FastFieldSolvers Forum
FastFieldSolvers Forum
Home | Profile | Register | Active Topics | Members | Search | FAQ
 All Forums
 FastFieldSolvers
 FasterCap and FastCap2
 Dielectric Interface def. for sandwiched structure

Note: You must be registered in order to post a reply.
To register, click here. Registration is FREE!

Screensize:
UserName:
Password:
Antispam question: What do MOONwalk and MOONdance have in common?
Answer:
Format Mode:
Format: BoldItalicizedUnderlineStrikethrough Align LeftCenteredAlign Right Horizontal Rule Insert HyperlinkInsert EmailInsert Image Insert CodeInsert QuoteInsert List
   
Message:

* HTML is OFF
* Forum Code is ON

 
   

T O P I C    R E V I E W
gpandit Posted - Aug 11 2021 : 18:56:48
Hello,

I have a special case of a metal disc(radius=6mm), some dielectric (er = 2.7, width=1.3mm) and two half circular plates (radius=6mm, not connected electrically). The plates and discs are in air. Since the plates and disc are circular, I have approximated them by using triangular sections for the top and bottom surfaces and quadrilaterals for the sides.

I have following questions when I am trying to prepare input files for this stack-up:

1. Should I use two dielectric interfaces sandwiched between the plates and the disc? If yes, what should be the width of the dielectric interface & reference point defined in the files?
Cross sectional description:
Conductor(plates)|Dielec(1,2.7)|Dielec(2.7,1)|Conductor(disc)

2. When I try to use FastCap2/FasterCap with 2 dielectric interfaces, I get following errors on iteration #4 (gmres iterations never end):
Capacitance matrix is:
Dimension 3 x 3
g1_myshape 5.71731e-013 -8.11308e-014 -9.60827e-014
g2_myshape2 -9.09785e-014 5.633e-013 -1.02333e-013
g3_myshape3 -3.98736e-013 -3.98984e-013 4.91374e-013
Warning: capacitance matrix is not diagonally dominant due to row 3

Also, the symmetry is not maintained in the output capacitance matrix. What can I do to ensure I get the correct capacitance results for this particular case?

Here's a pastebin link to the input geometries and txt file with conductor and dielectric definitions:
Input Geometries: ht*ps://pastebin.com/6q5EsNHF
test.txt: ht*ps://pastebin.com/KiWNJP0s

3. If the above mentioned stack-up is incorrect, what should the the plates and disc with sandwiched dielectric be defined as?

Thanks in advance!

P.S.: Please excuse me for not using the right terms/definitions, as this is my first time using FastCap.

Gaurav
1   L A T E S T    R E P L I E S    (Newest First)
Enrico Posted - Aug 17 2021 : 17:01:02
Dear Gaurav,

there are multiple issues with your geometry and the most important one is that you have overlapping panels.

You should NEVER overlap any panel in FastCap/FasterCap. E.g. your dielectric cubes are dispaced one on top of the other, so the surface on the top of one cube is completely overlapping the surface on the bottom of the other. The same happens with the electrodes.

In fact, seems that your electrodes are not in air, but touching the dielectric on one side, i.e. they are partially in air and partially in the dielectric medium. In this case you need to cut the surface of the dielectric corresponding to the electrode (same shape) and you need to build the electrode by two parts: one touching the air (with permittivity 1.0) and one touching the dielectric (with permittivity 2.7), connected together by the '+' statement at the end of the line in the 'test.txt' list file.

You may reference some of the sample files showing capacitors in mixed dielectric medium as a template to build your input file.

Other less critical issues you may want to fix is the usage of very thin triangles in the discretization. Thin triangles cannot be refined into better behaved triangles. You should make a constrained discretization via some software; e.g. FreeCAD may help you.

Finally consider that if you use FastCap, you are responsible of the correct level of refinement. The level of refinement you are using in your input files is too coarse for getting meaningful results.
FasterCap on the other hand can refine the structure for you; in this case, FasterCap is happy to receive a discretization as coarse as possible as it prefers to handle the refinement in a way to minimize some error measure.

Best Regards,
Enrico


FastFieldSolvers Forum © 2020 FastFieldSolvers S.R.L. Go To Top Of Page
Powered By: Snitz Forums 2000 Version 3.4.06