Page 1 of 1

When to run gpaver

PostPosted: Thu Apr 07, 2016 6:56 pm
by JCollier
In the basic CABB continuum data reduction tutorial that Jamie put together, he suggests running gpaver after having run gpcopy to copy the calibration solutions over to the source. This is suggested so that the full 2 minutes (or equivalent) for which the phase calibrator was scanned is considered when solving for the phase, rather than the 10 seconds worth of data from one integration cycle. In the tutorial, the visibility file corresponding to the source is used as input.

I would have thought you'd want to perform gpaver on the phase calibrator (not the source) and perform it before (not after) copying the calibration solutions over to the source. In my mind, if you've already copied over the solution to the source, running gpaver at this point would not affect the interpolation of the phases. This is based on the assumption that the interpolation happens during running gpcopy. But perhaps the interpolation happens during running uvaver or invert. Is this the case? Does it produce the same result either way?

Re: When to run gpaver

PostPosted: Fri Apr 08, 2016 10:24 am
by Mark.Wieringa
Hi Jordan,

All gpcopy does is copy across the gain table. The interpolation happens when the data is read. So it doesn't really matter when you do the gpaver.
If the weather is not too bad you can set the gpcal solution interval to 30s or 1 minute, in which case gpaver is probably no longer needed.
If you do need 10s solutions, you may also need selfcal on the source to get a decent image.

Cheers,

Mark