2007 October 30
Continued to look at broadband spectra data. On Friday, I did the stuff with looking up the 3c48 model and looking at the antenna gains that I was getting out. The gains were usually pretty flat over time, but there were some trends and then some serious outliers that indicated that the self-cal solutions weren’t very good.
Saturday, went observing at Lick for JohnJohn.
Monday, came in late but spent a little time looking at data from 3c84. Most of the effort was improving the flagging situation; now I can update the list of flags to apply to a bunch of different datasets at once. Also worked out some handy utilities to slice through the data in different ways to look for things that ought to be flagged — for instance, many spectra of a single baseline over time, or overlaid spectra at the same frequency to look for RFI. Then I stayed real late in the lab doing nothing productive at all.
Today I actually looked at the 3c84 data and made images. With the automatic flagging, pretty much everything came through OK, although occasionally there are baselines that cause problems for reasons that I can’t figure out. Then came the big test of copying the self-cal solutions from the 3c48 data to the 3c84 data. That didn’t go so well. The solutions aren’t completely off, but the amplitudes I get don’t agree well with my 3c84 model. Unfortunately my 3c84 model is something I made up from two data points from the VLA calibrator database, so its reliability is dubious at best. I searched around online for a while but couldn’t find any nice published spectral models for 3c84 in the 1-10 GHz range. So maybe, just maybe, the copied self-cal solutions actually work OK, but it looks like 3c84 has a very flat spectrum around the ATA band and I’m getting something extremely steep, so I’m not optimistic.
Then more time at lab. A little bit of traction on a revised, more useful OmegaPlot. Hmm.