I have all the scsi devices inside a box except for the TK50
that is in a separate container. So between the disk drives I
don't have any more than 15cm between each other of 50 wire
ribbon cable. Between the CPU and the box I have a 1.5 m cable
(DIGITAL brand name) and between the TK50 and the box about 60cm
(also DIGITAL). The terminator is DIGITAL (passive). The
sequence is:

5000/240 ------> SCSI BOX -------> TK50 -------> TERMINATOR

In my case the cables didn't make any difference. I also tried a
different ordering without change. It was definitively the
loading seen into the SCSI connector of the Chinon CD drive.

>That is a bit like my problems.  I have a box full of scsi cables
>(a dozen or so) that I routinely reach into for hooking up A-B or
>B-C or whatever.  Alas, a pair of them turned out to be funny.
>Get this.....they worked fine with Ultrix, yet died on NetBSD, on
>both pmaxen and VAXen.  I am beginning to think that something
>like cable capacitance or RF resonances are beginning to affect
>the transfer of data.  Are we pushing the data limits for the
>old hardware?  Anyone know?  What are the expected data rates
>on Ultrix 4.2A vs NetBSD?  I would be curious to know the exact
>type of cables you were using, the manufacturer spec, the length,
>terminator type, etc?  Theoretically 6 meters should be usable, but
>I get about 3 meters on old Suns before the signal dies.  My funny
>cables were new, but were no-name brand from the local cable mongers.
>I tried several other no-namers and they worked fine, as did an
>Apple cable, an IBM cable, and a couple DEC cables.  Any insights
>or suggestions are appreciated..... SCSI gets funky sometimes...(:+{{...
>
>.....
>
>> >I also get asc_intr errors on HD accesses occasionally, but they seem to
>> >be more spurious interrupt things.
>
>I am still getting a few of these, even after changing cables, so I still
>suspect some timing things need tweaking a bit.   I will write down the
>exact asc_intr messages tonight.