送信者: Alan R. Whitney 宛先: Cc: Wayne Cannon ; Brent Carlson ; Dick Ferris ; Dave Graham ; Nori Kawaguchi ; Tetsuro Kondo ; Sergei Pogrebenko ; Jon Romney ; Ralph Spencer ; Alan Whitney ; 件名 : Answers to questions 日時 : 1999年6月26日 2:50 Dear Nakajima-san, I am very sorry for such a late answer to your questions, but did not realize until recently that they had been asked. So, let me try to answer as best as I can: >After the intensive discussion in TDC-CRL Japan this week. >We all know that your first VSI idea is basis to study furthur. >But it is a little bit confusing since the idea is abstract >actual existing cables problem. >Here I exclude discussion with optical transmission in future. >I just ask you the following question based on my action items related. > > >1. Will we make standard cables for the VSI or not? > VSI will define a interface. Between the interface, we need cables in any case. > In this frequency, the cable which maintain digital signal shape is the key > component for inter compatibility. > Neither connector or pin assignment, the cable is important. > I would like to see a standard connector and pin assignment for VSI if people can agree. Based on the extensive work that you have done in Japan, the D-sub 50 appears to be a good candidate. We have had good experience with the Mini-D ribbon connectors from 3M. We are using an 80-pin version at 53 MHz with 8m shielded flat cable with no problems. They have sturdy screwdown locks and appear to be well made. 50, 68 and 100-pin versions are also available. I do not know cost offhand. >2. Will the VSI define the cable length ? > Without consideration, The maximum length for 64MHz clock and data will > keep it shapes less than 10m(33ft). > From my experience, we will see many troubles around 10m. > This is not long enough, at large correlator site or > in sampler mounted inside telescope. > It seems to me that 10m should be long enough for almost all applications. Normally, the second downconversion of the RF to BBC can take place at some distance from a telescope and be within 10m of the DIB. As for correlators, perhaps the DOB output signals should be converted (outside the DOB) to high-speed optical or coax serial lines, which can be used over many tens of meters. This is the method we use in the Mark IV correlator and it seems to work quite well, where longest distance is of order 20m with coax [fiber could be considerably longer]. >3. Should we exclude 128MHz CLK in this discussion? > Since the ECL100K which handle the high frequency > is different technology and expensive. After my study, I know this will not be > unified currently used ECL10K,10KH. The clock step in the cable should be > maintained under 64MHz. > I think the complications of 128 MHz clock are probably so much that it is not worthwhile to consider them for this purpose. >4. Is the VSI define a future system compatibility or current system modification ? > For example, will the Mk-III or S-2 system will have a VSI interface ? > Do we have a target correlator or target experiment to perform compatibility > in near future ? > The goal of the VSI specification is to promote the use of this specification in new VLBI equipment, and to suggest retrofit of existing designs if possible. Currently, there is no target experiment to perform compatibility, but is definitely something we should think about doing. Others may wish to comment on these subjects, as well. Regards, Alan