Minutes US-VLBI/EVN telecon 2nd April 2009 ========================================== 'Present': Alan Whitney, Jon Romney, Chet Ruszczak, Steve Durand, Walter Brisken, Craig Walker, Adam Deller, Arpad Szomoru, Walter Alef, richard Porcas, Dave Graham. VLBA wideband recording status ============================== Spending freeze means that further 'roach' hardware for VDBE cannot be ordered. Firmware development for VDBE continues. Conduant has delivered 'daughter boards' for data input to mk5C, also firmware which has been installed. Testing is under way at Haystack, using a PC-based test fixture. Upgrading VLBA recording hardware will take a few days per station. Plans call for gradual installation (of the DBE and Mark 5C simultaneously, if possible) across the Array over a period of about a month, parallel to the existing operational data path. Once this process is complete, a nearly instantaneous transition making the new equipment operational should be possible within a week. The new correlator will be brought into operation before this transition. The 'roach' hardware can support 'VDBE' firmware (for line observations), or Haystack 'DBE2' firmware, typically for continuum. Haystack DBE2/Mark5C development ========================= Mark5C with daughter board and firmware is in testing. No CRCC check is foreseen in current firmware. Tests of 10GE transfer from DBE2 to Mark5C are being done. Zero-baseline tests of as many types of digital backend as possible are planned: DBE2/dBBC, Chinese CBAS, possibly a Japanese unit. Compatibility, out-of-band rejection and other characteristics are to be examined. After units are returned, real-world VLBI measurements between them will be carried out. During the meeting there was some theoretical discussion of out-of-band suppression and necessary number of filter coefficients, the Haystack practical end-to-end tests were judged to be a good way to check this out. European dBBC development ========================== See presentation from Wettzell meeting: http://www.wettzell.ifag.de/veranstaltungen/vlbi/frff2009/DBE_DBBC/DBBC%20status%20Mar09.pdf The unit implements 4BBCs per LX220 FPGA, currently driving Mark5B but a 10GE board is under development. A polyphase filter personality is also available but is as yet untested in the field. Two dBBCs are installed at Wettzell and have been used for tests. One unit is installed at Effelsberg, for tests when all the signals are connected up. Several other units are being constructed, although the 'spinoff' construction company is not yet set up. Several Mark5Cs are at Bonn, as yet without 'daughter' boards. It is expected that, after the company is set up and if cash flow starts, remaining EVN stations could be upgraded. They would record as Mark5B stations until the dBBC 'Fila10G' is complete. This ensures compatibility with JIVE requirements. VLBA Software correlator ======================== The operator GUI is close to readiness for operational use. The changeover to the new correlator is 'ready to happen'. Special test observations are planned to check operational readiness. The correlator must be in operation before recording hardware can be switched. Data formats currently supported are: VLBA and Mark 4 format on Mark 5A; Mark 5B up to 1 Gbps; and LBA. Others (including VDIF) may be added eventually. JIVE correlator =============== Mark5A and MarkB formats currently supported. Mark5B for eVLBI is still being perfected. VDIF format not supported, although some type of on-the-fly format translation -may- be possible. Compatibility issues ==================== Mark5C recordings can only be replayed as Mark5B format if the appropriate 'packetizing' firmware is installed in digital backends. This is a requirement to allow global observations to be correlated at JIVE. Coupling the NRAO VDBE firmware with the Haystack packetizing code is expected to be straightforward, and is included in current plans. A meeting of specialists from NRAO and Haystack was proposed to address the issue. A further compatibility issue is HSA (high-sensitivity array) operation: GBT will get new hardware, also VLA (for the moment without phased array). Effelsberg would like to obtain a 4-IF VDBE unit, sharing Mark5Cs with dBBC via a switch. NRAO will provide pricing information. Arecibo plans are unknown. Phase Cal and noise Calibration ================================ These are discussed together since the proposal has been made to calibrate observations by looking at the strength of pcal tones. Not many of those present believed this would work at the highest frequencies, and it must be turned off in spectral line observations, so it cannot totally replace the synchronously-detected broadband noise cal. Phase cal with 1MHz rail spacing is not useful above 22GHz. Changing the spacing to 5MHz (already available on VLBA) would help, also a new Haystack pcal unit: http://www.wettzell.ifag.de/veranstaltungen/vlbi/frff2009/Part5/BeaudoinFRFFPcal.pdf gives better signals at high frequency. However 5MHz spacing is not good with 8-16-32MHz subband spacing, unless interpolation is used, and brings the danger of ambiguity. In the new digital systems, pcal extraction will be available both at stations (for VDBE) and at the correlator. The dBBC offers 80Hz calibration as in current VLBA equipment. The synchronous detection in the VDBE should be able to deal with a variety of switching frequencies.