I realize that a SCSI terminator is normally required.  Is there any
reasonable explanation why the scanner would perform flawlessly
without a terminator, but not run reliably WITH a terminator?

Alan


>  In <[log in to unmask]>, on 03/02/00
>     at 09:50 PM, Alan Eugene Davis <[log in to unmask]> said:
>
>  >I connected a Microtek X6EL to the SCSI2 channel of an Adaptec
>  >2940U2W.  The scanner is currently the only device on that card,
>  >and is connected through the DB25 to DB25 cable supplied with the
>  >scanner, through a converter/bracket from a 50pin internal cable.
>
>  >I installed this scanner before I saw the terminator in the box,
>  >without the terminator.  After recompiling the kernel to
>  >recognize SCSI generic devices, I installed relevant packages on
>  >my Debian GNU/Linux system, the SANE-related packages.
>
>  >After running
>
>  >      $ scanimage --list-devices
>
>  >the scanner was identified, and, ever since, there has never been
>  >a need to fiddle---whenever I start the software, either through
>  >The GIMP or via xscanimage itself, the scanner works flawlessly.
>
>  >After I discovered the DB25 SCSI terminator in the packing, I
>  >installed it.  Not only will the scanner not work correctly with
>  >it installed; but the terminator is  quite warm to the touch.
>  >Since I removed the terminator, the scanner has been performing
>  >flawlessly.
>
>  >Can someone explain the seeming anomalous behavior regarding  the
>  >terminator?  I can only  think  that the Adaptek 2940U2W doesn't
>  >need a terminator, while perhaps the Adaptec 2902E/I does need
>  >it.
>
>  >What gives?
>
>  >Alan Davis
>
>
>
>  >--
>  >[log in to unmask]
>
>  Alan:
>  The SCSI bus requires proper termination to operate correctly. The
>  normal configuration is to terminate each end of the bus. Some
>  devices have terminators built in, while others require external
>  terminators. The same is true of the cards, although most have
>  terminators on board. The terminators dissipate considerable
>  power, so warm is ok.
>
>  As a note, some SCSI cards allow both internal and external
>  devices. In this case, the boards termination should be disabled,
>  and the ends of both device chains terminated.
>
>
>  jan lambert
>




--
[log in to unmask]

"An inviscid theory of flow renders the screw useless, but the need
for one non-existent."    ---Lord Raleigh (John William Strutt),
                          or else his son, who was also a scientist.

                Curious about the people moderating your
                   messages? Visit our staff web site:
                     http://nospin.com/pc/staff.html