From: roucaries bastien <roucaries.bastien@gm...>  20080722 13:25:26

On 7/21/08, aepalea@... <aepalea@...> wrote: > Just stumbled upon qucs this weekend and I'm quite excited about its > potential in my problem domain. > > I'm working with a water sensing technology which sends pulses with a > subnanosecond rise time through a chamber of coaxial construction > containing various mixtures of fluids as the dielectric medium. The > transmission line I would like to model would have a couple of feet of > 50 ohm feeder cable at either end with a short (less than 2 ft) > chamber in between of highly variable impedance, depending on fluid > conditions. > > The length of our chamber is shorter than most wavelength components > of the injected signal. I was wondering if the sparameter algebra > used in qucs will properly model this situation, or if there is > anything I need to be aware of to properly set this up. Sparam alway works. It is a modelization. Beware only that phase is not numerically null. > As a further elaboration, it would be interesting to simulate this > with the impedance set up as a varying function along the length of > the sensor chamber. I could for example connect together a > transmission line with a dozen 2" segments, each with a difference > impedance. It could work only if segments are << lambda/10 where lambda is the c/f and f is 1/(5*tau)= where tau is the rise time. It is called a quasi static approximation. > Is there any hope that the math used in qucs would produce a > meaningful answer for a transmission line constructed of many > ridiculously small unmatched segments? It work, for instance it is how exponential taper are computed by hand. they are computed using infinitesimal small segment of straight line and taking the limit. Idem for FEM solution they are computed using triangle even in case of smooth part. BTW what is your background in microwave and what is your current position (I am a uni teacher and I do not like to do the homework of the students) Regards Bastien 