送信者: Alan R. Whitney 宛先: Wayne Cannon ; Brent Carlson ; Dick Ferris ; Dave Graham ; Nori Kawaguchi ; Tetsuro Kondo ; Sergei Pogrebenko ; Misha Popov ; Jon Romney ; Ralph Spencer ; Alan Whitney 件名 : Fwd: Re: Fwd: Re: VSI-H questionnaire (MDR connector) 日時 : 2001年3月21日 1:49 Gentlemen, I forward this note to you from Dick Ferris regarding the issue of the lockdown method for the MDR-80 connectors. Please let me have your comments. Regards, Alan >X-Authentication-Warning: venice.tip.CSIRO.AU: dferris owned process doing -bs >Date: Tue, 20 Mar 2001 20:13:56 +1100 (EST) >From: Dick Ferris >X-X-Sender: >To: "Alan R. Whitney" >Subject: Re: Fwd: Re: VSI-H questionnaire (MDR connector) > >Hi Alan, > >Herewith my comments re Nakajima-san's queries. > > > > > 1. MDR connectors > > > > > > > > 1-a: The VSI-H direct MDR-80 with 4-40 jackscrews. The 4-40jackscrew > > > > is not metric and they are impossible to find here. > > > > Only the 3M in U.S. supplying the connector mating inch screw. > > > > In addition, even they are in metric, the MDR shell with metric > > > > jackscrew is not popular at all. > > > > I remember that I stressed jackscrews lock for the DAS and > > > > DIM which mounted in tilting telescopes. Now I firmly believe > > > > the jack screw should be optional for these special situation. > > > > Latch lock is normal usage for the MDR users. Otherwise many > > > > VSI-H engineers will troubled with the inch/metric, connector > > > > and cable configuration as I had been confused recently. > > > > Fortunately, the VSI-H refer 3-M receptacle N102XX series has > > > > studs for latch. We will use similar one with metric screw hole > > > > in Japan and we will use the latch to lock a connector. > > > > The jackscrew shell will be used in special case. > > > > > > > > I propose you to change the VSI-H to exclude the word 'jack > > > > screw' or change them to 'optional: jack screw lock'. > > > > Please understand that, provisionally Japanese group will use > > > > latch lock receptacle and connectors. We have no other choice. > > > > (BTW, The 3M changed their Web and VSI-H refer URL does not > > > > work.) > >The jackscrews do not engage directly with the socket connector, therefore >they do not need to have the same thread. The connectors are permanently >secured to the panel with 'screwlocks' ('jacksockets'), and the plug's >jackscrews locate into the heads of these. > >Thus connectors with M2.6*.45 panel-mount threads, typically those of >Japanese origin, can be mounted with P/N 3341-13 (3M) jacksockets and >accept plugs with 4-40 jackscrews. Likewise, connectors with 4-40 panel- >mount threads, typically those of American origin, can be mounted with P/N >3341-17 (3M) or P/N 205817-1 (AMP) jacksockets, and also accept plugs >with 4-40 jackscrews as specified in VSI-H. See attachments. > > >All socket connectors have lugs for spring latch locking, and all may be >fitted with the appropriate jacksockets. A plug fitted with jackscrews >will mate with any socket, but would not be secure if jacksockets are not >present. A plug fitted with spring latches will lock to any socket if >jacksockets are absent and might also if AMP jacksockets are fitted, but >probably wont if 3M jacksockets (which are taller) are fitted as they will >foul the outer casing. There is insufficient detail in their respective >'customer drawings' to be definite about this. > >This suggests that the net effect of allowing 'optional jack screw lock' >would result in partitioning VSI-H into two incompatible camps, one using >spring latch only on plugs and sockets, and the other using jackscrews. > >Should we settle for spring latch only? They work well for small >connectors with light, flexible cables, but the bad experience at CRL >reported by Nakajima-san (16 Jul 1999) would suggest not in our case. We >would also not be able to access the manufactured cable assemblies from 3M >-(and Gore?) which only support jackscrew locks on 80 pin formats. > >Change to metric jackscrews? Which 'metric'?: Japanese M2.6*0.45 only >works one way with American M2.5*0.45 threads. And where from?; they are >hard to source from 3M and the only other international supplier of 80-pin >MDRs I have been able to find is AMP (recently gobbled up by Tyco) which >only lists spring latch housings. (Most MDR manufacturers top their range >with 68-pin connectors. While a few also list 100 pin types the >relatively recent 80 pin format seems to be supported only by AMP and 3M.) > >It would appear that we are stuck with 4-40 jackscrew connectors with the >cable connectors coming from 3M, the jacksockets from AMP or 3M as >described above, and straight or right-angle socket connectors from >either. Other than that we need a totally new connector..... > > > > > > > 1-b. The VSI-H direct a 3M foil pleated MDR cable assembly as > > > > an example. Although they expecting 100 ohm impedance, > > > > it seems the specified differential pinout will not locate > > > > neighboring location in the flat cable. > > > > > > > > Our local cable vendor does not used pleated foil type. Cables > > > > which bring into IVS-TOW by Koyama-san will be a round shielded > > > > cable. > >When using 0.025" spaced ribbon coax cables opposite pairs will be >interlaced within the cable. Thus pairs (a+,a-) and (b+,b-) will be >distributed in order a+,b+,a-,b-. This is not a serious problem since >each line is essentially a 50Ohm coax with minimal leakage to its >neighbours. (Even so for maximum length cables at maximum bit rate it may >be desirable to use individually shielded 100Ohm pairs, in Spectrastrip's >Skewclear or Gore's Twinax, in which case the pairs can be terminated in >the obvious manner.) > > > > > >1-c: Frame Ground missing. In the VSI-H pin out there is no > > > Flame Ground (FG) description. This is faulty situation as > > > an electric interface specification. A manufacture engineer > > > claimed us to write. Without the FG, floating voltage between the > > > units will damage LVDS devices when they are connected in worst > case. > > > I remember we discussed the issue in early stage and we agreed to > > > allocate > > > the FG to cable connector chassis/shell. the chassis are connected > > > shielded > > > wire each end. Since most of the pin are already occupied by > optional > > > purpose this is the favorite way to do. > > >This aspect is already covered in VSI-H 12.1.6 Note 1, which is not >perhaps the most obvious place to find it! No other connection is >required. > > >Cheers, >Dick > > > > > > > > >