Path: utzoo!news-server.csri.toronto.edu!cs.utexas.edu!uunet!aria!sanford
From: sanford@aria.ascend.com (Curtis Sanford)
Newsgroups: comp.dcom.sys.cisco
Subject: Re: Appliques can fail ... and other V.35 problems
Message-ID: <415@aria.ascend.com>
Date: 4 Mar 91 16:39:06 GMT
References: <32744@boulder.Colorado.EDU> <1991Mar3.225534.6358@bwdls61.bnr.ca>
Organization: Ascend Communications -- San Francisco
Lines: 21

If you ever unearth your original messages, I'd be interested in them.  I am
working on a problem using a DSU that can vary the clocking from 56kbit/s
to 3Mbit/s, depending on the amount of dial-up bandwidth obtained.  So,
simple clock inversions are unlikely to work over the whole range of speeds.
I'm looking for a solution.

The symptom I'm seeing is at high speeds (500K-3M) I will occasionally lock
up the cisco's, to the point where even pings do not go across between the
nominally 'up' links.  They remain in this condition even after I drop the
bandwidth to 56kbit/s.  The only way out is to reload the routers.  I'm
looking for a solution.

I'm using V.35 interfaces, with about 20' of cable.  I've put a scope on the
transmit data lines in the DSU (transmit from the cisco) and the data seems to
be within spec with the clock pulses.  I'm really interested if anyone is
familiar with this lock-up state of the ciscos.  Even if there is a clocking
problem, it should clear up when the speed is lowered back down.  So, there is
a cisco bug lurking here as well.


I'm using V.35 interfaces, with about 20' of