HWe
2008-03-22
Hi all,
hope you can help me with the following problem:
I want to do a transient simulation with a resistor, which changes its value over time:
e.g. R=abs(cos(2*pi*1000*time)).
This resistance is changing over time, but therefore it needs a time-variable.
From Stefan exists a solution to produce a time-dependent voltage with a current source integrator and a VCVS, but I think this does not help in this case, because then I would need to use this time-dependent voltage in a equation, which is not possible.
I would really be happy to find a solution for this problem, because I am tying to convert some ADS project to Qucs.
Does anyone have a idea/solution ?
How can I use the time as a variable in equations ?
Thnx
Holger
Stefan Jahn
2008-03-25
Hello Holger,
I just uploaded an example schematic, which can be used as sub-circuit. It realizes a cos()-shape time-dependent variable resistor with given amplitude and frequency. It can be found at <http://qucs.sourceforge.net/examples.html> "time_resistor.sch".
Please tell if it was useful...
Cheers, Stefan.
HWe
2008-03-26
Hi Stefan,
thanks for your useful reply.
This example does the job very well, and I will use it for now.
But it really would be much more convenient if there would be a time variable "time" which can used!
Is there a plan to make the "time" accessible as a variable, or are there conceptual obstacles in the qucs-design, which disable the use of a "time" variable ?
Bthw. Qucs-software and this project is a great thing ;-)
Rgds.
Holger
Stefan Jahn
2008-03-27
Hello Holger,
for the EDD equations (or any other) having "time" at hand means that the schematic would only produce useful results for transient simulation. This is point one.
Also it would mean to evaluate the equations on schematic (if allowed there) for each transient time step calculated. This would really slow down simulation.
For an EDD this could be done. But the problem is still: What is the EDD device meant to represent in frequency or quiesent domain with a "time" variable at hand?
As for frequency domain only we already created a new component called RF-EDD where use of "s" (laplace-variable) and "f" are allowed. It will produce useful results for the linear frequency domain simulations. And probably in time domain someday when convolution algorithm is implemented and tested. Anyway we do not allow "s" and "f" in other equations than in the RF-EDD.
So right now we have the two possibilities to get time and frequency into equations without too much implications and side-effects. Anyway I see the request for more flexibility. It is somewhere on my TODO list, but I can't promise anything.
Regards, Stefan.