+++ATH0
And those of us who remember why this did not work on some modems but others?
Well, it certainly won't work on modems that don't respond to commands
sent over the serial line [1] or modems that have a command set other
than the Hayes one.
It also won't work on true Hayes modems and those that perfectly emulate
them. They require a pause between the +++ and the rest of the comamnd.
This was deliberate so that having such a string in the data stream won't
have any effect on the moden. I seem to remember there was some legal
issue (copyright, patent, something like that) covering this bit of
design which is why some 'Hayes compatible' modems implemented the same
comamnd set without requirign the pause.
[1] Like the one I've spent the last couple of weeks restoring. A Modem
13A and associate Telephone 740. The modem is a PCB in a plinth under a
740 type telephone which is a 746 telephone with the space to mount 4
swtich buttons, here 2 are used to select between the telephone and
modem/ The 746 was the standa UK telephone in th 1970s, think of it as
being the UK equivalent to a Bell 500 set.
Anyway, the mdoem has no spmarts at all. It does contain ICs, 9 Op-amps
whic seem to be 748s and a 74L03 quad NAND gate (here used to XOR the
received signal with a pahse-shifted versio nof it, the mean voltage at
the output of the XOR depends on wheter hte incoming tone is mark or space).
The basic disign of the mdoem is as follows :
Trransmit : TxD line buffered by an Op-amp (who needs 1489s :-)), this
drives som transistors whic change the tap on an LC resonant circuit
which is the 'tank' for an oscialltor built round another op-amp. The
outptu of that goes to a complex LC filter stage, then to another op-amp
which drives the line coupling transformer.
Receive : Incoming signal from line transformer is amplified (another
op-amp), then another LC fiter, then an active filter (you guessed it, a
748). The output of that goes to (a) a charge pump, the output of which
goes to a couple of op-amps to provide the carrier detect signal; (b) to
a limiter, an op-amp which saturates to give a square wave. This sqaure
wave goes to an LC phase shift network, the outptu of that is XORed with
the unshifted version. The output of that goes to an averager (discrete
transitors), then to an op-amp to drive the RxD line (1488s? Who needs
them?).
Power supply : Amazingly this thing can be powered from the telephone
line (although I bleive this was only usable on short lines, it does work
fine on my line simulator). There's an SMPUS built from discrete
transistors (multivibrator, error amplifier, chopper) driving a
transoformer. votlage feedback is taken from the primary winding, not
ideal, but it saves an opto-isolator (which were not common when this
thing was designed). Outptus from the transfomer are rectified and
smoothed, giving +/-5V and +/-8V
FWIW, the modem PCB works fine (for 'fine' meaning 30 characters per
second :-)). The telephone has problems. Some I have put right, like the
broken-off screw in the ringer/ One I made worse (I repaired the case
with Plastic Weld, and managed to glue one of the buttons solif. I will
ahve to make a new button :-(). And one I am still thinking about, a
broken off plastic peg in the rotary dial which means the mainspring is
not achored.
I'll now go back to on-topic stuff (I hope).
-tony
