The WIDAR Correlator

 

WIDAR

The WIDAR correlator is inherently a spectral line correlator in any regular mode. A full description of the current WIDAR capabilities is provided in the respective section of the Observational Status Summary. The OSS also contains a spectral line configurations section.

There are two important issues when configuring the WIDAR correlator for spectral line observations. One is set the necessary spectral resolution. This can be achieved by baselineboard stacking and/or recirculation (the latter is under commissioning). Both are described in the OSS. For large bandwidths with high spectral resolution one can either use a number of normal subbands and stack them next to each other. A much better option, however, is to use a single, wide subband, up to 128MHz, and use baselineboard stacking (and/or recirculation) to obtain a high number of channels. This avoids the "stitching" process and provides a much better spectral baseline.

A second issue is the existence of the 128MHz boundaries. Lines should not be placed across or very near these boundaries since subbands cannot span across a boundary and the sensitivity drops near the boundaries. In particular note that the very center of the baseband always falls on a 128MHz boundary. The spectral line under consideration should never be placed in the very center of a baseband. Multi-line observations also need to ensure that none of the lines fall on or near a boundary. This can be challenging at times but is usually a solvable problem and the OPT provides some tools to do so. If it is not possible to obtain simultaneous coverage of all of your lines, or if the exact position of the line is unknown (e.g. for line searches), it is possible to observe with two basebands shifted by 10-64MHz apart. This will ensure that one baseband covers the boundaries of the other baseband with full sensitivity. An example is given in the figures, where the top figure shows the rms of a single baseband. The 128MHz boundaries stick out as having high noise. The bottom figure is a combination of two basebands where they have been separated by 64MHz. The noise spikes are clearly suppresses by adding (with the appropriate weight) the two basebands or even by simply replacing the noisy channels of each baseband with data from the other.

 

blankFieldRMS.AC.png

 

 

blankFieldRMS.interlace.png

 

Subband 0

The baseband shape can be very low at on one side of the spectrum. This causes bandpass functions to be close to zero in the first 20% of subband zero in this baseband and is mostly observed in subband 0 in that baseband. The affected baseband edge is at the lowest sky frequency in the baseband when using upper sideband, and at the highest sky frequency in the baseband when using lower sideband. This part of the spectrum should be avoided for spectral line observing. See EVLA memo 154 for details. NRAO staff, however, is constantly looking into improving this behavior.

 

Data Rate Limits

Baselineboard stacking, recirculation, and time resolution can add up to an extremely high data rate in the correlator. Please see the Observational Status Summary for the allowable data rates and volumes for each observing semester. The OPT instrument configuration calculates data rates based on the spectral line setup and the maxima for data rate and volume must not be exceeded for any observational setup.