Interpolation problem between orbit files from merge.py

[sguillot]

Following an email discussion with @paulray:

When calling pint.photonphase from merge.py, for a merged event file with a list of orbit files, there seems to be an interpolation problem with the orbit files. This gives different results than when the PULSE_PHASES are calculated on each individual ObsID event files. This only happens when using merge.py independently on events files that do not have the PULSE_PHASE column. This does not happen when the options '--par parfile.par' and '--merge' are used in psrpipe.py.