- April 8 -- the flux for 3C84 has been set at 12.5 Jy, although there is some hint the flux may have been decreasing a bit lately.
- March 28 -- the last dataset (D-13) has been taken.
- April 28 -- I (Marc) have been working on combination of the singledish and interferometric data, mostly with the "Stanimirovic Method" which is coded in $MIS/bin/lincomb.csh. It is very memory intensive because the entire 3167-plane beam is held in memory, so 8GB RAM required. It will also keep one core of my i7-870 3GHz CPU at 100%. The results have been ok, not great. One sees a "mottling" effect in the broad, low level emission in the combined map. See below . This is a result of clean components being found over a large area that have an average separation of larger than the synthesized beam. Is it not enough S/N in the singledish data? I have also played a little bit with immerge and mosmem with similar results.
- It helps to mask the outer edge of the map where the primary beam response goes down. The mask= keyword in lincomb.csh handles this.
- Changing Jy/K (jyperk= keyword) by +/-10%, to modify the factor by which the SD data are added to the INT data, doesn't make a huge difference in the final map.
- For playing around, try just a single plane and maybe just the inner quarter to reduce the turn-around time. We have been using channel 79, anywhere within a few channels of channel 80 has good mix of large and small scale structure.