1964 Antique MODEM Live Demo

[Eric Smith]

Gordon JC Pearce wrote:
> So, you can't see a possible set of circumstances where having the wrong
> voltage coming in the serial port could cause the frequency shift to be
> wrong?
>   
If you put enough voltage through a serial port signal, you'll break the 
modem.  Then you'll probably get a fixed frequency regardless of the 
input, or no frequency at all.  Other than that, no, having the wrong 
voltage won't cause the frequency shift to be wrong.
> How clever do you think the tone generator side is?
>   
Clever enough that its frequency isn't a linear function of the serial 
port signal voltages.  Modems with EIA-232 (formerly RS-232) interfaces 
detect whether the transmit data signal has a voltage above or below a 
threshold, and generate one of two tones based on that.  They don't feed 
the transmit data signal directly into a VCO, if that's what you were 
thinking.  Doing it that way wouldn't result in a reliable modem, since 
an EIA-232 mark signal can be anywhere from -3V to -25V, and a space 
from +3V to +25V.

For current loop, it senses current rather than voltage, but the 
principle is the same.  It doesn't base the tone on the precise amount 
of current flowing, because even in a loop that has a nominal current 
(i.e., 20mA), the actually current may be significantly different.