Author 
Topic 

magent1900
United Kingdom
3 Posts 
Posted  Jun 11 2015 : 11:18:29

Dear colleagues I have recently known about fastcap and fasthenry, where i need to implement stripline in 90nm tech. i need your help to get started with this software, where my job is on two stage, firstly get the parasitic of the stripline alone and secondly with coupled one. my work dimensions w=4um, t=1um, and l=10mm in the center of two gnd plats in the top layers where the dielectric thickness 4um, and Er=3.9. are there any ready codes or the way to determine the units to implement like this work? I ask you for help as you could thanks in advance.
moh 

Enrico
435 Posts 
Posted  Jun 14 2015 : 00:41:57

If you need to model a stripline, I strongy suggest you to use FasterCap in 2D mode. Actually, FasterCap / FastCap2 and FastHenry2 are quasistatic simulators. This means that the results are valid if the dimension of the structure is small with respect to the minimum wavelength of interest. Now, this is usually not true for a stripline or a transmission line in general, as you are interested in the signal propagating along the line. However, if the crosssection of the transmission line is small w.r.t. the wavelength of the maximum frequency of interest, can still use quasistatic simulators to extract the parameters perunitlength of the structure.
Said that, since you are interested in the crosssection, you should assume that you have an infinitely long line. This is why FasterCap in 2D mode is the perfect way to extract these parameters.
You can check any of the examples in the 'Samples>2D' directory of the FasterCap installation. Please be sure to read the comments inside the files, and possibly the papers referenced there.
Best Regards, Enrico



magent1900
United Kingdom
3 Posts 
Posted  Jun 16 2015 : 10:44:13

Error: Generic error, program execution stopped!
I have this message when i tried to execute the following code
*
C microstrip_top_4um.txt 4.2 0.0 0.0000015 + C microstrip_bottom_4um.txt 4.2 0.0 0.0000015
D diel_microstrips_4um.txt 4.2 0.0 0.000004
C gnd_plane_n.txt 4.2 0.0 0.0 + C gnd_plane_n.txt 4.2 0.0 0.000004
could any one help me to know what is the problem?
best regards moh 
Edited by  magent1900 on Jun 16 2015 10:45:39 


Enrico
435 Posts 
Posted  Jun 18 2015 : 12:17:11

As explained in the embedded help, in case of 2D simulation, the first line should start with
* 2D
otherwise FasterCap will assume you are running a 3D simulation; however as the input files will be invalid (they are 2D), they will eventually lead to an invalid solution, with consequent error.
Best Regards, Enrico



magent1900
United Kingdom
3 Posts 
Posted  Jun 19 2015 : 18:05:45

many thanks, it's working. but what if i want to find the inductance (L) for the same stripline, should i write a new code in fasthenry or what?
moh 
Edited by  magent1900 on Jun 19 2015 18:15:19 


Enrico
435 Posts 
Posted  Jun 22 2015 : 23:47:36

quote: but what if i want to find the inductance (L) for the same stripline, should i write a new code in fasthenry or what?
Actually you could do both: i.e. either perform a FastHenry2 simulation, or use again FasterCap.
The reason for the former, i.e. using FastHenry2, is straightforward I think. However, unless you are interested in the behavior of the line in a low frequency range i.e. when skin effect is not so pronounced, you can avoid using a 3D simulation to extract 2D parameters. On this topic you can read the parallel thread under the FastHenry forum with the title "impact of length on perunitlength calculation".
The reason for the latter, i.e. using FasterCap again, is not so straightforward, but if you read the articles I pointed you to, you should have some clues about it. Actually if you read the paper from A. R. Djordjevic, T. K. Sarkar, "ClosedForm Formulas for FrequencyDependent Resistance and Inductance per Unit Length of Microstrip and Strip Transmission Lines" IEEE Trans. on Microwave Theory and Techniques, vol. 42, no. 2, February 1994, you will understand that L and C per unit length in a transmission line are related by the speed of light. So you can derive the value of L (at high frequency) from the C extracted when no dielectric is present. You already know C when there is dielectric medium (your first FasterCap simulation), so you have all your data for using the resulting values in, say, a Spice simulation.
Best Regards, Enrico




Topic 


