From: "Christopher R. Carlen"
Subject: Who thinks this?
Date: Thu, 31 Oct 2002 16:52:38 -0800
Organization: Sandia National Laboratories, Albuquerque, NM USA
NNTP-Posting-Date: Thu, 31 Oct 2002 23:51:30 +0000 (UTC)
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.0.1) Gecko/20020826
X-Accept-Language: en-us, en
I find many people have the following conception:
A transmission line is like a capacitor. If you have a long line, and
feed it a digital logic transition, the output signal at the other end
will look like a RC exponential response. The slowness of that response
will be related, for the most part proportionally, to the length of the
cable and the output impedance of the driver.
Furthermore, that putting a resistor in series with the output of a
driver, and the input of the cable, will somehow slow the response even
more, which seems logical if you think the line is capacitive.
Of course all of these notions are very incorrect, assuming one is
talking about an almost ideal line with a purely real complex
propagation constant, and thus a purely real characteristic impedance,
and a non-dispersing, non-distorting line, such as typical controlled
impedance cables that we use every day.
What do you think?
Fun questions are then derived from these considerations like:
You have a driver generating a very stiff 5V output step, and you
connect it to a 50R line with a 50R series resistor.
Why does the edge at the unterminated output end of the line snap to 5V
after the propagation delay, with the same risetime as if the driver
were driving a simple 50R resistor, no matter how long the line (again
assuming that the line length is such that the line is very close to
ideally non-distorting)? And why then is the input voltage to the line
only 2.5V when the output voltage steps to 5V?
Or doesn't it?
(I'd say it does.)
Christopher R. Carlen
Principal Laser/Optical Technologist
Sandia National Laboratories CA USA