ZIP drive on SCSI segment w/ other devices -- problem or not?

[Mike Isely]

For a month or so I ran a SCSI Insider ZIP with a Quantum Fireball ST
4.3GB drive, hosted from a 2940UW.  The Fireball was running at Ultra
(20MHz) speeds and terminating the bus (ZIP drive was in the middle).  The
kernel version was 2.0.33 with a much older (and buggier) aic7xxx version,
though I can't remember off the top of my head which version it was. 

I never experienced any "damage".  I did experience a lot of SCSI timeouts
followed by resets, but in all cases, the driver was able to clean up and
continue, with no data loss.  The timeouts were "relatively" rare but they
did happen with regularity, while accessing the Quantum drive.  If I just
unplugged the ZIP drive from the bus and left everything else exactly the
same, all the timeouts went away.  Disabling Ultra speeds for the Quantum
drive didn't affect the problem. 

I also tried the above system but with a SCSI external ZIP drive via a
DB-25 to micro-D50 cable plus a micro-D50 to micro-D68 adapter instead of
the insider ZIP drive.  Of course, the ZIP drive terminated one end of the
bus (and the controller still terminated the high byte).  This
configuration was completely unusable.  The system would boot, but the
error (timeout) rate was far too high. 

It's possible the above problems may have been related to the version of
the aic7xxx driver, though I fail to see why disconnecting an otherwise
idle ZIP drive would have an effect in that case.

I've since moved the SCSI Insider ZIP drive to another system with a 3940
(dual channel, non-ultra) controller.  The ZIP drive now shares one bus
with a Nakamichi MJ-8S 4-disc CD changer and an old Tandberg SCSI QIC tape
drive.  (The two Quantum Fireball hard drives on this system share the
other bus).  I'm running kernel 2.0.35 + aic7xxx 5.1.12.  Works perfectly. 


On Tue, 1 Dec 1998, James W. Durkin wrote:

> I've been warned off of placing a SCSI Insider ZIP drive on the same
> segment (the 50-pin narrow connector of an Adaptec 2940U2W PCI card)
> as other devices such as CD-ROMs.  The expected problems have been
> variously described as "will damage the other devices on the segment",
> or "will cause the controller to reset indefinitely".  Nothing
> particularly specific, and nothing I've been able to confirm from
> looking through the various unofficial Iomega sites on the web.
> 
> Instead, it's been suggested, that it should be driven from the
> supplied (Adaptec sourced, I guess) controller.  Seems a waste of an
> ISA slot, plus cabling hassles, etc.  But if there is a legit reason
> for it, I guess it's tolerable.
> 
> So, truth or fiction here?
> 
> FWIW, I'm no fan of ZIP drives.  They're sort of a sneaker-net
> standard here, so I live with them.  I would prefer to live with them
> in the most efficient manner possible though.
> 
> If it matters, it's a PII system with all other storage devices (save
> floppy) hung off the 2940U2W, running RH 5.2.
> 
> -- 
> James W. Durkin
> Program of Computer Graphics
> Cornell University
> jwd at graphics.cornell.edu
> 
> 
> To Unsubscribe: send mail to majordomo at FreeBSD.org
> with "unsubscribe freebsd-aic7xxx" in the body of the message
> 

                        |         Mike Isely          |     PGP fingerprint
    POSITIVELY NO       |                             | 03 54 43 4D 75 E5 CC 92
 UNSOLICITED JUNK MAIL! |   isely @ pobox (dot) com   | 71 16 01 E2 B5 F5 C1 E8
                        |   (spam-foiling  address)   |



To Unsubscribe: send mail to majordomo at FreeBSD.org
with "unsubscribe freebsd-aic7xxx" in the body of the message