> is Ok, except the RS-232 readback is really tricky,
>> ( CTS is never < -3V what it should be to RS-232 specs)
>> but well it seems to work that way too.
>
>Yes, I know this problem.
>The PC's serial port is known for detecting levels below 1 V
>as LOW and above 1.25 V as HIGH.
>
>That's the reason why this circuit works and why it won't on some
>of the 'not so compatible' PC's.

One solution (as long as we are in clever hacker mode here) on serial ports with
"TRUE"
RS-232C / V.34 drivers would be to put a fairly large capacitor accross a
resistor in series with the serial line.

That would cause the data to swing to (up to) +5V on a 0-1 transition and -5 V
on a 1-0. On a
port with hysteresis, this might work.

NB: I haven't tried this, it's just a thought.
-Don