V2C Notes, Thursday May 23, 2013 ---------------------------------- In attendance: Bill Petrachenko, Arthur Niell, Zinovy Malkin, Dan MacMillan, Johannes Boehm, Gino Tuccari, Mamoru Sekido, and Chris Beaudoin. 1. Broadband system developments and tests (Arthur, Chris, Gino,...) The first ever 24-hour operational broadband session was completed yesterday on the 600-km Westford - GGAO baseline. A single scan (at the end of the session) has been processed. Fringes were found so it is assumed systems were running well throughout. Clocks were well-aligned so fringes were found with clock offsets < 0.1 us. The D-band delay was somewhat different from the A-, B-, C-band delays which is attributed to the use of the hi-band downlink fiber for bands A, B, and C and the lo-band cable for band D. The Field System controlled all sub-systems except the Up/Down Converters and the Analog Level Controllers (ALC's) used to set sampler levels. The sessions was manned but required no manual intervention. There was an average of 48 scans per hour for a 24-hour total greater than 1100 (compared to 300-400 for a typical R1/R4). Scans were 30-s long; data rate was 8 Gbps and minimum SNR per band was 15. With these parameters the minimum correlated flux was 300 mJ which lead to a source list of about 100 sources. A total of 36 Tbytes of data were recorded at each station. The goal of the experiment was to compare correlated fringe amplitudes with predicted amplitudes. Arthur cautioned that a very large amount of data is produced in the broadband mode and processes for efficiently handling and examining the data need to be developed. For this session, four Mk5C's (one per band) were used at each station but it will be possible in the near future to replace these with a single Mk6 at each station. The Mk6 can record continuously at a sustained 8 Gbps rate into a single record module for periods of a few minutes. This needs to be followed by a short down time to refresh the RAM buffer. A data rate slightly less than 8 Gbps can be sustained indefinitely. The D(data)-plane of the Mk6 has been completed while the control-plane and correlator interface still need work. It is hoped that the Mk6 will be ready for field test this summer and available commercially shortly afterwards. The next version of the RDBE, RDBE 3.0, is also expected this summer. The new version includes support for complex sampling, pulse cal detection and 1pps to GPS monitoring. Currently changes to the sampler clock synthesizer are being worked on. The FPGA code seems to work but still no zero baseline fringes have been detected in complex mode. The control and monitor interface to support the new modes still needs to be developed. Gino continues to work on the Digital Broadband Receiver (DBBR). The receiver will be compatible with typical 12-m VGOS antenna optics but adaptable to Noto through the use of a tertiary mirror. The dewar is complete with vacuum ok but work progressing on the RF window. The broadband feed is based on an Italian university/industry collaboration for a different application but adapted to the VGOS frequency band. The feed has been modified somewhat and needs additional changes to the output section. The use of a small (2cm x 2cm) broadband patch antenna on flexible PCB is being investigated for injection of pcal and noise cal into the feed. LNA's from "Low Noise Factory" in Sweden are being considered. Because of strong RFI at Noto below 2.5 GHz a cooled high pass HTSC filter is being considered ahead of the LNA. Gino is now fully involved in developing FPGA code for the DBBC3. A large commercial board with a single Vertex-7 FPGA and a large number of high speed signal connections is being used for the development. 2. RFI (Bill, Chris, Brian) Work continues on the final draft of the EVGA proceedings paper covering preliminary analysis of the V2PEG RFI survey. Based on the anlysis it appears that most stations, especially the new VGOS sites, will not be significantly affected by LNA saturation from RFI. At the same time, thought is being given to an affordable (or perhaps transportable) RFI monitoring system (and related processes) for carrying out a more uniform and robust RFI study. A model based on the La Plata/BKG survey involving month long statistics would be desirable. However, it has been difficult to come up with an RFI monitoring system that will detect interference levels down to a few (or few tens of) K which is required to evaluate RFI degradation in bands that are known not to saturate. In terms of intra-site RFI from DORIS and SLR aircraft avoidance radar, work conitinues (perhaps slower than hoped) on the design and evaluation of physical barriers but some thought is also being given to HTSC filters placed ahead of the LNA - similar to those suggested by Gino. Chris mentioned some work proposed at Haystack for cognitive radio which could be adapted for real time adjustment of channel and band allocation to adapt to a rapidly changing RFI environment. Chris also mentioned that he continues to work on the QRFH cut-off potential for RFI avoidance. 3. Source structure (Richard, Arnaud) - simulations of source position accuracy (Dan) Johannes reported that Stas Shabala and Jamie McCallum have applied their source structure model to simulations using CONT11 schedules. The idea of the source structure model is that, for a simulation, each source position is offset in a random direction in an amount proportional to the source structure index. Then it is seen how much the combined source position offsets effect of station coordinate estimates. For CONT11 schedules it was found that, depending on the station, the degradation varied between 5% and 20%. At the Haystack TOW, a group of participants had a spirited source structure discussion, leading to the suggestion of many possible source structure related projects. Arthur offered to summarize the discussion and distribute to the V2C. 5. Optimum schedules (Jing Sun, John) Dan has been working, along with Ericos Pavlos on SLR/VLBI simulations for networks expected 5 years in the future. Up until recently, the model for mixed (S/X and broadband) observations has been for networks that are roughly equal between S/X and broadband antennas or dominated by S/X. The problem was that the data rate of traditional (non DBE/Mk6-based) S/X observations is low. As a result, fringe detection for mixed S/X and broadband baselines requires long integrations to compensate for the smaller diameter of the VGOS antennas. As a first attempt to solve the problem it was proposed that schedules be divided into time regions using separate S/X and broadband subnets and time regions where all antennas operate together. However, in five years it is expected that 15 or so broadband antennas will be operational. As a result a better network model might be a network with 15 or so broadband antennas and 2-3 S/X antennas. With this model the S/X antennas can be integrated continuously into the schedule with only about 10% degradation in performance for the broadband antennas. 6. Antenna deformations and site ties (Dan) Dan was asked about a site tie error budget he prepared for the NASA Space Geodesy Project. He said he had reported on it at the Paris Site Ties meeting. He said he could distribute the paper to the V2C along with some info on error sources for the other other techniques. Chris reported on a concept he has been promoting (and was also reported at the Paris Site Tie meeting) involving an colocated airborne platform including SLR, VLBI, DORIS and GNSS transponders/receivers all working under a common time base. Conceptually the idea is similar to the underlying concept for the orbital colocation satellite proposals GRASP and micro-g but airborne instead of orbital (and hence more cost effective).