We present a new way to solve the platform deformation problem of co-planar interferometers. The platform of a co-planar interferometer can be deformed due to driving forces and gravity. A deformed platform will induce extra components into the geome
tric delay of each baseline, and change the phases of observed visibilities. The reconstructed images will also be diluted due to the errors of the phases. The platform deformations of The Yuan-Tseh Lee Array for Microwave Background Anisotropy (AMiBA) were modelled based on photogrammetry data with about 20 mount pointing positions. We then used the differential optical pointing error between two optical telescopes to fit the model parameters in the entire horizontal coordinate space. With the platform deformation model, we can predict the errors of the geometric phase delays due to platform deformation with given azimuth and elevation of the targets and calibrators. After correcting the phases of the radio point sources in the AMiBA interferometric data, we recover 50% - 70% flux loss due to phase errors. This allows us to restore more than 90% of a source flux. The method outlined in this work is not only applicable to the correction of deformation for other co-planar telescopes but also to single dish telescopes with deformation problems. This work also forms the basis of the upcoming science results of AMiBA-13.
The Y.T. Lee Array for Microwave Background Anisotropy (AMiBA) started scientific operation in early 2007. This work describes the optimization of the system performance for the measurements of the Sunyaev-Zeldovich effect for six massive galaxy clus
ters at redshifts $0.09 - 0.32$. We achieved a point source sensitivity of $63pm 7$ mJy with the seven 0.6m dishes in 1 hour of on-source integration in 2-patch differencing observations. We measured and compensated for the delays between the antennas of our platform-mounted interferometer. Beam switching was used to cancel instrumental instabilities and ground pick up. Total power and phase stability were good on time scales of hours, and the system was shown to integrate down on equivalent timescales of 300 hours per baseline/correlation, or about 10 hours for the entire array. While the broadband correlator leads to good sensitivity, the small number of lags in the correlator resulted in poorly measured bandpass response. We corrected for this by using external calibrators (Jupiter and Saturn). Using Jupiter as the flux standard, we measured the disk brightness temperature of Saturn to be $149^{+5}_{-12}$ K.
