Return-Path: <Derek.Beatty@COSMOS.VLSI.CS.CMU.EDU>
Received: from cosmos.vlsi.cs.cmu.edu by K.GP.CS.CMU.EDU id aa13751;
          23 May 90 10:36:45 EDT
Received: from BatMail.robin.v2.10.CUILIB.3.45.SNAP.NOT.LINKED.COSMOS.VLSI.CS.CMU.EDU.pmax.30
          via MS.5.6.COSMOS.VLSI.CS.CMU.EDU.pmax_30;
          Wed, 23 May 90 10:35:01 -0400 (EDT)
Message-ID: <saKdyJ200jccEG0NEt@cs.cmu.edu>
Date: Wed, 23 May 90 10:35:01 -0400 (EDT)
From: Derek.Beatty@CS.CMU.EDU
To: bb-unix-forum@PT.CS.CMU.EDU
Subject: Re: Dirty phone lines
CC: Harry.Bovik@CS.CMU.EDU
In-Reply-To: <10371.643382228@PROOF.ERGO.CS.CMU.EDU>
References: <10371.643382228@PROOF.ERGO.CS.CMU.EDU>

I used to work for a company specializing in instruments to test phone lines
an systems to track repair orders.  Here are some ideas.  As in auto repair,
fixing phone problems is largely a matter of guessing the proper parts to
change.

  1.  Something's wrong with your modem.  Borrow another and try it.

  2.  Your modem or terminal is improperly grounded, and the noise is being
	introduced on a ground line.  This is unlikely, since on the digital
	side of the modem the signal levels are quite robust, and on the
	analog end the noise is common-mode, and easily rejected (though
	common-mode noise could be saturating something).  Be sure that your
	terminal and modem are plugged into the same outlet, and that the
	outlet is properly grounded.  Be sure your RS232 cable between the
	terminal and modem has both the signal and protective grounds connected.

  3.  Something's wrong with your inside wiring.  Unlikely, though a tip-ring
	reversal could concievably act similarly to a grounding problem.
	One way to check this is to disconnect the inside wiring at the
	entrance to the building, and try the modem there.  This can be
	cumbersome.  Borrowing a laptop computer might make this test more
	convenient. 

  4.  Something's wrong with CMU's modems.  But since you see the problem when
	you dial any of several modems, this isn't causing your problem. 

  5.  Something's wrong with the phone company's equipment.  Dial 611.
	(If they visit your home and determine that the problem is in your
	equipment or inside wiring, they'll charge you.)  There is an
	astonishing variety of possible problems.   The most likely is that
	over the winter a seal has deteriorated, allowing water to get
	someplace where it shouldn't be.  Water can corrode contacts, or it
	can simply sit in a cable and raise your line's shunt capacitance.
	But there is a host of other "outside plant" (phone line) problems
	that could be the cause.
      The most subtle problem I can imagine offhand could be lack of clock
	synchronization in some piece of digital switch gear, causing it to
	drop an occasional bit and introduce a phase error.  The ear is quite
	insensitive to phase errors, but (obviously!) phase-shift keying such
	as is used in modems is more sensitive.  But all this is the phone
	company's worry, not yours.  
      I've found Bell of Pennsylvania to be very responsive to my trouble
	reports in MCN's early days, and also when I had real MCN
	problems that looked like my equipment had failed but turned out
	to be their problem.  So if it's their problem they probably can and
	will fix it.  But if they can't, order another phone line.  When it is
	installed and working and noise-free, have the original service
	discontinued, with calls directed to your new number.
