Velocity Rest Frame

There are various rest frames which might be appropriate. The following table lists their name, the motion for which one has to correct in order to reduce an observed velocity to that particular rest frame, and the magnitude of the velocity correction. Each subsequent rest frame is obtained by adding the effects of the preceeding ones:

Correct for:	Amplitude (km s$^{-1}$)	Rest frame:
Nothing added	0.0	topocentric
Earth's rotation	$\leq$0.5
Earth's motion around	$\leq$0.013 km s$^{-1}$	geocentric
earth/moon barycenter
Earth's motion around	$\leq$30 km s$^{-1}$	heliocentric(z)
the Sun
Solar motion around the	$\leq$0.012 km s$^{-1}$	barycentric
Solar System barycenter
Solar motion	$\sim$20 km s$^{-1}$	local standard of rest (LSR)
Galactic rotation	$\sim$300 km s$^{-1}$	galactocentric

The most commonly used rest frames are heliocentric (to be precise, barycentric is used at the VLA) and local standard of rest (LSR). LSR is generally used in Galactic astronomy and heliocentric in extragalactic astronomy, although the latter is often reduced to galactocentric.

More often than not, one wishes to specify the velocity of the object and let the on-line system do the conversion to sky frequency for each scan. This is called ``Doppler tracking''. The user will have to specify, in observe, the rest frequency, choice of rest frame, and radial velocity. Doppler tracking is not implemented during a scan as the frequencies are set at the beginning of each scan. If very accurate tracking is required, one is advised to use short scans. Note that the on-line system uses the same algorithm as dopset whereas observe (version 3 and higher) uses a slightly different method and calculates the observing frequency to within a few tens of meters per sec to the values derived using dopset.

Ideally, one wants calibrators observed at the same sky frequency as the sources. This can be achieved by specifying ``no change'' in observe for the flukesynthesizer on the calibrators instead of a velocity. The effect of this is that the LO settings are not changed from what they were during the previous scan. If one wants to start a sequence with a calibrator, it is necessary to precede it with a dummy 1-minute source scan to force the on-line computers to set the LO chain to the proper values. WARNING: if the system crashes and comes back up in the middle of a calibrator scan this scan will be useless because the frequency setting will be in error. The VLA Operator should be alerted to the use of the ``no change'' option, so he can restart an interrupted observation by including again a 1-minute dummy scan. Because of this overhead, it is advised that the use of ``no change'' be restricted to those cases in which it is essential, such as for obtaining high spectral dynamic range; in general, the frequency difference for nearby calibrators is negligable.