Maximum Time Span Window for Measurements

Question

Accepted Answer

There are two separate limits to be aware of when measuring long time spans: how long a measurement can run before its timestamps roll over, and how large a time difference you can measure accurately. These are unrelated limits, roughly eleven orders of magnitude apart, so it is worth keeping them distinct.

Maximum Representable Time Span (Rollover)

Time tags are represented as 64-bit integers in picoseconds, so the maximum representable time window of a measurement is approximately 107 days in both directions. This does not limit how long you can continuously acquire data with a Time Tagger; acquisitions can run indefinitely, for months or years. The rollover only affects the raw timestamp value itself, and is handled transparently by the software except in the rare case where a single measurement’s time difference calculation spans exactly across the rollover instant.

How do the time tags look after 107 days?

The time tags are integer numbers in picoseconds relative to the creation of the Time Tagger object. After 107 days, the time tags jump to negative numbers (-107 days) and increase until they reach 0, at which point the counter restarts.

Practical Accuracy Limits Over Long Time Spans

Separately from the rollover limit above, measuring a time difference accurately over a long interval requires a correspondingly accurate and stable clock signal.

The internal clock of our Time Tagger 20 is sufficiently stable to achieve jitter-limited measurements of time intervals up to about 100 ns, whereas with our Time Tagger Ultra and Time Tagger X, the internal clock is sufficiently stable to achieve measurements with a time span of up to 1-10 ms. Our Time Taggers can certainly measure larger time intervals, but users should expect increased time uncertainty beyond these referenced limits. For precise measurements over very long time intervals, users can introduce an external clock source.