Simtel MSDOS 1992 June/SIMTEL_0692.cdr/msdos/modem/modnoise.arc/MODPROBS.TXT

Text File | 1988-05-01 | 13.0 KB | 232 lines

NOTE CAREFULLY:-

The "call waiting" and "call forwarding" information in this article is
correct for Bell Canada's territory which is primarily the provinces of
Quebec and Ontario in CANADA.  It has not been checked in areas controlled
by other telephone companies.

=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
MODEM COMMUNICATIONS PROBLEMS
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

Written by  Graham Newton
P.O. Box 672
Don Mills, Ontario M3C 2T6
C A N A D A
Occasionally on Compuserve;- 71076,111
Also on Canada Remote Systems


Some people have experienced trouble (typically, being 'dumped' in mid-
transmission) with some systems yet have not had similar problems with others.
The natural thing is to blame the system BUT before you do, read the following
information. You may be surprised that there could be a very simple cure for
your problem:-


The various telephone companies have been offering a useful service called
'Call Waiting' which for practical purposes gives you two phone lines on one
phone.  In other words, if you are on the phone, and someone else calls you,
they hear a 'ring' instead of getting a busy signal.  You (but not the person
you are presently talking to) hear a momentary 'beep' to tell you that
another call is waiting and you can then answer their call by momentarily
depressing the hook-switch which puts the current caller 'on hold' while you
talk to the new caller.  You can switch back and forth between callers also by
momentarily depressinmg the hook-switch each time.

This is a great service, BUT IT HAS A SERIOUS PROBLEM FOR MODEM USERS! That
'beep' you hear to signal that another call is waiting ALSO momentarily
'blacks out' the audio...  and if it is a modem at the other end, yours or the
other end, will think it has lost carrier for that brief duration, and it MAY
DISCONNECT, depending on the terminal software and/or modem types involved!
Fortunately, there is a cure for this problem within another service, 'Call
Forwarding' that the phone company usually offers in a package deal with the
'Call Waiting', and this method DOES work!

NOTE: There is mis-information being circulated to the effect that Call
      Waiting can be disabled by a simple user code *70 or 1170.  This
      comes from a hackers BBS in the USA and may work there, but it does
      DOES NOT WORK here!

This service allows you to 'forward' calls coming in for you, to another phone
number that you can enter whenever you activate the service.  Once forwarded,
your phone rings a 'half ring' once only on every call forwarded to the
number you have entered, just to remind you that your line is 'on forwarding'.
Your line can then be used for outgoing calls WITHOUT causing a 'busy' to the
forwarded calls!

Now the interesting point... once on forwarding, your outgoing calls are no
longer interrupted in any manner, so your modem calls will now, not be
affected.  You must, however, originate the call since incoming calls are now
forwarded to another number AND YOU CAN'T ACCESS THEM.  If you don't have a
second line in the house to forward to, (whereupon, you could still answer the
forwarded calls) you can also 'forward' calls to your own number, which will
then ring 'busy' to anyone calling you WHETHER OR NOT YOU ARE ON THE LINE.
If you are not using the line, it will do a 'half ring' everytime someone
attempts to reach you to remind you that your line is 'on forwarding'.
A simple code entered to the phone line disconnects forwarding and all becomes
normal again.

     Some troubles of the 'I got dumped' variety can be traced to carrier
levels that are too low, either at the sending end (yours!) or the receiving
end (the system you are calling).  Remember that there is a telephone line
between that does have a variable amount of signal loss depending how many
exchanges the call passes before reaching the desired destination.  You don't
have any control over this aspect, but it seems that people who operate from a
Touch Tone equipped line do consistantly better than those who have a simple
'basic black' dial telephone service!  The reason, it appears, is that Touch
Tone lines are more carefully controlled than the dial pulse lines, and hence
they are more consistant.

Another common problem appears in acoustic coupled modems.  Those are the ones
where you stuff the handset into an earmuff type of receptacle after you dial
and hear a carrier being sent to you by the system you have called).
Your telephone mouthpiece can suffer from 'compacted carbon granules' which is
the modem users equivalent of going deaf in both ears!  In a word, the
mouthpiece becomes insensitive to sound and the distant end hears a low level
carrier.  There is a quick and easy cure for this, but unfortunately it only
works with the older model 500 type telephones which use a carbon type of
microphone.  Simply unscrew the mouthpiece and remove the little round capsule
(it will fall out into your hand) and rap it sharply a few times around its
perimeter and replace it.  This will shake loose the carbon granules and make
for better transmission levels.

In some cases, with lower speed modems (110 - 300 baud) it is possible to have
undesirable differences in your modem mark and space tones, thus making the
receiving end modem appear deaf to yours, or your modem putting out a lower
level than is needed, making your operation mariginal, subject to the small
variations in amplitude of one telephone line vs another. In some of the
cheap modems, these are very real problems because the manufacturer can't pay
attention to these matters without raising the price substantially.  This is
partly the reason why good modems cost money.

If you have a frequency counter and audio generator available, you can check
your low speed send and receive frequencies from the following list:-


   ORIGINATE MODE:  MARK  = 1270 Hz
   ==============   SPACE = 1070 Hz
      (YOU)


   ANSWER MODE:     MARK  = 2225 Hz
   ===========      SPACE = 2025 Hz
    (SYSTEM)


You can see that there is only 200 Hz difference between the MARK and SPACE
frequencies and only 755 Hz between the originate and answer filters.  This
means that the filters MUST be accurate and that very little variance can be
tolerated.


