Expert Advice: A Leap into the Unknown?

March 1, 2015  - By

 

By Mark Sampson

A leap second will be introduced this year at 23:59 on June 30. This phenomenon comes around periodically and is necessary for keeping Coordinated Universal Time (UTC) in line with the small vagaries of the Earth’s slowing rotation. Although it is an event that will pass unnoticed by the majority of people, it has implications for anyone involved in the development of GNSS-enabled devices. For some, it can be the cause of a major headache.

Part of the problem with the leap second is its irregularity. Occurring every two or three years, it means that receiver technology moves on in between — and because the Earth’s slowing rotation is not at a constant rate of change, it cannot be predicted when the next one will be announced. A rapidly developing market of GNSS products having to deal with random alterations to its time framework is not an ideal situation. Suitable preparations, clearly, should be employed.

The behavior of a new receiver when subjected to a leap second may prove critical in certain instances, and without robust characterization it can lead to inconsistent performance. It has already happened this year: on January 21, GPS signals started to include information which effectively announced this year’s leap second event, with the relevant data for future delta time, and week and day numbers. This caused issues with some receivers that weren’t expecting it: some units applied the additional second immediately. It would be interesting to see how these systems might have reacted during an actual leap second transition.

Receiver logic flow requires testing so that any GPS receiver can remain compliant with the IS-GPS-200 standard, and potential problems must  be mitigated and controlled. The use of a GNSS simulator — which outputs a scenario containing the leap second event — allows for the receiver and any systems around it to be exercised over and over again, ironing out any anomalies, to ensure total reliability.

The recent issues with those non-compliant GPS engines highlights the advantage that simulation provides. The consistency it delivers enables a very thorough testing schedule, which will in turn lead to a straightforward application of the time change.

One school of thought holds that leap seconds should be abandoned, and that we should stick to atomic time from now on. Their removal would mean that by 2100, the Earth’s rotation would be some two to three minutes behind humanity’s precise, atomic-powered, 24-hour clock, and half an hour or so by 2700.

The World Radiocommunication Assembly, which has control over such matters, had been postponing a decision on whether to abolish the leap second for over a decade; another vote is due this year. It wouldn’t be any great wonder if this prevarication continues, so whilst it still exists, it is best to concentrate on what this June’s extra second might have in store for anyone currently developing a GNSS product. Armed with a simulator, the unpredictability of leap second scheduling should no longer be a major concern. Should this year’s vote be again inconclusive, those who have taken the positive step of acquiring a GNSS simulator will be in good shape to deal with the next time the clocks show 23:59:60.


Mark Sampson is LabSat product manager for RaceLogic.