[daip] Re: SDA in VLA datasets

[Ken Sowinski]

Douglas, et al,

I have some comments to offer at the bottom of this message.

> My request was made as part of an investigation into some errors in the
> data from our observations of the Vela pulsar region in September (the
> data you have). I would have liked to check that the data were observed
> with the correct flags in the 3rd card of the array file (i.e., T in
> columns 13, 15).
> 
> However, perhaps you or others can instead tell me:
> - that the flags must have been set or we would have seem nothing (or all
> the pulsar),
> or alternatively
> - that their state cannot have given rise to our errors
> 
> or perhaps you can shed other light on the situation. So let me describe
> it.
> 
> We made phased-array mode observations of the Vela pulsar with the
> correlator gating on Sep 22, 26, 28, 29 (we phased up on a nearby strong
> source, not the pulsar). On Sep 22 we set the phase center on the pulsar.
> There was a substantial artefact at the phase center. Our working
> hypothesis was that this artifact had something to do with the phased
> array mode (i.e. the zero fringe rate might have meant some DC
> terms/errors were not being subtracted). Perhaps the pulsar gating mode
> was related.

A strong phase center artifact should not occur just because the
array is phased.  Phasing only adjusts the phase for each antenna, 
not the fringe frequency.  More likely, the 'T' on cc15 referenced 
above was not set.  This has the effect of applying a DC offset to 
the correlation coefficients, hence a phase center artifact.

It is possible to read the archive tape, as already discussed, to see
whether the 'T' was set or not.  If not the data taken while the 
correlator was gated is hopeless(*).  If it was set, more thinking is 
required.

Does Tim think that he has to place a pulsar away from the phase
center when gating the correlator?  I suppose not.  That leaves a
missing 'T', or a correlator malfunction such that the wrong V_S
is reported.


Ken

(*)  Perhaps I should write almost hopeless.  I suppose in principle
the visbilities could be restored to correlation coefficents, a 
correction made, and scaling to visibilities re-applied.  A dubious
business, but possible in principle.