The WSRT as Tied Array

21 August 1996

Introduction to the use of the Tied Array

The normal use of the WSRT (Westerbork Synthesis Radio Telescope), is to take visibility data to make maps of a piece of sky up to about the size of the primary beam. The visibilities are complex correlations, and the correlation process involves multiplying signals from different dipoles and averaging (typically for 10-120 sec). For more information see The WSRT User Documentation.

The WSRT can also be operated in a tied array mode, i.e. a mode where the signals from the receivers have their delay, fringe and phase compensation applied and are then added together. `Tied Array' is another name for a `phased array'. In the tied array, the data are not averaged in time nor multiplied to form a power level, and the output of the tied array is thus a voltage. These unaveraged data can be used for VLBI (Very Long Baseline Interferometry), Pulsar observations with the Flexible Filter Bank (FFB), and 'long spectra' where there are a very large number of frequency channels.

For our tied array we do the delay, fringe and phase compensation for the telescopes in the tied array with cables and digital circuits. The signals from all telescopes get added in the adding box, but at this point there are 8 separate frequency bands, corresponding to the 8 hardware channels of the Digital Continuum Backend (DCB). The fringe stopping is applied for a frequency in the middle of the DCB bands, which means that each of the 8 channels needs a (small) residual fringe correction. This is applied in the Residual Fringe Demodulation System (RFDS) immediately before adding. For more explanation about the RFDS see Note 472.

Note that any errors that happened at the adding stage cannot be corrected for afterwards, because the original input signals are lost. Any signal to noise ratio lost cannot be regained. At best an overall amplitude scale factor can be applied.

For VLBI purposes the output of the adder needs to be upconverted to a frequency in the range 100-200MHz, and this is done in the Adding Sections Terminal Interface (ASTI). The pulsar backend takes its output from the ASTI input. The ASTI input is also where signals are taken to the spectral line correlator (DXB) to produce spectra with large numbers (>1000) of channels.

For VLBI we also need to combine the signals of different polarization and this is done in the IF/Pol select box. After this the signal enters the Mark 3 Data Acquisition Rack.