Holdover is a technique used in telecommunications to maintain accurate timing and synchronisation of equipment in the event of a temporary loss of timing signals. Accurate timing and synchronisation of devices is crucial for ensuring efficient and reliable data transmission in a telecommunications network. With 5G, synchronisation is required for the system's basic functionality and higher-level collaborative performance enhancements of the radios.
Holdover enables devices to maintain accurate timing in the event of a temporary loss of timing signals, such as when a signal is interrupted due to source failures, network congestion or a network fault. During a synchronisation holdover, a device continues to generate its timing signal based on its last valid timing signal received before the loss of signal, using internal oscillators and other mechanisms to maintain the accuracy of the timing signal. The holdover period can last from a few seconds to several hours, depending on the requirements of the specific network.
Holdover is an important concept for network operators, as it enables them to maintain service continuity and reliability during network disruptions or outages. It also helps keep service costs down. Truck rolls can be expensive and time-consuming for network operators, as they require resources such as personnel, equipment, and travel expenses. Unplanned truck rolls are even more expensive because of the urgency and unscheduled nature of the work. Having the ability to have synchronisation holdover can help minimise service disruptions and reduce the need for unplanned on-site technician visits.
In practice, network operators typically decide on a holdover period to be long enough to ensure uninterrupted service during synchronisation disruptions. The holdover period is usually shorter than the maximum tolerable outage duration for the specific service to ensure the equipment can re-synchronize with the primary synchronisation source before the outage duration exceeds the service requirements. The holdover period can vary depending on the quality and stability of the internal oscillator in the network component, as well as the temperature and other environmental factors that may affect its performance. Depending on the equipment and conditions, the holdover period generally ranges from several minutes to several days.
The typical duration of holdover desired by the operators in the 5G DU equipment range from 4 hours to 72 hours, considering various application scenarios. For a campus or enterprise network, equipment can be serviced within a few hours of failure once notified because of the availability of technicians onsite. For remote stations, days are required to apply for access permission, approval of access and actual maintenance and restoration work. A 24-hour holdover for the 1.5uS of phase stability is a starting point for the long–term holdover capability.
Global Navigation Satellite System (GNSS) based synchronisation is the primary synchronisation source in many telecommunication networks. In several network deployment scenarios, there may not be backup synchronisation from the network, and GNSS-based synchronisation will be the sole synchronisation provider of the system. However, a GNSS-based synchronisation failure can occur due to several reasons:
To mitigate these risks, systems that rely on GNSS for synchronisation often have backup systems, such as atomic clocks or other independent timing sources. They may also use techniques to detect and mitigate interference or spoofing attacks.
A typical synchronisation system consists of a primary reference clock (such as a GPS or atomic clock) or a clock derived from the network (using protocol or physical methods), which provides a highly accurate timing reference to network elements. However, synchronisation holdover comes into play in the event of a loss of the primary reference clock due to external factors like a GPS signal outage or loss of connection to the network primary reference clock.
During the holdover period, the network element uses its internal oscillator to maintain timing and synchronisation. The primary reference clock may be more accurate than the internal oscillator. Still, it can provide sufficient accuracy for a certain period, typically a few hours to days.
Atomic clocks are commonly used for long holdover, with Caesium and Rubidium-based clocks being the most popular. Caesium atomic clocks are widely used due to their stability and accuracy. Rubidium clocks are smaller and less expensive but have a shorter holdover duration. Chip-scale atomic clocks (CSACs) are smaller and more affordable versions of atomic clocks that focus on low power consumption. GNSS-disciplined high stability oscillators also provide 24-hour holdover stability, but they are expensive and may not be suitable for widespread deployment. The cost of Caesium atomic clocks ranges from tens of thousands to hundreds of thousands of US dollars depending on the model and features, while Chip Scale and Rubidium clocks are generally less expensive ranging from a few thousand to tens of thousands of dollars. However, the cost and size of these solutions may be too high for widespread adoption in 5G networks.
Equipment vendors need cost-effective solutions that fit into the current form factor of the equipment, thus likely a similar profile of clocks and oscillators to go into the systems. Many vendors provide a single skew of components so as to minimise the component loading profiles with Contract Manufacturers who are building the equipment.
Quartz based crystal oscillators are commonly used in most of the equipment clocks today. Until recently, the stability of clocks for a certain form factor acceptable for the equipment are one order of magnitude less than what basic atomic clocks could achieve. The short-term stability of the Quartz based solutions that previously measured minutes to hours has now been extended to tens of hours. A 25 x 22 mm size SMD footprint has been the standard stratum 3E stability level oscillator for more than 20 years now and features in most telecommunication equipment.
Several factors can contribute to oscillator instability; however, the two most significant contributors are temperature fluctuations and time. Temperature changes can affect the frequency of an oscillator, causing it to drift. Ageing is caused by gradual changes in the physical properties of the crystal that affect its resonant frequency. Recently, techniques have been developed to perform temperature compensation to stabilise the oscillator's frequency over a range of temperatures and use ageing models to predict and correct for frequency drift over time by using a traceable reference inside the oscillator.
In addition, thermo-mechanical design improvements, fundamental innovations on resonator design and integrated high-performing electronic circuitry, have propelled Quartz based solutions in recent years to achieve 24-hour holdover at a minimal cost increase. The solutions fit onto existing form factors with minimal design changes, enabling equipment to achieve holdover capabilities at economical and affordable costs.
If you would like to know more about our Stratum 3E oscillator with 24-hour holdover, visit our PPS Disciplined OCXO page or get in touch with us.
Subscribe to our emails now!