Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
Incorrect capacitance matrix
Posted 2017年1月19日 GMT-5 11:06 Low-Frequency Electromagnetics Version 5.2a 10 Replies
Please login with a confirmed email address before reporting spam
The model is very simple : a block of silicon, with rectangles on the surface. Two of the rectangles are Terminal objects, and the others are Ground objects.
This is the capacitance matrix I get as a result :
es.C11= 4.6233e-13 [F]
es.C12= -4.6233e-13 [F]
es.C21= -4.6168e-13 [F]
es.C22= 4.6168e-13 [F]
As you see, the matrix is not symmetrical : the mutual capacitance between terminals 1 and 2 is different from the mutual capacitance between 2 and 1. Furthermore, the terms C11 and C22 should be equal to the sum of the capacitance to ground and the capacitance to the other terminal, but this is not the case.
How to explain these results? I am inclined to believe that for some reason, the capacitance matrix given has the elements in a different order than the usual Maxwell matrix definition.
Please login with a confirmed email address before reporting spam
please have a look at this blog post
www.comsol.ch/blogs/analyzing-capacitive-touchscreens-in-consumer-electronics/
to see the convention COMSOL is following in capacitance matrix calculation. Note also that in the Derived Values > Global Matrix Evaluation, you can define alternative output transformations, e.g. into the "from Mutaul to Maxwell" format.
Best regards,
Sven Friedel
Please login with a confirmed email address before reporting spam
thank you for the link. I encountered another problem when trying to use a Global Matrix Evaluation : the "Expression" field is not clickable, and I cannot enter anything into it.
I can get each component of the matrix individually by using a Global Evaluation, but I cannot use the Global Matrix Evaluation feature to see the whole matrix and use the "Mutual to Maxwell" transform.
How can I set up a Global Matrix Evaluation for capacitance?
Thank you,
-- Louis Haeberle
Please login with a confirmed email address before reporting spam
Hello Sven,
thank you for the link. I encountered another problem when trying to use a Global Matrix Evaluation : the "Expression" field is not clickable, and I cannot enter anything into it.
Click on "Replace expression" in the Expression section header's toolbar and select the matrix variable in the tree.
Please login with a confirmed email address before reporting spam
make sure to set "Activate terminal sweep" in the Electrostatics interface. If in doubt, you can comparte your model with our touchscreen model. There all necessary settings have been made.
Sven
Please login with a confirmed email address before reporting spam
I have tried to replicate the settings in the touchscreen model, and am able to select es.C in the Global Evaluation "expression" field, but all the elements are NaN. Here are the steps I have taken :
1. Create a parameter named "terminalName" in Global Definitions
2. Activate Terminal Sweep in the Electrostatics node, with sweep parameter "terminalName"
3. Create a Parametric Sweep in the Study 1 node, for the parameter "terminalName" with value list {1,2}
4. Compute the study.
5. Create a Global Matrix Evaluation of the expression es.C
All the elements in the capacitance matrix are now NaN. I have looked at all the settings in the capacitive touchscreen model, and cannot find any relevant settings I failed to duplicate.
Regards,
--Louis Haeberle
Please login with a confirmed email address before reporting spam
Best regards,
Sven
Please login with a confirmed email address before reporting spam
thank you very much for taking the time to build this example model! It has solved my problem.
My mistake was omitting to exclude the terminal domains from the selection in the Electrostatics node.
When the domains are included, the study will compute, and the solution will be the same, but the capacitance matrix cannot be calculated.
Thank you and regards,
-- Louis Haeberle
Please login with a confirmed email address before reporting spam
Sven
Please login with a confirmed email address before reporting spam
I have encountered a situation where I have gotten non-diagonal positive elements (under Maxwell form) and asymmetric capacitance matrix, which is quite weird to me. Do you have some idea why does it happen?
Please login with a confirmed email address before reporting spam
this would be a quite weird behavior, indeed. I could only imagine that your system matrix is very ill-posed for some reason (arising from geometry or meshing) so that numerical errors could accumulate.
Have you contacted support about it?
Best regards,
Sven