Digital Equipment Corporation                    X3T9.2/88-77
    146 Main St.
    Maynard, MA 01754-2571


    +-+-+-+-+-+-+-+
    |d|i|g|i|t|a|l|      M E M O R A N D U M
    +-+-+-+-+-+-+-+

    TO:     X3T9.2 Membership               Date:   7-JUL-1988 
                                            From:   John A. Morse
                                            DTN:      223-5801
                                            Dept:     MSD
                                            Loc/Ms:   MLO5-5/E71
                                            Enet:     THUNDR::MORSE

    SUBJ:   Definition of OPEN pins in SCSI-2 Cables
    REF:    PINOUTPROP.RNO



    I was asked at the June Plenary meeting to justify the definition
    of  "OPEN"  pins  in  the  proposal  I presented for SCSI-2 cable
    pinouts (Document X3T9.2/88-048 Rev. 1).  In that document,  OPEN
    was  defined  by:   "Pins  assigned  OPEN shall present a minimum
    impedance of 10K ohms to either ground or any non-OPEN  pin."  It
    was correctly pointed out that no justification for the selection
    of 10K ohms was given, and  that  there  was  no  assurance  that
    allowing  such  "OPEN"  pins in the cable would not cause induced
    noise and/or cross-talk problems in the cable.

    Ideally, we should  do  full  transmission  line  analysis,  both
    theoretical  and  experimental,  of all the proposed SCSI-2 cable
    and connector designs.  This would give a firm basis  for  either
    accepting   or   rejecting  the  10K  minimum  impedance  figure.
    Although DEC is currently engaged in signal integrity analysis of
    the  existing SCSI single-ended cable, it is beyond the scope and
    timeframe of our work to treat the full range of SCSI-2 cable and
    connector proposals.

    Therefor,  lacking  sufficient  data  to  support  the   proposed
    definition of OPEN, I wish to withdraw that portion of the pinout
    proposal.  In the interests of still preserving  the  possibility
    of  future use of the OPEN pins for TERMPWR, I would propose that
    one of the following 2 definitions of OPEN be adopted:

         1.  Pins assigned OPEN  shall  be  terminated  in  the  same
             manner  as the other signal pins in the cable.  No other
             connection shall be made to an OPEN signal.

         2.  Pins assigned OPEN shall be grounded at  the  points  of
             cable termination.  No other connection shall be made to
             an OPEN signal.

                                                               Page 2


    Either of these alternatives will insure the signal integrity  of
    the cable, since each OPEN signal will behave exactly the same as
    some other existing signal on the cable.  With alternative 1, the
    OPEN  signals  will  behave  the  same as an "Or-tied" signal not
    being driven by any device, while with alternative  2,  the  OPEN
    signals will behave the same as any other GROUND signal.

    Note that each proposal specifies that the OPEN signals  will  be
    attached  only  at  the terminators, and nowhere else.  This will
    permit these signals to be re-assigned to  TERMPWR  at  a  future
    date,  at  which  time only the terminators need be modified, and
    not the devices.

    I submit that adoption of either of these proposals will  achieve
    the  desired  goals  of  retaining  the  existing level of signal
    integrity  on  the  proposed  SCSI-2  cables,  while  effectively
    permitting the re-assignment of the specified pins for additional
    termination power at a future date without requiring re-design of
    SCSI-2  devices.   I  would  like to place these proposals on the
    agenda of the August 15-16 Plenary meeting  at  Colorado  Springs
    for a vote by the full X3T9.2 committee.