you to measure the delay time between the master node and slave nodes. The basic idea is to use signal like the GPS one to estimate the transmission delay.

You have a specific problem and it needs investigation to concept the system and implement it within the scope of the existing systems.

Best wishes

I think one must first determine over what time frame you need to maintain synchronization. Long-term sub-ns synchronization among mobile platforms using wireless is quite a challenge. I am not aware of anyone achieving that - even GNSS systems with their significant expense and infrastructure require constant effort and adjustment. White Rabbit can be applied over any link, but assumes some constraints on propagation time variability. GNSS systems can provide highly accurate time or highly accurate position, but generally not both simultaneously unless they are operated in an over-constrained mode such as using additional channels and constellations. (The typical high-accuracy GNSS time source operates in over-constrained mode by knowing it's location very precisely. This is often achieved by allowing it to map its location over a long period before it starts operating as a time source.) For tropospheric links local atmospheric variations such as temperature and pressure can effect propagation times in the ppm range, primarily due to path bending. Plus it is difficult to approach the level of clean detection of edges from an RF signal that can be achieved with a fiber optic link due to the fiber link's much greater S/N. Short-term synchronization is relatively easier to achieve; one method is to start with two highly stable sources that are synched while stationary and co-located. But sub-ns is still difficult and I have not seen it done outside of a research setting.

Please look at the book

Digital Communication

by

Proakis

at

Northeastern University

Boston, MA

USA

Thank you Reggie A Ratcliff to summarize the problem and Ziad Sobih for the book.

Dear Ait Mansour El Houssain ;

The Network Time Protocol (NTP) is a networking protocol for clock synchronization between computer systems over packet-switched, variable-latency data networks. ... NTP is intended to synchronize all participating computers to within a few milliseconds of Coordinated Universal Time (UTC).

Please look at this article:

Clock Synchronization for Wireless Networks

December 2004 , DOI: 10.1007/11516798_29

Thank you Abdelghani Rouini for you answer