At 11:30 PM 8/11/97, Colin Wood wrote:
>Amitai Schlair wrote:
>> Norton 3.5 checks out the Quantum fine -- no problems whatsoever. I'm
>> having some minor difficulties with the SCSI chain, which might be
>> throwing a wrench into the whole thing... namely:
>>
>> 1) Can't boot the HFS partition on the internal Seagate, even when
>> specified with *both* Startup Disk and System Picker, when the external
>> Quantum is on.
>>
>> 2) When booting Mac OS 7.6 from Quantum, the internal HFS partition does
>> not auto-mount like SCSI volumes should. I have to use SCSIProbe to get
>> it mounted.
>>
>> 3) The internal Seagate drive does not spin up at all (!) when I power
>> up the IIci without the Quantum and its terminator on the SCSI bus. The
>> Quantum doesn't have to be on -- it just has to be plugged into the SCSI
>> bus and have a terminator on the end.
>>
>> In fact, the only way I can boot NetBSD as intended (boot internal Mac
>> OS partition, open booter on that partition, boot) is to leave the
>> Quantum connected to the SCSI bus and leave it off. Pretty strange!
>
>This is why people call SCSI a "black art" ;-)
>
>> Should I remove the 3 terminating resistors from the internal drive? Or,
>> is there a 25-pin SCSI terminator somewhere which I can put directly
>> into the Mac's SCSI port? Or is there something else which I am
>> completely missing?
>
>I'm not too sure what you might be missing unless you have a bad internal
>cable somewhere.  Is it possible that you have a bad jumper setting on the
>internal drive?  Check the owner's manual and see what it says about
>jumper settings.
>

I probably missed part of this thread since I've been gone for a while, but
my experience with SCSI says that first you make sure you're following all
the rules:
        no duplicate ID's
        only first and last device terminated (this means the internal
drive and the last external drive)
        make sure the external terminator is powered (if the last external
box is internally terminated then the terminator is probably powered
internally as well which means the SCSI bus is only terminated when that
box is turned on.  I prefer to use external terminators, even though they
may not work as well, because they should get their power via the bus from
the CPU, unless one of the boxes broke the terminator power line.)
        minimize the total cable length
        minimize the number of cable type transitions (Since changing from
50-pin ribbon on the internal drive to the 25-pin case connector to the
50-pin clinch connector on the external case to the 50-pin ribbon inside
the external case are all transitions it helps a *lot* to use boxes that
hold multiple devices.)

When all else fails go buy an active terminator.  I was able to get 5
external devices (7 or 8 ID's in use) to work on a II cx by putting all of
them except the CD-ROM in a single box on a single ribbon cable.  But it
wasn't reliable until I stole an active terminator from an SGI.

I know some people (including me) have been able to get things working by
e.g. putting an extra terminator in the middle.  If you can follow the
rules then things should be more reliable and more resistant to things like
the phase of the moon and whether your neighbor's air conditioning is
running.  My three terminator set-up stopped working when I moved the
computer to a new house.

Signature failed Preliminary Design Review.
Feasibility of a new signature is currently being evaluated.
h.b.hotz@jpl.nasa.gov, or hbhotz@oxy.edu