SAS watchout

Why does phasecalc not give me the correct phase reference epoch?

When running the phasecalc task, the user is expected to indicate the date of a reference phase using command line parameters of the form:

     phase=0, epoch="2013-07-05T01:23:45.168"

where the epoch MUST be entered in the UTC time system.

However, the task contains an error which leads to the reference phase being incorrectly treated when the epoch is converted into other time systems.

Workaround:

  1. If the TIMESYS keyword in the event file header = "TT" (the default produced by evselect):
     
    - Subtract the number of leap seconds which have occurred between 1998-01-01T00:00:00 (TT) and the reference epoch, from the time entered as the epoch.

    e.g. in the above example, there have been 4 leap seconds between 1998-01-01T00:00:00 (TT) and 2013-07-05T01:23:45.168 (UTC) and so the parameters:

         phase=0, epoch="2013-07-05T01:23:41.168"

    would give the correct result.

    A list of leap seconds can be found at

    http://www.ietf.org/timezones/data/leap-seconds.list

  2. If the TIMESYS keyword in the event file header = "TBD" (e.g. after Barycen has been run on the event file)
     
    - Convert the reference epoch (from UTC) into TT.

    - Then subtract the number of leap seconds from the reference epoch (in TT) which occurred between 1998-01-01T00:00:00 (TT) and the reference epoch [see 1.].

    e.g. in the example above, you would convert the epoch to TT by adding the appropriate offset for UTC=2013-07-05T01:23:45.168, which is 63.184s and then subtract 4 leap seconds. The final call would then be:

         phase=0, epoch="2013-07-05T01:24:44.352"

Further information on the XMM-Newton time scale and reference time may be found at:

http://xmm-tools.cosmos.esa.int/external/xmm_user_support/documentation/dfhb/timescale.html