On Jun 24 20:01 96, Rod Speed of 3:711/934.2 wrote:

DH>> Sigh.  I've now got another episode of the one-way-only disconnect
DH>> blues, so I've gone to work on trying to identify the trigger mechanism.

DH>> My theory re V8 appears to be full of the proverbial - I managed to
DH>> produce a solid case of the fault under the USR v32Terbo protocol. No
DH>> V8, no calling tones, no fancy negotiations.  ??????  What the _blazes_...

RS> Do you actually have any evidence for anything other than just a
RS> USR problem, albeit one that is only seen in SOME circumstances ?

Yes - the Hayes vFC and V34 both produce the symptom, although with
different degrees of incidence.   Eventually, so did the Microcom - it
appears to be about 3 orders of magnitude better, not immune as I first
thought.

I gradually come to the conclusion that the information available is
nowhere near enough to even partially indicate what the bloody problem is.

DH>> Scratch one theory, IMHO.   Wonder what to try next?

RS> You could try sacrificing some chickens |-)

DH>> Same shitstorm is still on - I now have a Microcom Deskporte Fast
DH>> on the same line, with an occasional retry.  (Two retries so far, in
DH>> 1.4Mb.) Holds dead solid in VFC, at 26.4k.   Yet in
"normal" conditions,
DH>> the USR craps all over the Microcom for line holding ability and speed.

RS> Sure, but all that shows is that the USRs are sensitive to something.

Yeah - but it's intransitive.  Other than "The" fault, the USR is
far superior on bad lines.

DH>> Now if I just knew what "normal" meant, and
DH>> what the #{at}#{at}# is sour at the moment...:-(

RS> Are you still getting the audible ticking in that fault situation ?

Too hard to reproduce at the moment - I have to have control of two lines,
and shut down the modem signals on both, at the critical time, AND have the
line I can hear be the direction it happens in.   Three cases heard now,
over several months worth.  Enough to be "I think so" - but not
enough to jump up and down and say "I know for sure".

DH>> Neither USR I have is capable of holding that line at the moment
DH>> for longer than 2 minutes maximum.  _However_ both negotiate happily
DH>> up to 28.8, and report s/n at over 35db, until the fault hits.

RS> Corse you dont actually KNOW that the S/N is valid in the fault
RS> situation, just that thats what the USR claims is the S/N.

Point taken.  The USR has gone gaga.  Some of the readout figures are
simply impossible - for instance, the far echo goes to several thousand db
of loss on some of the logs.   It is apparent there's interference - and
that the interference affects only one direction.  That's about the
totality of accurate information.

DH>> The fault is in severe mode, and as soon as
DH>> it's initiated, lock into retrain, and disconnect.

DH>>I think I'm gonna take up golf. Easier, cheaper, and more likely 
DH>>profits..:-

RS> Life wasnt meant to be easy boy |-)

Yeah, but this?  :-)

 Dave Hatch

--- Msgedsq 3.20