[GreenKeys] A reliable current-loop-to-RS232 converter

[Jones, Douglas W]

On Aug 14, 2015, at 4:00 PM, Mike Douglas via GreenKeys wrote:

> On transmit to the TTY, I was worried about the +15v swing allowed by the RS-232 input exceeding the Vbe spec of the PNP transistor, so I used input transistors to hide the RS-232 signal from the PNP transistor that drives the TTY receive loop.
> 
> On receive from the TTY, I used a transistor to buffer the signal to the RS-232 output. This solution drives it directly from the loop.

Back in the 1970's, the last time I designed an RS232 transmitter
receiver from scratch, I used an op-amp as the RS232 receiver, with
one input grounded and the other input tied (through a sufficiently
large resistor) to the data line. Then, I clamped the output of the
op-amp between 0 and +5 (again, with a resistor in series) using a
pair of diodes and fed that into the TTL world.

For an RS232 transmitter, I also used an op-amp to compare a TTL
input with 1.5V (derived by high impedance voltage divider from the
5V supply).  The output of the op-amp was connected to the outside
world through a resistor that dropped about 10V when short circuited
to ground, and the op-amp could handle the current through the
resistor when it was dropping 30V.

Both op-amps were run open loop from the same 15V supply (this was
back in the era when you expected your computer's power supply to
put out +/-15 and +5 on the backplane.

We were usually working at 9600 baud back then, although if I recall
correctly, this particular design was pushing data at 1200 baud from
a Magnavox PLATO IV graphics terminal to a GE Terminet 1200 printer.
The data came to the PLATO terminal in 21-bit packets, of which 18
were delivered to the accessory output at 1200 baud in a wrong-endian
format.  We used software to format those 18 bits as a pair of 7-bit
ASCII characters (bit order reversed), including one start bit and
one stop bit per character in the packet.  Great fun, but it
worked just fine and the users didn't need to know about the
weirdness involved in adding the accessory printer to selected
graphics terminals.

		Doug Jones
		jones at cs.uiowa.edu