Xref: utzoo sci.electronics:4460 comp.dcom.modems:3064 comp.dcom.lans:2089
Path: utzoo!utgpu!watmath!clyde!bellcore!rutgers!gatech!ncar!boulder!sunybcs!bingvaxu!leah!itsgw!steinmetz!uunet!wucs1!kumar
From: kumar@wucs1.wustl.edu (Arun Kumar)
Newsgroups: sci.electronics,comp.dcom.modems,comp.dcom.lans
Subject: trouble with TAXI
Keywords: TAXI, FDDI, fiber optics
Message-ID: <612@wucs1.wustl.edu>
Date: 9 Dec 88 21:57:38 GMT
Distribution: na
Organization: Washington University, St. Louis, MO
Lines: 24


	I would like to get in touch with people who are working
to implement a communication link with the TAXI chip-set from AMD.

	I have built a transmitter/receiver pair, using TAXI's, that
attempts to transmit a 512*512*8 picture from one frame buffer to
another over an optical fiber. For electro-optic and opto-electric
conversion I use the ODL200 chip set from AT&T. The circuit used to
control the TAXI is built with FAST chips from Fairchild. The FAST
logic is wire-wrapped, yet it appears to work as expected. The fast
ECL connections between the TAXI's and the ODL200's are on a plain
vector board (no ground plane), but are short as short can be. I will
soon be putting everything on a 4-layer PCB. I seem to lose about 
1 pixel per hundred at transmission speeds of about 70 million bps.
I would like to run at the FDDI speed of 125 mbps.

	The TAXI chip set I have is not the long-promised final-
version from AMD, and I suspect that therein lies all the problem.
Is there anyone out there working with TAXI's, or implementing ANSI's
FDDI standard, who can help or sympathize? 
---------------------------------------------------------------------
Arun Kumar					kumar@wucs1.wustl.edu
Box 1045, Computer Science				(314)889-6160
Washington University, St. Louis, MO 63130