Hello real-time people. This is how I test the real-time performance
of my system:
Have the system respond to an interrupt by toggling a pin on the printer
port. Interrupt the system periodically with a square wave. Record the
output pin on the parallel port to the left channel of the sound card on
another machine, and the square wave (interrupt) signal on the right
channel. Then import the stereo audio file into MATLAB and develop the
probability density function of latency. I want to get away from using
MATLAB, but it is very easy to use. (I see Octave on the horizon.)
Note 1: I like to look at the output pin in the above experiment in the
frequency domain using the spectrum analyzer in my lab. The spectrum of
a rectangular wave is very sensitive to the types of jitter that are
introduced by interrupt latency.
Note 2: It has been my observation that the density function of the
interrupt latency is actually multi-modal. I first noticed this in the
time-domain looking at the signals on an oscilloscope. The jitter looks
like multi-path interference on a TV. There are "forbidden regions" for
the transitions.
-kurt
______________________________________________________
Get Your Private, Free Email at http://www.hotmail.com
--- [rtl] ---
To unsubscribe:
echo "unsubscribe rtl" | mail [EMAIL PROTECTED] OR
echo "unsubscribe rtl <Your_email>" | mail [EMAIL PROTECTED]
----
For more information on Real-Time Linux see:
http://www.rtlinux.org/~rtlinux/
