CST EM-capacitance matrix-Urgent reply please

[hapciu]

hi all,
I am trying to calculate the capacitance between 2 parallel capacitors with CST EM.
imagine that: 2 parallel capacitors with surface=1cm^2 (pec) distance between d=1mm. theoretically, the C=8.85*10^-13F. I made the simulation with cst em, and I got a capacitance matrix: C11 C12
C21 C22.
What that means??? none of these terms equals 8.85*10^-13F.
One more thing, if I change the background (dimensions) this matrix changes .
Can anyone help please, need urgent reply,
thanks all

 

[timof]

hi all,
I am trying to calculate the capacitance between 2 parallel capacitors with CST EM.
imagine that: 2 parallel capacitors with surface=1cm^2 (pec) distance between d=1mm. theoretically, the C=8.85*10^-13F. I made the simulation with cst em, and I got a capacitance matrix: C11 C12
C21 C22.
What that means??? none of these terms equals 8.85*10^-13F.
One more thing, if I change the background (dimensions) this matrix changes .
Can anyone help please, need urgent reply,
thanks all

Capacitance matrix completely characterizes a system of conductors from capacitive coupling viewpoint.

An element of this matrix, Cij, is equal to the change of charge on conductor "j" when you change potential of conductor "i" by 1V. Thus, Cij is a (coupling) capacitance between conductors (nodes, nets, contacts...) "i" and "j". The signs of Cii and Cij may (or should) be different (opposite) according to this definition.

Obviously, Cii is the total capacitance of conductor "i", which is equal to the sum of all coupling capacitances of conductor "i" to all other conductors (including bounding box or ground-plane).

If your capacitance matrix is sensitivive to where you define your bounding box - that means that the bounding box is not defined correctly. The boundaries should be placed far enough so that their exact placement does not affect the simulation results. If the size of your plates is 1x1 cm^2, place the boundary at a distance 10cm or more from the plates.

In your case, C12 should be close to your analytical result. The difference may be caused by the edge or fringe capacitance, which should be of the order of 4*0.1=0.4 cm^2 of the parallel-plate capacitor, i.e. contributing ~40% (a very rough estimate) - so if you want your simulated result to be close to analytical value, decrease the thickness of the dielectric between the plates to 0.1 or 0.01 mm.

 

[hapciu]

hi,
thanks for your answer.
very helpful