Msg: 6796 *Conference*

09-12-96 18:39:39

From: RON WIESEN

To : RICHARD HANSON

Subj: REPLY TO MSG #6793 (CALLER ID PROJECT)

Got some info out of Bell System literature:
.  Frequency Shift Keying tones are:
   .  1200 Hz Mark  (logic 1)
   .  2200 Hz Space (logic 0)
.  emitted during first silent interval of the ringing cycle.
   .  silent interval persists for 4 seconds.
 
What I didn't find is b.p.s. and character framing info.  But knowing the
bandwidth limit of a phone line, then with these FSK tones it's safe to assume
that the keying rate is no more than 1300 baud, and likely far less.  In the
case of FSK, baud and b.p.s. are equal.
 
Before you get too excited thinking the M100 modem circuit can do the whole
job, note its bandpass filters and compare them to what we need. 
.  Frequency Shift Keying tones expected are:
   .  ANSWER mode:
      .  1270 Hz Mark  (logic 1)
      .  1070 Hz Space (logic 0)
   .  ORIGINATE mode:
      .  2225 Hz Mark  (logic 1) - close enough for 2200 Hz
      .  2025 Hz Space (logic 0)
 
But get this, the M100 cassette firmware "looks for" these FSK tones:
.  2400 Hz Mark  (logic 1)
.  1200 Hz Space (logic 0) - right on the money! note opposite state!
 
Of course the Ring Pulse detect circuit option (Inside the Model 100 book,
pages 266 to 269) must be used to detect onset and cessation of the first ring.
After cessation, the 4 second "silent interval" ensues.  In this case it's not
silence that's expected but rather the "caller ID" character string sent in
1200/2200 Hz FSK tones.
 
The modem in ORIGINATE mode can detect the onset of 2200 Hz (a Space which M100
interprets as a Mark) giving Carrier Detect at port 208 bit 5.  This is of
little use because carrier detect circuits usually have slow response times.
Don't know exactly but I bet the M100 needs to hear carrier tone for more than
1/15 second before it's considered valid.  Likewise cessation of carrier would
not be qulified until at least 1/15 second after carrier ceased.  I doubt Bell
uses less than 15 b.p.s. transmission!
 
It looks like all FSK demodulation must be done at the Cassette input port.
Perhaps some ROM code can be "lifted" or even linked.  I guess that the
cassette read code has a rather wide tolerence for Mark/Space tones to
accomodate a lot of tape speed inconsistancy.  The usability of ROM code
depends on the character framing Bell uses as well as the b.p.s.  transmission
rate.  Probably the two aren't compatible so sepearte machine language code is
needed to demodulate the FSK that Bell uses for call id transmission.
 
Capacitive coupling from the telephone line to the cassette input jack is just
the ticket to block the DC and the low frequency (hi-power) ringing signal,
while passing the FSK audio tones.  Beware the regulatory aspects of connecting
a telephone line to a M100 cassette port.  Simple coupling ain't good enough to
satisfy Federal do-gooders.
 
Cheers,  -= Ron =-