[1649] in linux-scsi channel archive
Re: SCSI Cables and aha2940
daemon@ATHENA.MIT.EDU (Andy Poling)
Mon Mar 31 23:57:12 1997
Date: Mon, 31 Mar 1997 23:54:07 -0500 (EST)
From: Andy Poling <andy@realbig.com>
To: Dave Wreski <tel1dvw@is.ups.com>
cc: "James T. Kirk" <root@plantel.com>, linux-scsi@vger.rutgers.edu
In-Reply-To: <Pine.LNX.3.95.970331213135.6048A-100000@tony.plantel.com>
On Mon, 31 Mar 1997, James T. Kirk wrote:
> Well again it's probably a case of "It's not supposed to work"
> when you put all connectors into use, it creates a fragmented SCSI chain,
> with three terminations. SCSI is supposed to be a linear model with two
> ends. Using all three make a fork in the chain and can result in echo and
> other errors. Like I said, it's not supposed to work, and I wouldn't
> trust it for real world data.
>
> Morgan Nelson
> Plan Telecommunications
>
> On Mon, 31 Mar 1997, Dave Wreski wrote:
>
> >
> > On Mon, 31 Mar 1997, James T. Kirk wrote:
> > >
> > > any two of the connectors may be used, but not all three. The way the
> > > card is designed, using all three breaks SCSI termination rules and will
> > > not work.
The documentation that comes with the controller says explicitly not to use
all three connectors. RTFM m'boy...
As (ahem) Captain Kirk mentions, it is also contra the SCSI spec, which also
expressly prohibits any `T' fitting of any sort - especially one as long as
any cable would be.
It's probably worth mentioning that using both the internal and external
connectors (even in any legal combination of two) is likely to put you at
risk of exceeding a legal length (which exact value escapes me at the
moment). Especially with that long cable that Iomega supplies with the Jaz
drive (in the grand tradition of suppliers of PC peripherals - "what, you
have other peripherals too?!"). At the least, a Forced Perfect Termination
terminator is advised...
-Andy
Global Auctions
