EVLA Project Status

Mark McKinnon and EVLA Project Team

Antennas continue to be converted to the EVLA design. The conversion of the 23rd EVLA antenna was completed in August 2009, although a maximum of 22 EVLA antennas can be used with the VLA correlator. EVLA antennas accounted for 79.7% of all antenna hours used in scientific observations in July 2009.

A pinion gear in one of the two azimuth gearboxes in EVLA antenna 5 failed in April 2009. The EVLA antenna conversion schedule was interrupted to replace the antenna’s gearbox and azimuth bearing. Although the replacement process caused a one month delay, the EVLA antenna conversions will still be complete in Q4 FY2010. A program was implemented to ascertain the extent of this problem in the other antennas. All antennas inspected to date show varying degrees of similar wear on their pinion gears, suggesting that the wear results from the method used to drive the antennas in azimuth. Once the inspections are complete in late 2009, we will develop a plan to repair the gearboxes in a prioritized fashion. Unless the gear wear is widespread and severe, we will make necessary gearbox repairs as part of the regular antenna overhauls beginning in FY2010.

The production and installation of fully EVLA-compliant receivers continues. To date, two L-band (1-2 GHz), four S-band (2-4 GHz), 12 C-band (4-8 GHz), and 13 Ka-band (26-40 GHz) receivers have been installed on antennas in the array. The “interim” versions of the L-band receiver are no longer being deployed. Detailed laboratory tests of an L-band receiver were conducted, and orders were placed for production L-band receiver components. The interim C-band receivers that have been used in the array for some time are being replaced by the fully EVLA-compliant versions. The prototype Ku-band (12-18 GHz) receiver was assembled and tested. The test results indicate that the receiver is exceeding performance specifications. Full production of the receiver should commence in late 2009.

Three different designs for the X-band (8-12 GHz) orthomode transducer (OMT) are being evaluated to optimize performance and to minimize cryogenic cooling requirements for the X-band receiver. A design for a turnstyle junction OMT was developed at the Central Development Laboratory. A planar-style OMT was built and tested in Green Bank. A quad-ridge style OMT with fins projecting from the OMT corners at 45 degree angles from the side walls is being built and tested in Socorro. In early October 2009, the performance characteristics of all three OMTs will be summarized and the final OMT selection will be made so that X-band receiver production can proceed on schedule.

Excellent progress continues with the production of feed horns for the EVLA. Nineteen S-band horns have been fabricated to date. This is on track for achieving the goal of having 20 horns completed by the end of the 2009 fiscal year. The production of Ku-band feed horns is underway at the Green Bank machine shop. The X-band feed horns are being fabricated by commercial machine shops. These horns should be completed well in advance of the X-band receiver production schedule. All other EVLA feed horns are complete.

The data transmission system of the EVLA employs a group of 3-bit, 4Gsps samplers to digitize the 8GHz of bandwidth per polarization from each EVLA antenna. The current sampler design employs a high speed digitizer followed by a de-multiplexer. The digitizer performs satisfactorily. The prototype de-multiplexer also performed satisfactorily; however, subsequent versions of the de-multiplexer have not met EVLA performance specifications. Further work on the current de-multiplexer design has been suspended, and alternative designs are being investigated. These alternatives fall within two general categories: (1) replacement with a de-multiplexer from a different vendor and (2) replication of the de-multiplexer within a fast floating point gate array (FPGA). The FPGA options tend to minimize cost impact and schedule delays. The technical evaluation of these options is underway. During a review in mid-October 2009, the options will be compared and a recommended path forward for the de-multiplexer design will be selected. The de-multiplexer issue will delay the availability of 8GHz observing capability with the EVLA for shared risk observing (SRO), but it is not expected to delay the completion of the EVLA project.

The production and installation of circuit boards for the WIDAR correlator is proceeding on schedule. The pre-production baseline boards were fabricated and successfully tested. After a successful NRC/NRAO production sign-off review on June 15, an order was placed to commence the full production of the final baseline boards. These boards should arrive in Penticton in September 2009, and their installation at the VLA site is scheduled for completion in March 2010. In early August 2009, 60 cross-bar boards and 54 station boards were delivered to the VLA site. The delivery and installation of all station boards should be complete by December 2009. The final correlator will contain 128 station boards, 128 baseline boards, and 64 cross-bar boards. To date, all 64 cross-bar boards, 24 station boards, and 10 baseline boards have been installed in the correlator racks.

Testing of this initial subset of the correlator, called WIDAR-0, continues. WIDAR-0 is currently capable of recording data from 12 EVLA antennas. In late July 2009, data were recorded in all four polarization products for the first time. Recent correlator testing has been complicated somewhat due to the increased complexity of systems integration issues with the additional circuit boards and the distribution of time code between them. We are making the systems integration effort more systematic so that correlator testing can be more efficient.

Software development for WIDAR continues with significant work and testing being done on the correlator configuration mapper. A non-real-time test version of the configuration mapper with a graphical user interface (GUI) was released. A cross-bar board GUI was also released. It allows for manual setting of the distributed cross-bar boards and full polarization (8-bit) testing. For the time being, this GUI will provide the method for establishing station-to-baseline switch connections.

A review of Science Support Systems (SSS) readiness for SRO was held on June 5. The review committee found that SSS is well positioned to support EVLA SRO and did not believe that SSS posed a significant risk to the SRO schedule. The committee also found that the overall design of the SSS applications seemed sound, and the processes used to implement the software were generally good. The committee recommendations included devote more attention to refining the specific SSS requirements for SRO and define software acceptance processes (by non-SSS team members) within the EVLA project.