Another possible problem, but less likely, is a mis-match of word length, stop
bits and parity check or no parity check in the coding your modem sends to the
system you are calling.  Most systems default to 8N1 or 8 bit words, no parity
and one stop bit to make up each information transfer 'word' that the system
deals with.  If your equipment isn't set to match this format, then you should
adjust your settings to conform.


An insideously subtle problem is the possibility of YOUR LOCATION being
responsible for revealing an apparent fault in the design of your modem, which
the supplier may not know about, or even think about, when trying to duplicate
the fault conditions under testing.  This can be caused by you being in close
proximity to an A.M. broadcast transmitting station (in the case of low
frequencies) and paging, F.M. or T.V. Broadcast or other 'common carrier'
service transmitters (in the case of high frequencies).
By way of an example, in one case a modem user was close (within a few miles)
to a 50,000 watt A.M. radio station transmitter which operates on 1010 kHz.
His modem was constantly acting up at his home, but worked OK at his office
much further away from the transmitter. The tip off was that he could hear
music and voice on his telephone line corresponding to the programs on the
station and, of course, this should not be happening to any self-respecting
telephone.

A complaint call to the Phone Repair Service got a service man to install
surpressors, at no charge, curing the music and spurious talk problem.  The
modem, however, was still acting up, although substantially less than before
the Phone Companys fix was applied.  This now confirmed the fact that the
entire problem was due to 'Rectification' or 'Break Through' of the radio
signal into the audio circuits of the modem.  There are various ways this can
be 'cured' but they all boil down to the necessity of locating the stage where
the rectification is taking place and eliminating it at that point.  There is
usually only one place in the circuit where the problem originates and it most
frequently is a high gain stage like a preamplifier or an active filter.
'Shot-in-the-dark' band-aid fixes usually only reduce the problem, but don't
eliminate it, where proper diagnostic techniques applied will result in a
complete fix for the problem.


More recently, with the advent of higher speed modems that use complex
algorithms to code the data, there are problems with line noise that did not
appear to substantially affect the slower speed modems. There is a lot
of comment regarding 'line noise' and related problems, but there is also a
great deal of mystery surrounding the mechanism of how and where the noise
originates and what the user can do about it, if anything.

Noise will usually take the form of short 'impulse' or clicking noises and
longer crackling or 'static' noises.  The effect of the noise can be directly
seen on your monitor and gives a clue to the type of noise and its probable
origin.  The following quote from the Hayes Smartmodem 1200 Owners Manual
shows what happens when noise corrupts data transmission:-

   "If an error occurs or a data bit is lost at the low speed (300 baud), the
    result on the terminal screen or printer is a single error - either an
    incorrect or missing character.  The same error at the high speed (1200
    baud) is multiplied, however, due to the path that characters take before
    being output for printing or display.  At least two and often three
    characters are incorrect as a result of a single bit error.  The
    descrambling algorithm alters the data so that an error is often produced
    as a left brace '{' or a lower case letter 'i'.  Do not interpret this to
    be a hardware defect."

The impulse noise will often produce the single bit error described above,
however, if the problem encountered is a 'burst' of anywhere from a few to ten
or more characters, you can assume that you are plagued with the 'static'
noise variety.

There is an alternate path selection used by the phone company for local and
long distance call completion.  The equipment tries to complete the call a few
different ways by 'Alternate Routing' through other available central offices
if it finds trunks to the desired end office are busy, before it abandons the
call and gives you a fast (trunk busy) busy signal.  Alternate routing MAY
run your call through an older, noisy office, and hence one possible answer to
why hanging-up and redialing often gets you a better line!

The cause of the noise determines the type of noise, and in the phone company
central office, some impulse noise is produced by older switching equipment.
If your call to a system is routed through one of these older offices, chances
are it may pick up some of this impulse noise.  Natural phenomenon such as
static discharges and lightning contribute to this as well as man made noise
from electrical switching circuits that may induce impulse noise into the
nearby phone line anywhere along the way.  These problems are usually minimal
because the phone lines are balanced to ground and thus tend to cancel this
kind of common-mode noise.

One very likely cause of the 'static' variety of noise is the 'line protector
block' which is installed at every phone line terminating location.  It is
designed to be a surge protector for lightning etc., and it has carbon fuse
elements in it which have been known to become noisy.  The problems caused
are not consistant because the higher ringing voltage and normal switching
transients when you pick up your line will clear the problem temporarily, for
minutes, hours, days or maybe even months!

If you experience the 'static' problem described, then a call to the telephone
repair service is in order.  Simply tell them you are hearing loud static
noises on your line which sometimes go away if you bounce the switch-hook up
and down a few times and you suspect the line protector block is defective
and would like it replaced.  Don't mention modems or computers or they
will start suggesting that you get an expensive data line which isn't needed!
Just make sure that you remove any and all attachments that aren't approved
when the phone company employee comes to do the job, just to avoid any likely
problems with them complaining that you were the cause and not their line.

Come to think of it, are you SURE you, or equipment that you have connected to
the line, are NOT the cause of your problems?  If the noise goes away when you
have all your gadgets disconnected, then one of them is the cause, and not the
phone company supplied equipment.