On Tuesday, September 28, 1999 1:52 PM, John Refling 
[SMTP:johnr@imageworks.com] wrote:
>
> A break has nothing to do with the DTR line.  A break is formed
> by holding the TX line inactive for a second or two.  Zero voltage
> is not inactive, so pulling the cable out of the port doesn't
> send a break.  The inactive signal is a voltage between 3-20 volts
> or so (and I can't remember if it is positive or negative).  The
> active signal is the same voltage, but opposite polarity.

Many RS232 drivers consider near ground (0.8 Volts) and below to be 
an inactive (I forget if that's marking or space) and greater than 
about 2 Volts to be an active state.  The spec (RS232C) defines [-25, 
-3] Volts and [3, 25] Volts as the valid signal levels.  I seem to 
recall RS232D dropping this to [-15, -3] Volts and [3, 15] Volts.

If the cable is pulled out, the signal may very well go to ground ( 0 
Volts).  This is NOT a valid signal level and could be seen as a 
break condition depending on where the RS232 drivers consider the 
dividing line between active and inactive to be located.  Some RS232 
receivers have a fail-safe mechanism to force the logic output to a 
known state when the cable is removed but not all RS232 receivers are 
required to have this fail-safe mechanism.

If an RS232 data transmit signal is not actively transmitting data, 
it will be held at the lower potential, usually around -10V.  If I 
had to build a system that would survive a PC console reboot, I would 
probably tie a 25 K Ohm resistor between the NetBSD box's data 
receive pin (pin 3 on DB25, pin 2 on DB9) and some signal driven by 
the NetBSD box to a -10V level.  I don't recall what signals are 
driven by the NetBSD box but $1.00 US worth of parts and a 
multi-meter should be enough to solve the problem.

Most of this is from memory, if you have any questions you can send 
them to me at steven_grunza@ieee.org