2007 October 25
Today I actually looked at my broadband spectra data for the first time in what feels like forever! Got a fair amount done: I generated images of 3c48 in our seven single-pol correlator slices at several sets of frequencies. The difficulty there was mostly flagging the data. Fortunately it seems like the visibilities that need to be flagged can be broken down into two orthogonal cases: something that’s bad at every frequency in a given slice (i.e., bad antennas or baselines), or something that’s bad in every slice at a given frequency (RFI). I worked out a little system that allows me to flag things and check the results pretty quickly and I was able to work through several data sets with a pretty consistently high mojo, where mojo = (speed) * (confidence in results).
So far, all I have is a bunch of images of point sources of varying quality. Tomorrow, I’ll look up a model of 3c48’s emission and actually scale my results by the correct amount; then I’ll look at the antenna gains that I’m finding and see how they vary from one frequency to another. I believe they should be more-or-less constant with some kind of linear trend. There are probably a lot of ways to spend time transferring the self-cal results from one dataset to another to try and improve the results. Once I look at and think about the gains, I may experiment a bit there.
If things go really fast, I might even be able to reduce the data for another source and see how the gains compare from one source to another. Unfortunately Miriad doesn’t provide great tools for manipulating and visualizing gain information, but I think I have the tools to put together the … more tools … that I need without too much difficulty.