I've integrated three night's worth of data synched to a sidereal time to test the interferometer's sensitivity and it looks pretty good. I can get fringes at any time and when avoiding daylight hours, fringes align very well.
Three out of six days had less RFI so I used them for integration to build a FX filterbank complex visibility reference. After cross-correlating this with individual days, I can model the sky well in the evening when Tau A is near the zenith.
Surprisingly, I get fairly strong residuals for the data collected before dawn. A guess is that the moon is moving through my field of view at this time and is close to the galactic plane. This makes it a 'negative' source that might be responsible for what I am detecting. More observation will be needed to explore this.
Modeling against the average sky is leading to very high sensitivity in the system such that more RFI (probably local) and amateur radio signals in the 2 meter band are really popping up.
David