Time-based synchronisation is being used by the Met Office for synchronising distributed instruments on a global scale, writes Victoria Murtland
For engineers working in design validation and automated test the precise timing and synchronisation of measurements is critical.
Using traditional rack and stack instruments, this has often proved difficult to achieve.
As precision requirements increase, higher levels of synchronisation are required, necessitating the use of technologies such as GPS satellites, dedicated trigger buses and even more advanced crystal oscillators.
PXI, the automated test bus, has capabilities for timing and synchronisation applications. It works with two different types of synchronisation architectures, time-based and signal-based, and features built in capabilities for extremely accurate measurement timing, such as a 100MHz system reference clock on the backplane.
Time-based synchronisation
Time-based synchronisation is used for synchronising distributed instruments, even on a global scale. An example of this is the work done by the Met Office’s Arrival Time Difference Network (ATDnet).
Lightning emits electromagnetic radiation over a broad frequency range. Our eyes detect it by the flash of visible light, yet it emits an even stronger signal in the very low frequency (VLF) radio-wave part of the spectrum (around 10kHz), which travels over thousands of kilometres.
The Met Office were upgrading a system that can detect the VLF radio-wave emission from lightning strikes, synchronised to less than 100ns, to accurately record the arrival time and the structure of the VLF signal. These signals are mainly used to locate thunderstorms for general weather prediction, although the outputs can also be used for other applications, such as monitoring volcanic ash clouds.
The Met Office required a system that could provide accurate synchronisation between instruments situated thousands of miles apart and that could be deployed for a number of years. This made a PXI-based system, using time-based synchronisation over GPS, an appealing solution.
Each VLF receiver was connected to a PXI-based software-defined instrument configuration synchronised with commercial off-the-shelf timing cards and GPS. This fulfilled their first goal by providing a synchronised clock so that data was recorded with a timestamp accurate to 20ns at each receiver location. GPS timing and synchronisation is often used to timestamp events from multiple instruments, to correlate measurements or to generate hardware events at user-specified times. However, this is only one method of time-based synchronisation.
Other absolute time reference protocols such as IEEE 1588, NTP or IRIG-B are also used with time-based architectures to achieve synchronisation over large distances and can be implemented with a similar approach.
Although the hardware requirement for implementing these protocols can be straightforward using standard PXI cards, often the challenge to success lies in the software development.
By using a high level development environment, like LabView, many of the protocol details between master controller and submissive time clocks could be abstracted, making application development quicker. Time-based systems experience jitter in transmission time, so for accuracy when triggering events, signal-based architectures must be deployed.
Signal-based synchronisation
Signal-based synchronisation architectures use on-chip synchronisation, such as system reference clocks that are physically connected between subsystems, typically resulting in the highest precision synchronisation, if chassis are within close proximity.
The PXI standard features integrated timing and synchronisation, incorporating a dedicated trigger bus, star trigger bus and a 10MHz system reference clock, which can be increased to a 100MHz differential system clock if a PXI Express chassis is used.
It is possible to synchronise modules within a single chassis, but also modules in multiple chassis by sharing the reference clock.
Through shared timing and synchronisation, the timing accuracy of measurements can be improved and advanced triggering schemes can be applied. With differential star triggers, increased noise immunity and industry-leading synchronisation accuracy (250ps and 500ps of module-to-module skew respectively) can be achieved.
As engineers face ever more demanding challenges, there is no doubt that even more tightly synchronised measurements will be required. On PXI, the system reference clock is driven by an independent oscillator frequency source, accurate to around 25 parts per million (ppm).
For applications requiring even greater accuracy, it is possible to use timing and synchronisation modules with an oven-controlled crystal oscillator. The backplane system reference clock can be replaced with an oscillator frequency source that is accurate to 50 parts per billion (ppb). When tracking lightning, the sky’s the limit when it comes to synchronisation.
Author is Victoria Murtland, applications engineer, National Instruments UK & Ireland