Questions about microstrip lines in HFSS
I have many questions as follows.If you are free,please give me advices.
1.
As I simulate a transmission like structure,such as Microstrip line,two solutiion types can be choosed.I run the simulatiion separatey and get different resluts(for Driven modal,the mode number I set up is one).Which one is correct?
2.
The number of modes in Driven Modal is the modes that wll propagate along the sturcture.Therefore,if I set up the no. of modes as one,this means only one mode will propagate.For Driven terminal type,we sholud draw a terminal line.As using Driven terminal type,Does it means that the modes will propage is just one or in fact I inclues many modes(for instance,Microstrip line) ?
3.
As I simulate the microstrip line about Zo using driven terminal type,I plot S parameters and Port terminal Zo.I find that it is not consistent between S-par and Port termnial Zo(ex:the s11 is very low at low frequency and increasing with frequency,but the Zo is approaching 50Ohm with frequency).What is the port termnal Zo in HFSS simulation reslut?Is it different from the characteristic impedance of Microstrip line?
4.
If I just wanna know the characterisitic impedance of Microstrip line,which solution types is suitable?And the same question is that in Driven terminal there is a data "port Zo".Is it the same as the characterisitc impedance of microstrip line?
5.We can deembed the waveport.If the transmission line is 500um,how long is the deembed length suitable.Different deembed length will have different reslut,how to make sure your setup is correct?We can deembed port 1or port 2or port1&port2 same time and it will have different resluts.If I simulate a microstrip line,shoud I deembed the microstrip lline.
5.
If I just wanna know the characterisitic impedance of Microstrip line,should I use radiation boundary or just leave as defalut as pec on the air box?
6.
The solution frequency is the frequency that the machine will do adaptive mesth operatioin.What is the solution frequency I choose as I want to simulate a microstrip line from1GHz to 80GHz?What is different in adaptive mesh from 20Ghz and 40Ghz?
7.
The frequency will simulate up to 80GHz.Should metals solve inside?In the "asign mesth operation"->skin depth based,HFSS can help us to calculate the skin depths.The conductivy unit is mhos/m.What is mhos?after "assign mesh operation",it has two choics,one is "on selection".the other is "inside selection".They all have length based.what is the difference?
8.
I know the waveport using semi-infinite waveguide to excite the wave to the structrue,but I still confuse how the HFSS calculate the fileds or solutions?If the waveport is too small,the filed pattern will wrogn due to waveport that many fields will have more density aroud the waveport edge.If the waveport is large enough,the filed pattern will be correct.But there are still some small fields around the edge,this is not the same in the real world.
9.
To simulate a microstrip line,I make two waveports.May I assign one is input port,and the other is output port?
10.
The final question is that as you simulate a structure not familiar before.How do you make sure your simulation is corrct?Do not say the it comes from the simulation,so it is right.
Thank you very much to reply the post.
Someone free?
please give me advices
I don't get everything, but I'll start to answer this anyway.
I used ports some time ago, when I learned to handle HFSS. The type of port impedance and the value must be chosen correctly. As far as I remenber, the help file or the handbook (pdf) includes a technical note on wave ports. This explaines how to set the ports correctly. Because a port always works in two directions, you should set the impedance equal to the value of the wave impedance (or not, depending on what you want to do).
For the solution every port is solved on its own. The solutions are later superimposed, depending on your source setup. This is used to plot and calculate fields. The S-Matrix includes (as you should know) the solutions for all directions of energy flow, so that it makes no sense to define input and output ports.
If you question the correctness of the simulation results (and you should always do so) you must first of all guess if it may be correct. You can check the convergence and see how the simulation behaved. To be more sure, make more adaptive passes. Experience is of high value for this.
The setup frequency is also carefully to be selected. It should be near the range of interesting changes in the frequency response, but not at the special point. At a peak for example you must not use the peak frequency for the setup, but a value from a lower point on the flank. Using a broad frequency range, I make several setups at different frequencies, and individual sweeps.
Try the following:
Simulate a simple structure, like a straight waveguide, and make several setups with different frequencies but all with the same sweep. Then compare the results, they will not look the same. It looks better if you put a frequency depency to the structure, so that you will find a resonance peak. Good luck!