Lognosh
Newbie level 1
Hi,
This question has asked before:
https://www.edaboard.com/threads/163486/ <-
What I do:
I use the plane wave excitation in CST MWS and then transfer the results to matlab, using the Resultreader.dll. For my project it is important to remove the Plane wave excitation from the grid.
Everything is done in half space at a fixed frequency.
My method now is to place a field probe far enough that the signal of the plane wave is shown in the time Signal, then separate incident and system response, do an FFT and determine the phase of the plane wave.
The information received can be used to determine the phase of the plane wave.
This was the first solution I could think of and it has a poor accuracy, since cutting the time signal is just bad and should be avoided if possible.
The FFT will loose information, except if you have a structure where incident and response are wide apart from each other in time.
I will try to improve this, since I know the time pulse at the probe and may be able to calculate everything with much more precision by only knowing the time delay of the first maxima.
All can be done with the template based post processing.
What I ask for:
I appreciate any input to make this less of a hassle. I was thinking it would maybe be possible to determine the position of the plane wave with the steepness of the phase at the probe.
The plane wave is shown in the simulation as a red plane, how this is done for different grids seems to follow certain rules and I am almost certain this is where the phase of the plane wave is zero.
Does anyone know how CST is doing this?
Also it would be helpful to me, to what the advantage of this mysterious procedure is.
To me it appears so arbitrary, but there should be some thought behind it.
CST MWS is rotating the whole field to meet this requirement and it surely differs from the time signal. Also there is a phase difference to the coordinate origin.
If anyone knows why or has any input to this, I would really appreciate it.
I am also afraid the FDTD is loosing some amplitude through the grid.
Even though everything is made from PEC and no losses are involved. This seems to be some numerical inaccuracy. I am not certain about that, since a good removal of a incident plane wave I have yet to produce.
Can somebody confirm this and is there a way to adjust for that?
Thanks a lot
Lognosh
This question has asked before:
https://www.edaboard.com/threads/163486/ <-
What I do:
I use the plane wave excitation in CST MWS and then transfer the results to matlab, using the Resultreader.dll. For my project it is important to remove the Plane wave excitation from the grid.
Everything is done in half space at a fixed frequency.
My method now is to place a field probe far enough that the signal of the plane wave is shown in the time Signal, then separate incident and system response, do an FFT and determine the phase of the plane wave.
The information received can be used to determine the phase of the plane wave.
This was the first solution I could think of and it has a poor accuracy, since cutting the time signal is just bad and should be avoided if possible.
The FFT will loose information, except if you have a structure where incident and response are wide apart from each other in time.
I will try to improve this, since I know the time pulse at the probe and may be able to calculate everything with much more precision by only knowing the time delay of the first maxima.
All can be done with the template based post processing.
What I ask for:
I appreciate any input to make this less of a hassle. I was thinking it would maybe be possible to determine the position of the plane wave with the steepness of the phase at the probe.
The plane wave is shown in the simulation as a red plane, how this is done for different grids seems to follow certain rules and I am almost certain this is where the phase of the plane wave is zero.
Does anyone know how CST is doing this?
Also it would be helpful to me, to what the advantage of this mysterious procedure is.
To me it appears so arbitrary, but there should be some thought behind it.
CST MWS is rotating the whole field to meet this requirement and it surely differs from the time signal. Also there is a phase difference to the coordinate origin.
If anyone knows why or has any input to this, I would really appreciate it.
I am also afraid the FDTD is loosing some amplitude through the grid.
Even though everything is made from PEC and no losses are involved. This seems to be some numerical inaccuracy. I am not certain about that, since a good removal of a incident plane wave I have yet to produce.
Can somebody confirm this and is there a way to adjust for that?
Thanks a lot
Lognosh