Precision timing is an essential component of modern communications and navigation systems. While the signals from GPS satellites are sufficient for many applications, often these systems require a local precision timing reference. This is particularly important for applications which require higher stability and/or availability than GPS can provide. In these applications, it is often valuable to form a composite clock, which exploits both the short-term stability of the local clock and the long-term accuracy of GPS. The local ensemble is generated by a technique known as “disciplining” in which the timing output of the precision clock is compared to that of the GPS system and the frequency of the clock is gently steered into compliance with the GPS. Whether the local clock is a crystal oscillator or a high precision atomic clock, optimum disciplining requires an understanding of the noise properties of both the clock and the GPS reference source in order to avoid the risk of accidentally degrading the clock’s performance while trying to improve it.
GPS is a valuable reference source because of its ready availability and long-term stability and accuracy. On shorter time scales, however, relatively low cost clocks, such as quartz or rubidium oscillators, have significantly better stability, as shown below in the Allan Deviation (ADEV) stability chart of Figure 1.
Another good example of this is Symmetricom’s Quantum SA.45s Chip Scale
Atomic Clock (CSAC). Particularly considering its low power (<120mW)
and small size (<16cc), the CSAC demonstrates short-term stability.
Nonetheless, its long-term accuracy and stability can be improved by
periodic recalibration to GPS. Even intermittent disciplining to GPS can
improve the long-term drift of the system while the CSAC can “clean-up”
the jitter on the GPS signal.
As seen in Figure 1, the CSAC instability at one-second averaging time is sy(t =1)˜10-10and improves with averaging time as 1/t1/2, typical of atomic clocks. At longer times, here t >1000 seconds, the CSAC performance degrades due to temperature sensitivity and long-term drift. On the other hand, the GPS signal is relatively unstable at shorter averaging times, sy(t =1)˜10-8, but improves as 1/t such that it has superior stability at longer averaging times. To optimize the system performance on all time scales, the CSAC should be disciplined to GPS with a loop time constant equal to the averaging time at which the two noise processes intersect. From Figure 1, the two lines intersect at ˜3000 seconds and thus this should be chosen as the loop time constant. An example of the combined stability of GPS and CSAC can be seen in the ADEV chart below where the CSAC was optimally disciplined with a time constant of 3000 seconds.
Join our online Radio Show on Friday 11th July starting at 2:00pm Eastern, when EETimes editor of all things fun and interesting, Max Maxfield, and embedded systems expert, Jack Ganssle, will debate as to just what is, and is not, and embedded system.