Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!uunet!wuarchive!psuvax1!rutgers!mit-eddie!snorkelwacker!spdcc!xylogics!world!jstone
From: jstone@world.std.com (Jeffrey R Stone)
Newsgroups: comp.sys.ibm.pc.misc
Subject: Re: Using a PC to monitor an external signal (How?)
Message-ID: <1990Jul6.015701.22051@world.std.com>
Date: 6 Jul 90 01:57:01 GMT
References: <1990Jul4.215014.29212@mintaka.lcs.mit.edu>
Organization: The World @ Software Tool & Die
Lines: 22

streeter@theory.lcs.mit.edu (Kenneth B. Streeter) writes:


>I would like to use an ibm pc to monitor an external signal.  The
>signal in question is from a single digital wire.  The signal on the
>wire is essentially a square wave, from 0V to 5V, with an irregular
>period.  The input signal frequency would never be above about 150 Hz.
>Essentially, what I want to do is simply count cycles.  (Rising edges
>of the wave, actually.)  It would be nice to be able to count LOTS of
>these pulses (for about 8 hours).

The serial port handshake signals make OK digital I/O ports (2 bits each
direction, per port) if you can arrange your signals to fit meet the 
voltage specs (0 to 12v, I think).  It sounds like you'd need only one
of the inputs.  Monitoring it wouldn't take much CPU power at < 150Hz.
A simple polling loop could maintain a pulse count, and by hooking the 
timer interrupt, you could do whatever count vs. time stuff you need to
do (just keep it short).  A 32 bit int can count to over 2 billion before
overflowing, even with a bit reserved for the sign.  That's 20 million
seconds at 100Hz, or about 8 months.  Should meet your requirements! :-)

-jeff-