Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.2 9/18/84; site usna.UUCP Path: utzoo!linus!gatech!seismo!usna!tcs From: tcs@usna.UUCP (Terry Slattery) Newsgroups: net.lan,net.dcom Subject: Re: More on Fiber Enet Rptrs Message-ID: <119@usna.UUCP> Date: Sun, 10-Nov-85 21:53:26 EST Article-I.D.: usna.119 Posted: Sun Nov 10 21:53:26 1985 Date-Received: Mon, 11-Nov-85 08:06:05 EST References: <377@nrcvax.UUCP> Organization: U.S. Naval Academy Lines: 32 Xref: linus net.lan:989 net.dcom:1251 From postnews Sun Nov 10 21:47:18 1985 > A short follow up to the net on the article I posted concerning Fiber > Optic Repeaters for Ethernets. > > Another vendor has come on the scene : > Fiber Com The strategies used by Fiber Com differ considerably from those used by Codenoll and Siecor. The FiberCom network is really a ring in which a single uni-directional fiber connects all xcvrs into a ring. The maximum distance between xcvrs is something like 1 or 1.5 km (my literature is at work). If an xcvr fails, a relay couples the input and output fibers. I havn't had the pleasure of opening one up so I can't say how it's done. The Codenoll and Seicor products both use transmissive star couplers at the center of a star configuration. Codenoll's product is totally passive. The optical signals must be within 4db for collisions to be detected so detailed flux budgeting must be done. Maximum arm to arm distance is 1.5Km with either a 4 or 8 port star available. Seicor goes one step further in that they use an active sensor at the star to sense collisions and then broadcast a collision signal (increasing the probablity of sensing a collision.) Fiber Com probably does well in point-to-point stuff and the others in places where several buildings need to be interconnected. I'm in the process of installing the Codenoll stuff here. I should know more in another couple of weeks or so. -tcs Terry Slattery U.S. Naval Academy 301-267-4413 ARPA: tcs@brl-tgr.arpa UUCP: decvax!brl-bmd!usna!tcs