Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!allegra!mit-eddie!genrad!decvax!decwrl!pyramid!pesnta!epimass!oliveb!glacier!jbn
From: jbn@glacier.UUCP
Newsgroups: net.dcom
Subject: Re: Thin Ethernet really works...
Message-ID: <7117@glacier.ARPA>
Date: Sun, 4-May-86 23:08:43 EDT
Article-I.D.: glacier.7117
Posted: Sun May  4 23:08:43 1986
Date-Received: Tue, 6-May-86 06:58:32 EDT
References: <735@hoptoad.uucp> <136@portal.UUcp>
Organization: Stanford University, IC Laboratory
Lines: 15
Summary: Thin Ethernet; things to watch for


      Thin Ethernet has its uses, but there are a few things to watch
for.  First, of course, just because there's a T-connector at each
station does NOT mean that you can have a section of cable between the
side of the T and the tranceiver; the T connector has to be right next
to the tranceiver.  Any length of cable beyond an inch or two introduces
out-of-phase reflections.
      Second, after extending your net, make sure that the two most distant
stations can still talk.  When you have too much loss in the cable, the
observed symptoms are that some pairs of stations cannot communicate and
that undetected collisions occur, resulting in high retransmit counts.
You're only allowed 8db loss across the whole cable.  If the two most
distant stations can talk, the cable is probably OK.

					John Nagle