On 2/14/2020 2:00 PM, Paul Koning via cctalk wrote:
On Feb 14, 2020, at 4:54 PM, jim stephens via cctalk <cctalk@classiccmp.org>
wrote:
...
The SCSI spec and cabling have a specific way that the conductors have to be
rolled to make a round cable. Each cable type has a recommended way that
signal and grounds should be paired and in what proximity in the cable.
For SMD I never saw a formal spec with as much detail as the SCSI spec, and I
don't know if they standardized the cabling. Mainly to speculate about whether
you can use a generic 25-25 or 37-37 straight thru.
I suspect the 25-25 would be sensitive to the type of conductor pairing and
fabrication would work.
That reminds me of a situation I ran into about 20 years ago, working for a
small router company. They had a long spool of Ethernet cable to verify
correct operation with max length cables. But things were not working right.
At some point I inspected the connectors and noticed the pairs were connected
to pins 1-2, 3-4, 5-6, 7-8. So two of the Ethernet pairs were split, since the
correct pairing is 1-2, 3-6, 4-5, 7-8. I cut off the connectors and described
the correct way; once new connectors were put on correctly the tests passed.
paul
With all the "cat" foolishness the shysters have gotten in and it's
about like the old nonsense with stereo amps "watts" and the current
"oxygen free copper" cabling for speakers. Hard to trust any advertised
cables and pick ones that actually adhere to any specs. I am not very
good at terminating network cables, and prefer to buy all
pre-terminated, so what you are referring to is very frustrating. And
I've not even started to touch the 10gb networking realm to figure out
how to get optimum cabling and networking set up with that technology.
thanks
Jim
