John,

>> I have two different manufacturer's PCs here that both work differently and
>> both need different levels of negative voltage before they accept an RS-232
>> "Mark" (1 or negative voltage).
>
>That is interesting; are they laptops, or desktop models?  Is the serial
>I/O on the motherboard or on a card?  Are the receiver chips identifiable?

An AST 824 mini-tower and IBM ThinkPad 755 CE.  I haven't opened up the IBM,
and in the AST, the drivers are obscured by a metal frame; I haven't had any
reason to take apart the PC other than this idle curiosity.  Sorry, I'm a
bit curious myself, but there just hasn't been a good enough reason to start
tearing everything apart.

>On one of my PIC projects where I wanted to be able to receive RS-232 but
>also be able to tell whether it was plugged in or not (i.e. so an unplugged
>cable would read as a "break") I wired a 4.7K resistor from RB0 to the input
>and another 4.7K resistor from RB0 to VDD.  Perhaps your PC is doing some-
>thing similar?

I haven't checked the Laptop, but the AST doesn't seem to have any parallel
termination or pullups/pulldowns - at least none that could be measured with
a DMM.

>> If you're designing for yourself and it works with TTL levels, great.  Just
>> don't assume that it will work for everybody.
>
>It's still useful to know how common machines are on which something like
>that will break.  For example, in some cases it can be EXTREMELY useful to
>use the four control wires on a printer port as inputs, even though this
>technique will fail on Toshiba laptops.

Isn't there an "Murphy's Law" that states "Standards aren't"?  Actually in
this case, what I've found by dinking around with other systems, a Sun
Sparcstation 20 will not take anything as a "mark" unless the voltage is
less than 3 Volts.

This whole exercise convinced me that with RS-232, stick with the standards.

myke

"There are only three kinds of economists in the world.  Those who can count
and those who can't." - Eddy George, governor of the Bank of England