微波EDA网,见证研发工程师的成长!
首页 > 研发问答 > 微波和射频技术 > 电磁仿真讨论 > Symetry plane in HF$$9

Symetry plane in HF$$9

时间:03-23 整理:3721RD 点击:
Hi,

How can I set symetry plane in HF$$9? I know that first I have to use Driven Modal, but than I don't know how to set it. Everything I have tried the program showed errors. Maybe someone knows how to do this?

BTW, what is the difference between Driven Modal and Driven Terminal solution. What is the usage of both of them? Which is better?

Greets,
Redi.

Hi,Redi
Set symmetry plan in HF$$9 as the same as in ver8.First you must select the face you want to assign( be sure you are in select face mode not select object mode), the use boundary menu to assign symmetry condition.
Modal driven means the stimulus into device is the EM mode excited on the port ,it's decided by the feed waveguid physical geometry and is the same as real microwave device work condition.
Terminal driven is more relation with circuit network concept, the stimulus is described by voltage and current.If your simulation involve in signal integrity ,it maybe useful to use terminal driven to export a spice compitable netlist for further analysis.

Best regards!

Hi,

About symetry plane.. I want to put a symetry plane in the middle of my structure, so I can't pick the face, which will be that plane. Do I have to make the half of the structure first and then assign a symetry plane?

Greets,
Redi.

hi Redi!

You are correct on the last comment... In HFSS if you set symmetry planes you have to split (chop) your geometry so that the symmetric parts will NOT be present. In CST MWS on the other hand you dont have to do this (the geometry is sliced in half automatically).

The best way to do this is by selecting all your objects and do this:
3D Modeler -> Boolean -> Split

and select the slicing plane :)

Then you have to select the faces and assign the symmetry plane boundaries.

hope this helps!
mogwai.

Hi mogway,

That was I wanted to know. :D

BTW. Isn't that some kind of artificial thing to cut the whole sructure just to make a symetry plane? In C/S/T M/W/S it is really simple and you don't have to even touch the structure. But good to know how to do this also in H/F/S/S. :)

Best regards,
Redi.

No,it's not convenient in HF$$ to set symmetry plan,but it's more powerful in HF$$ when you want to set such plan.You can cut your structure to leave 1/4, 1/8, 1/16,.... as long as your structure has such a symmetry property. HF$$ will know your structure by impedance multiplyer.

Yes. You pointed on very interesting thing, impedance multiplier. Could you explain me how it works? I know that it is used in HF$$ for defining the symetry planes, but how to set it correctly? There are some clues in the help, but only for simple symetries.

Greets,
Redi.

For H-plan symmetry boundary, impedance multiplier should be less than 1, i.e split structure with 1/16 left, the multiplyer should be 1/16=0.0625
For E-plan symmetry boundary,, impedance multiplier should be great than 1, i.e split structure with 1/16 left, the multiplyer should be 16*1=16
Because for H-plan only 1/16 current be included in calculate and E-plan only 1/16 voltage be included in calculate in my example
Best regards!

pewang is right!

Impedance multiplier in the symmetry plane definition will include all multipliers for all symmetries...

That is, if you have a perfect-E symmetry and a perfect-H symmetry so that you split your geometry by a factor of four, then the multiplier will have to be:

0.5 x 2 = 1.0

(0.5 for perfect-H split in half and 2 for perfect-E split in half)...

That's very true...

cheers!
mogwai.

Hi,

For the mixage of E-plane and H-plane symetry, you just have to multiply multipliers, right? When there is one E and H symetry you just put 1?

How to split the structure in 1/32 for instance? Can you for instance split the structure by 2 and put symetry plane and than split just the half that left by the symetry plane in the same direction? As an example of this it could be antenna array of 4 elements in the row. You can split it in the half and then this half split in two, so than only one element left. Is it possible?

Greetings,
Redi.

usually HF$$ use three orthogonal coordinate plane
XY,YZ,XZ as split plan,so you can rotate and offset your coordinate to a appropriate position for splitting.

I wasa trying to use symetry planes, but unfortunately without results. :( So.. I have got further questions. What to do with the port when it is in the middle of symetry plane? Can I use in that case symetry plane or not? If yes, how to do it properly? And, what to do with the structure after spliting, can I throw away one half of it, for instance?

Greetings,
Redi.

I can't understand your quenstion well.But I think you must have not threw away the half.
Thanks!

Hi,

Maybe, I wasn't clear. So.. could you explain me how to put a symetry plane ste by step. Let's say, I have a structure with one symetry plane. Symetry plane is also crossing the port. What should I do?

Thanks,
Redi.

The port multipliers are exactly for the case in which the symmetry plane is slicing the port in half.

The type of symmetry plane (perf H or perf E) depends on the geometry and the modes travelling in the structure. See help in "Perfect E boundaries -> and symmetry planes" for guidelines.

Even if you split your geometry in half, you should get the same results if the symmetries are set correctly (perf E & H and port multipliers).

Take a look on the wg_combiner example in HFSS9... There's a perfect-E symemtry plane on top of the geometry and the impedance multipier is set to 2.

Hope this helps!
mogwai.

Hi,

Thanks for the answer. I can only add a perf H symetry plane, which i slicing port in a half, so multiplier is 0.5. I have splited my structure in two halfs by "split" option, and than what? Should I draw a big rectangle and set it to symetry boundary or just pick the faces in the structure and set it to symetry? How to do this correctly. I haven't seen anything when there was done the symetry and this is important in term of calculation time.

About the example you are saying. I havent found it in my instalation. There is only T junction without saying a word about symetry. I also have the examples which can be downloaded from Elektroda and there is noe such example you are saying about. Could you eventualy send it here or PM me? Example is the best learning material. :)

Greetings,
Redi.

hehe I don't remember where I got the example

It is attached below...

As for you question, you have to pick all the faces that appeared by splitting your geometry. Not drawing a new rectangle.

mogwai.

Thanks for the example. Like I said the best way of learning is the example. Now it works. :D Another thing is choosing the solution setup. But, from the begining. My structure is fed by coaxial cable and in every example I have found there is used Driven Terminal solution. In this solution however you don't have the posibility to use symetry, which cuts the computation time. What is the reason then to use Driven Terminal instead of Driven Modal for ports with the coax cable? The only reason I found to use Driven Terminal solution type is multipin port. I haven't found anything reasonable about differences between this solution types and usage of them. There is said that Modal is working with wave modes and Terminal with voltages and currents, but what are the types of structures to use one or another? What are advantages and disadvantages of both of them?

Greetings,
Redi.

Hi,

Now I have another problem with symetry in HF$$. When I have simulated the structure in normal conditions and then I have made the simulation on only half structure using H symetry surface with multiplier 0.5 and the resonance frequency moved up by around 200 MHz. I don't know what could went wrong. Maybe someone will know what might happend? I am feeding my structure with coax and when I used as outer conductor perfect E boundary there was more or less ok, but when I changed to real conductor with normal thickness it changed a lot. But, maybe there is something with setting symetry for radiation boundary? I don't know. Aby sugestions gladly welcomed.

Greetings,
Redi.

loss metal in HF$$ will shift frequecy from the ideal value.

Copyright © 2017-2020 微波EDA网 版权所有

网站地图

Top