Comments by the Space Geodynamics Laboratory (SGL) on the Draft Proposal for a VLBI Standard Interface (VSI) Specification Cannon, W.H., Feil, G., Feir, B., Newby, P., Novikov, S. June 28 1999 1. IRIG-B Time Code ---------------- While it is recognized that VLBI operation requires timing information which has a precision better that a single bit time, it is our opinion that the VLBI Standard Interface (VSI) should provide for only the MINIMUM requirements of an interface with this support capablity. In our opinion the minimal means of providing for VLBI timing is provided by: 1 pps, a fast digital clock, and radio astronomy data streams (sampler output). The advantage of such a minimal definition is its fundamental nature. The timing resolution to the level of a single bit is essentially achieved by counting clock cycles between known epochs. This method obviously requires that the 1 Hz epoch be established without error. Beyond that the communication of time information can be handled in software via the control/status link, and because of the existing 1 Hz signal the software timing precision need only be good to the nearest second, which is easily achievable. We feel that such an arrangement would constitute the minimal interface specification and is consistent with with the philosophical statement #5 presented in the Summary of the Draft Proposal by Nori Kawaguchi et al of June 15 199: ``The VSI specifications do not exclude any additive functions but only defines the MINIMUM functions''. In our view the use of the IRIG-B time code would constitute an extension of the interface specification and would constitute an ``additive function''. In this regard our recommendation would be to adopt the use of IRIG-B as an ``additive function'' only when necessary and otherwise avoid the use of IRIG-B if possible. We would suggest that IRIG-B could be replaced with a standardized software timing interface to address the problem of time code transfer. 2. Parallel Recording and Playback ------------------------------- Referring to the Figures 2 and 3 of the Draft Proposal by Nori Kawaguchi et al of June 15 1999 showing high bit rate data recording by the parallel use of DIB's, it appears to us that the radio astronomy data is entering both DIB's in parallel (as would be expected in this application) from data acquisition system (not shown) while the timing signals are travelling between the DIB's in series. We would recommend that the use of DIB's to accept parallel data streams also require the use of parallel timing signals. That is to say we recommend against the use of ``daisy chained'' timing signals in such an application in which the 1 pps and data clock might acquire unknown and hard to control delay after passing through the first DIB unit and the extra cabling required to reach to the second DIB unit. Instead it is our recommendation that the 1 pps, high speed clock, and data signals be passed through a data distributor box with multiple outputs which serves to insure that there are no relative delays between the radio astronomy data and the VLBI timing signals. For the S3 VLBI data record and playback system we plan to use such a box, called the S3 VLBI Interface Adapter (S3-VIA), to connect multiple S3 systems. It is similar to the S2 VIA (S2 VLBI Interface Adapter) used at many S2 record stations except that it will have multiple outputs to feed more than one recording system (DIB), and multiple inputs to accept data from more than one playback system (DOB). 3. Standardized Control Specification ---------------------------------- We support an initiative to define a standard software control protocol, including physical and message layers. However, since this effort has not begun in earnest and may take some time to gain momentum we suggest that software control be covered in a separate specification document so that it does not delay the development and acceptance of the hardware interface spec.