Got loads of responses to this one. Seems that the 'industry' isn't fully
up to steam yet. The majority of the responses mentioned three things to
1) ensure twisted pair is true twisted pair that has the correct polarity,
number of twists per foot, and is truely twisted. Seems that there are
some TP that don't have the required number of twists, use flat cables,
or use analog rather than digital tests.
2) Verify that the transceivers DO NOT run with heartbeat (SQE). In our case
this isn't even an option, but the concentrators are running with SQE.
No one on a followup knew what to make of this 'twist' :-)
3) Check for faults in the equipment. Shorts in the wires, connectors, etc.
Especially with the unshielded wire, it picks up external signals.
Other comments included:
Keep the vendor you purchased from in the loop til things get fixed.
Run routed -q, especially on 3/50's. This was impirically found.
Skip twisted pair and go to fiber (too expensive still).
Go to thinnet (not an option on our campus).
Switching to UTP is a mistake, fire somebody.
Things we're doing:
Any problems with the wiring we're replacing with LAN grade 4 UTP.
Switching many systems at once to UTP, seems to be SOME correlation
between the mixing of thick and UTP.
Slowing down installation on our heavy-use networks until a better
feel for new 'twists' to the problem can be found and corrected.
On the up-side, some of our flakier systems and connection via thick net
are working BETTER with UTP. Just another example of the kind of problem
you will see with the UTP switchover.
I`d post the responses I received, but due to a problem with either Mailtool
or Mail or something else, part of them got clobbered when I was saving them.
I'll send the list to those who explicitly ask for it.
Thanks Net for the help and info.
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:12 CDT