🚀 go-pugleaf

In article <1agi51INNc58@usenet.INS.CWRU.Edu> at847@cleveland.Freenet.Edu (Jeff Hunsinger) writes:
>
>This has got to be an FAQ, but I just read about 100 subject lines
>and didn't find it. I am trying to hook up a DTMF decoder (SSI202)
>and a voice synthesizer (amplified with an LM386) to a telephone line.
>The DTMF decoder works fine so long as I don't hook up the amp.
>I tried putting an amp before the decoder, but that doesn't seem
>to work either.
>Here's my current set-up:
>
>----Cap--- 600:600 ohm matching transformer-> amp -> ssi-202
>                                           <- Amp -< voice synth.
>

My experience is with a CS8870 DTMF decoder and an LM358 (cheap op amp)
driving the line.  Hooked up like this:

                                      100kohm
line ---\||/---------- protection -*--/\/\/---||--- 8870
        /||\           diodes      |
        \||/                       |
line ---/||\-GND                   +--/\/\/--- amp output
                                      600 ohm

When the voice synthesizer is not talking, the incoming audio on the
line is terminated by the 600 ohm resistor and tapped off by the 8870,
which is capable of dealing with telephone-line-level audio directly.
When the voice synthesizer is talking, there is a loss of power in the
600 ohm resistor but that is not a problem, you need less than one
milliwatt of power!  While the voice synthesizer is talking the 8870
is SWAMPED with audio and will not recognize any DTMF.  Fixing this would
require a hybrid (circuit to subtract outgoing audio from incoming audio)
but such a good one that it's way beyond the scope of this project.  A
simple one (from Motorola telecommunications databook) did not help.
I ended up leaving small pauses in my voice menus so I can interrupt
them.

Potential problems with your circuit:

(1) You are trying to decode DTMF while voice is going onto the line.
    Forget it.

(2) Your amplifier is OVERKILL.  Like I say, one milliwatt of power is
    plenty.  If your amplifier is too big just its idle hiss might
    swamp the DTMF decoder with noise.  Try attenuating the output
    of the LM386.  Also your voice synthesizer might produce too much
    idle hiss.

(3) There are no protection diodes so line transients (like someone else
    hanging up a phone or, heaven forbid, lightning) could blow up your
    circuit.  In mine I have a couple of power diodes, back to back, 
    right across the transformer secondary.  That limits the audio to
    about 0.7V peak which has not been a problem.  If even that is not
    enough, the DTMF decoder is protected by its 100K ohm input resistor
    and the line driver is cheap to replace.

Another problem:  My audio transformer is very small and when I pass the
telephone loop current through it it begins to saturate.  This distorts
the audio slightly.  The voice synthesizer sounds tinny as a result but
it still works.  The only simple fix is to use a bigger transformer or
shunt the loop current past it with a "holding coil".

I would have sent this in e-mail but I can only post from here.

Other experts, please leave the flame throwers on "medium rare".  I didn't
say this circuit is good, and it isn't legal.  It just works.

Markus Wandel
markus@pinetree.org <-- NOT the source of this posting