Absolute timing of the Crab pulsar with the Rossi X-ray Timing Explorer

We have monitored the phase of the main X-ray pulse of the Crab pulsar with the Rossi X-Ray Timing Explorer (RXTE) for almost 8 years, since the start of the mission in 1996 January. The absolute time of RXTE's clock is sufficiently accurate to allow this phase to be compared directly with the radio profile. Our monitoring observations of the pulsar took place bi-weekly (during the periods when it was at least 30degrees from the Sun), and we correlated the data with radio timing ephemerides derived from observations made at Jodrell Bank. We have determined the phase of the X-ray main pulse for each observation with a typical error in the individual data points of 50 mus. The total ensemble is consistent with a phase that is constant over the monitoring period, with the X-ray pulse leading the radio pulse by 0.0102 +/- 0.0012 periods in phase, or 344 +/- 40 mus in time. The error estimate is dominated by a systematic error of 40 ms in the radio data, arising from uncertainties in the variable amount of pulse delay due to interstellar scattering and instrumental calibration. The statistical error is 0.00015 periods, or 5 mus. The separation of the main pulse and interpulse appears to be unchanging at timescales of a year or less, with an average value of 0.4001 +/- 0.0002 periods. There is no apparent variation in these values with energy over the 2-30 keV range. The lag between the radio and X-ray pulses may be constant in phase (i.e., rotational in nature) or constant in time (i.e., due to a path-length difference). We are not (yet) able to distinguish between these two interpretations.