Continue to Site

Welcome to

Welcome to our site! is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

how to define finite substrate in IE3D

Not open for further replies.


Full Member level 2
Jun 18, 2002
Reaction score
Trophy points
Activity points

I am wondering if I can define finite substrate in IE3D. I find the default substrate size is too big. Thank you.


Hi, danda821:

1. Define some dielectric type(s) in Basic Parameters.
2. Draw the shape of the finite dielectrics as a group of polygons.
3. Select them and select ADV EDIT->DEFINE DIELECTRICS CALL.
4. Choose the dielectric shapes and choose whether you want to remove the polygons you use for defining the shapes. If you choose to remove it, the polygons will be gone and they are used to define the finite dielectric shapes only.

Best regards,

hi jian,

i have a doubt in any structure in default has infinite substrate in IE3D?...

Then if we reduce the substrate thickness to nm range say 500nm(0.5micrometre)......what should be the strip thickness and width?......whther is there any optimum range for simulation in IE3D?...will it work in FIDELITY ?.......

Please specify the range ......expecting ur reply...

with regards,

Hi, mailgvenkat: On IE3D, finite size thin dielectrics can be modeled quite easily. It should not have any problem in handling dimensions in nm range. In fact, such a thin dielectrics should cause problem in FIDELITY which is based on FDTD. FDTD algorithms normally have difficulty in modeling structures with electrically smally dimensions and structures with large aspect ratio structures. Thanks!

hi jian,

i have simulated a stripline structure with nanometer dimensions as:
strip width -60nm
strip thickness -10nm
substrate thickness(bottom)-50nm
substrate thickness(top)-40nm

On simulating this i have the s-parameters and transmission line parameters as shown in attachments...
i infered that the impedance is quit large exceeding there any constaint or limit in this value of impedance?........

give me the ideal values os s-parameters(s11,s22,s12,s21).....

also i could not plot the near field pattern for this structure...could u please help me in plotting the same...i read the manual but on following it iam getting certain errors...

so please specify me the steps for plotting the near field pattern......

with regards,

Hi, Mailgvenkat: I downloaded your resultnm.doc file and open it on MS-Word. It shows empty window even the file size is 800+ bytes. What is the file format or is it corrupted? Regards.

hi jian,

sorry for the trouble....herewith i attach again the new file for your reference.....
please have a look at it.. and help me for the above questions in previous reply.....

anticipating your help....with regards,

Hi, Mailgavenkat: I checked your .doc file. Following are the comments:

1. I noticed you have defined 3 substrates (50 nm, 10 nm and 30 nm). However, your strip is defined on z = 245 microns. It is very far away from the ground. It will be a very high impedance TLN no matter what.

2. You use Fmax = 20 GHz and Ncells = 200 cells per wavelength. It is very coarse for such a tiny structure. I tried to increase it to Fmax = 20000 and Ncell = 30 to mesh a length of 1 micron into a few cells in the longitidual direction. When the meshing is too coarse, it may cause accuracy problem.

3. The strip thickness is 10 nm. For such a thin strip, the loss is very high. It is talking about 2 ohms/square. You will find the Zc to be very big even for a normal dimensions which you will expect the Zc to be around 50-ohms.

The thickness of the strip is no longer very small compared to the width. It is possible building the thickness model may improve accuracy.

I have attached my models in this posting.

Best regards,

hi jian...

thankyou for the valuable and useful i have defined 3 substrates in my case.....but for a stripline it will normally have 2 i have to remove the middle one...i.e.10nm substrate......

If i remove it and specify a length of 500nm for the entire structure...will there be any improvement in char.impedance and the loss.......also i have to increse the Fmax to around 20000GHz(is it not a large value...will it be a practical value) and reduce ncells to 30 cell/wavelength.......

also i have to analyse for near field pattern .....but for such tiny substrate how to do the same.....plz give me steps to follow......i have the manual but it shows some errors when i execute thro' it.....

expecting ur reply...
with regards,

Hi, mailgvenkat:

1. Removing the middle substrate will only change the Zc slightly. It will not help.
2. The strange behaviour of Zc is due to the thin strip. It is so thin (10 nm) that even gold has 2 ohms / square. It is way too lossy. It changes the behaviour of the metallic strip completely.
3. The meshing frequency Fmax = 20,000 GHz is very high. It is completely due to the small structure you have. If you choose significantly smaller Fmax, your whole strip will be meshed into one cell. It may degrade the accuracy.

Best regards,

hi jian,

help me for the following questions..

1. i could not plot the near field pattern for the structure dimensions given above...could u please help me in plotting the same...i read the manual but on following it i am getting certain errors...

so please specify me the steps for plotting the near field pattern.....can we animate it for each time step so as to view the result...

2. Also i have a doubt...whther the operating freq.has any effect on structure dimensions or viceversa.....(is there any optimum range for freq.used in micrometer range)?....

reply me....with regards and thanking you...

Hi, Mailgventat:

I tried the example for you and it is working. Here are the files. Please read the appendix for the near field calculation and visualization.

Best regards,

hi jian,

i simulated those files given by u works welll.......but i have a doubt sir if the char.impedance is too large, does it infer that the loss is high........surely it is a disadvantage for those structures?.....

Should i need to increase the dimensions of the strip & as to reduce the char.impedance....normaally we have char.impedance to be 50 to 100 ohms...

can u give me the significance of large/small value of char.impedance ?.....plz...

Wtih regards,

Hi, Mailgvenkat:

At least, I can see you need to increase the cross-section of the strip. It is very lossy with such a small cross-section even for the best normal conductor. Certainly, you may be ok if you use HTS.

Best regards.

ok ,very good! simulation with ie3d, RFID ,are we setup groud? my result ' error is %65

Not open for further replies.

Part and Inventory Search

Welcome to