openEMS.de

Have you tried the MSL port? I must admit I didn't pay much attention to the port setup. There should be an example on the tutorials page of the wiki showing how to use the MSL. I'm away from my computer for a few days so won't be able to check it until I'm back next week.

Here is an example showing how to set up a simulation of a parabolic reflector antenna being illuminated with a simple feed over the Ku satcom band. To make it more realistic you can change the design of the feed and add support struts. The aim of this example is to show how to set up such a simulat...

Try the attached file. I got the mesh smoothing to work somewhat. There are more lines in between the gaps and the edges are properly represented. When you run it, it will bring up a number of warnings but let it run. The mesh still requires further improvement but I think it's a step forward. Perha...

I cut and paste the 3 lines defining the port box directly after the three lines with the SmoothMeshLines functions for mesh.x, y and z. Immediately after that and before the DefineRectGrid, I insert the line: mesh = DetectEdges(CSX, mesh); That will add mesh lines to the edges of the model features...

I checked the mesh and I can see that the mesh lines do not coincide with the edges of the metal or the etched area of the split rings. Also the mesh lines do not lie on the front of the port box. If you check the wiki for the FDTD mesh setup, there are some guidelines that discuss the 1/3 2/3 rule ...

I've not had chance to check it yet, but can you check that the port extents lie directly on mesh lines?

Also make sure there are mesh lines on the edges of the CSRR and perhaps one or two lines in the gap.

Thanks, Janis. I was an EagleCAD user for a long time and was considering to buy the full version until I realised that you can only have it as a rental unless you stay with the free version. As a result I will be making the transition to KiCAD, so I'm sure this will be useful for me in the future. ...

I've just installed more memory in my computer and it's now quad channel and it seems to have made quite a difference. Using the same, large simulation with two large data dump planes, I did a number of tests increasing the number of cpu cores. I find a setting of 7 cores gives the greatest MC/s for...

After making the reflector a solid object the farfields are much more like what I'd expect.

I also noticed, in the line you pasted, there is -wg.length+wg.length which is just 0! I was afraid that the parabola as a thin sheet may not mesh properly so I will make it a solid and check the PEC dump. I'll also add field monitors to view the fields over the whole structure. The reflection back ...