Rich Alderson wrote:
From: jim s
Sent: Saturday, January 10, 2009 8:54 PM
The HVD bus is designed such that if you plug it
into a SE bus it will
go into reset and go passive. There is a signal DIFFSNS on the HVD that
will disable the bus. Not good, but better than nothing if you mess up
and plug them together.
I don't think so. The XKL Toad-1 was designed with FASTWIDE differential SCSI,
and any single-ended peripheral that got connected accidentally was guaranteed
to lose the magic smoke and go off to live with <insert $DEITY here>. We were
in complete accordance with the SCSI specs.
We had a SCSI development system at Peer Protocols, as well as wide
sniffer. on the sense side, with the sniffer, we did hadd an electrical
design that could tolerate the misconnection of devices and not catch
fire, though we had a standard adapter to go from differential to SE
that you were supposed to use.
The design used at least with our equipment would be in reset.
I don't know what happend if you plugged something in with the power on,
but if you connected the pile up and it was wrong and then powered
things up, you were covered. Also I don't recall if there was any
issues in the device order.
My observation in the previous post had to do with seeing the reset on
our analyzer when the HD device was connected and wondering why the bus
was held reset. It was a non smoking situation.
I wonder if perhaps the problem was with the fact it was wide and not
the narrow variety of bus? The additional 8 bits and parity were sort
of the bastard child added. on after the fact of the original design.
It was quite a trick to figure out w/o watching the SCSI Domain info
whether all 16 bits were active or just 8 in a give transfer. This
might apply to hardware as well.
Jim