TEMPO2: The world’s most powerful
pulsar timing package

George Hobbs (ATNF)

Figure 1: Using TEMPO2 to view timing residuals dominated by a simulated gravitational wave background (click on image for larger version).

The technique known as “pulsar timing” has already led to numerous important scientific results. For instance, the discovery and timing of the first binary pulsar system, PSR B1913+16, and the testing of Einstein’s general theory of relativity led to the award of the 1993 Nobel Prize in physics to R Hulse and J Taylor. During the year 1990, Alex Wolszczan was studying the pulse arrival times from PSR B1257+12 and managed to deduce the existence of the first planetary companions to a star detected outside of our solar system. More recently, by timing the double pulsar PSR J0737-3039A/B, the Parkes pulsar multibeam survey team obtained the most stringent tests to date of the theory of general relativity (Kramer et al. 2006). The basic pulsar timing idea is very simple. We create a model of the pulsar and then compare the actual pulse arrival times with the times expected from this model. Any discrepancies imply that there is a problem with our model, which can subsequently be improved. The accuracy with which this technique allows us to measure pulsar parameters is outstanding. For instance, the rotational period of PSR J0437-4715 was 0.005757451840111090 seconds on 14 January 2006, with an uncertainty of 4 in the last quoted digit.

Pulsar timing predictions are required for almost all pulsar observations made worldwide. For instance, the new ATNF digital filterbank folds the data from the telescope on-line and requires knowledge of the pulsar phase and pulse-period at the time of the observation. Huge increases in signal-to-noise ratio are obtained for VLBI pulsar observations by recording data only when a pulse is expected to arrive. Again, this technique requires pulsar timing software in order to enable such predictions.

New high-precision pulsar timing experiments, such as the Parkes Pulsar Timing Array (PPTA) project, are pushing the standard pulsar timing code (TEMPO) beyond its limit. This older program cannot predict arrival times accurately enough for our requirements for the on-line hardware, nor model all the phenomena that affect the pulse signal at the levels required. TEMPO is also limited in that the algorithms implemented are not compatible with International Astronomical Union (IAU) 2000 resolutions and it can only analyse one pulsar at a time. In order to detect gravitational waves we need a more powerful and updated software package.

We have now produced a new package, unimaginatively called TEMPO2. This is the most powerful pulsar timing software available and takes into account all known issues that can affect the pulsar signal at the 1-ns level. This involves, for example, correcting for the delay that the signal experiences as it passes by the limb of Jupiter, polar motion, propagation effects in the Earth’s atmosphere and the solar wind. TEMPO2 can analyse multiple pulsars simultaneously allowing global signatures that are common between pulsars to be detected, thereby making pulsar timing array experiments realisable. TEMPO2 uses International Celestial Reference System (ICRS) coordinates and is fully compliant with recent IAU resolutions that require updated precession and nutation models. The software also includes numerous visualisation and analysis tools.

TEMPO2 has already been used to place the most stringent limits to date on the existence of predicted gravitational-wave backgrounds (Jenet et al. 2006), detect non-Kolmogorov turbulence in the interstellar medium (You et al. 2007) and is now being used on-line for our PPTA observations. Already adopted internationally, TEMPO2 will supersede all existing pulsar timing packages and provide the accuracy needed for pulsar timing into and beyond the SKA era. More details on this easy-to-use software can be found in our papers (e.g., Hobbs, Edwards & Manchester 2006) and from our web-site: www.atnf.csiro.au/research/pulsar/tempo2.


Hobbs, Edwards & Manchester 2006, MNRAS, 369, 655

Jenet et al. 2006, ApJ, 653, 1571

Kramer et al. 2006, Science, 314, 97

You et al. 2007, MNRAS, in press