[843] in linux-scsi channel archive
Re: NCR Problems
daemon@ATHENA.MIT.EDU (Michael Weller)
Wed Oct 23 14:36:13 1996
Date: Wed, 23 Oct 1996 20:10:42 +0200 (MESZ)
From: Michael Weller <eowmob@exp-math.uni-essen.de>
To: John Sutton <john@scl.co.uk>
Cc: linux-scsi@vger.rutgers.edu
In-Reply-To: <2.2.16.19961023182640.2e1f71d2@mail.scl.co.uk>
On Wed, 23 Oct 1996, John Sutton wrote:
> >supply terminator power, and should NOT have internal terminators -- that's
>
> You are saying that the box itself should supply termination power? Surely
> if at least one of the units in the box supply it, that should be
> sufficient?
Yep, in theory. However, from my very own experience I learned that
supplying it several times makes no harm and actually improves things. Esp.
for external devices. Thus it is nowadays a recommendation (I heard that
from several places) to supply termination power wherever you can (and
use active terminators).
Well, I don't know enough about high frequency electricity to give an
explanation. Reasons might be the resistance of the TERM-PWR wire in the
external cable and some wire noise caused by the different power supplies in
the host and the external case.
> >the job of the terminator you plug into the open end of the external bus.
> >If you still have these "pass-through" terminators, take your hot glue gun
> >and close the open end. Now. There's absolutely no use for it.
>
> I've seen mention of these so-called "pass-through" terminators but can't
> fathom out what it can mean! Surely it is a contradiction in terms -
> if the bus has been passed through, it should never have been terminated?
Well, yes, however, what is a terminator.. just a pull up (or pull down?)
resistor to ground (or power); at least for a passive one. So, if the end
of the bus is terminated, you actually get a Y on you bus there. A few
millimeters to the resistor and a few millimeters to the input drivers of the
SCSI device (and a bit around within the chips).
So, from the electrical standpoint, you can have such a pass through
connector sitting on the last device doing the termination right before the
input driver chips on the SCSI device. This is perfectly well, as long as the
wires from that connection into the input drivers are really short (about
the same length as the pins of the termination resistor pack on a
conventional SCSI device).
So yes, in general (if you can) avoid these pass through beasts as they
might add additional problems due to the wiring on the scsi device. But
in theory they work. Just because there is no real end on a terminated
SCSI bus, only a very tiny Y.
Michael.
(eowmob@exp-math.uni-essen.de or eowmob@pollux.exp-math.uni-essen.de
Please do not use my vm or de0hrz1a accounts anymore. In case of real
problems reaching me try mat42b@spi.power.uni-essen.de instead.)
