| T O P I C R E V I E W |
| faizanhaaq |
Posted - Sep 05 2014 : 09:00:44
I am having some problems in FastHenry simulations of a microstrip trace. I want to calculate inductance of microstrip trace.I have following questions regarding that. It will be really good if you can help me in this regard.
1) In Fasthenry examples 'simple_gp', i saw that one end of L trace has been connected to ground plane. And the overall inductance is calculated for the L trace and ground return path. Why is this case?. Can i not find the trace inductance only by not connecting it to ground plane?. Which one will be the right value? One with the ground plane path or one without ground plane return path?
2) Ground planes in Fasthenry are defined in two ways. One with the segmentation (in 'simple_gp') file and one without segmentation as in '3d_example2'. The values of inductance comes different when i define ground plane with and without the segmentation as follows:
Without segmentation
G1 X1 = Y1 = -1500 -1500 0 Z1 =
x2 + y 2 = 1500 -1500 = z2 = 0
= x3 + y3 = 1500 1500 Z3 = 0
+ THICK = 430 sigma = 0.015
+ spin = NONE
With Segmentation:
g1 x1 = 0 y1 = 0 z1 = 0
+ x2 = 1000 y2 = 0 z2 = 0
+ x3 = 1000 y3 = 1000 z3 = 0
* thickness:
+ thick = 1.2
* discretization:
+ seg1 = 20 seg2 = 20
Moreover, the inductance value also differ with different segmentation values. What should be the right way to define the ground plane?. With or without segmentation and with what segmentation value?
faizan |
| 2 L A T E S T R E P L I E S (Newest First) |
| faizanhaaq |
Posted - Sep 10 2014 : 09:39:33
Thanks Enrico. It really helped
faizan |
| Enrico |
Posted - Sep 08 2014 : 18:46:28
quote:
1) In Fasthenry examples 'simple_gp', i saw that one end of L trace has been connected to ground plane. And the overall inductance is calculated for the L trace and ground return path. Why is this case?. Can i not find the trace inductance only by not connecting it to ground plane?. Which one will be the right value? One with the ground plane path or one without ground plane return path?
Inductance is a property of a closed loop. If you only consider the line, you get a partial inductance value. For details and the definition of the concept of "partial inductance" you can refer to the very good explanation by M. Kamon, "Fast Parasitic Extraction and Simulation of Three-Dimensional Interconnect via Quasistatic Analysis", PhD thesis at the Massachusetts Institute of Technology.
You can download the paper from the M.I.T. site, whose link is in the 'External Links' page you can reach through the 'Services' page (of you can google it).
Bottom line: you need to connect the plane and the line at the far end.
quote:
2) Ground planes in Fasthenry are defined in two ways. One with the segmentation (in 'simple_gp') file and one without segmentation as in '3d_example2'. The values of inductance comes different when i define ground plane with and without the segmentation as follows:
Actually '3d_example2' has a very fine segmentation, but it uses the non-uniform ground panel definition to do so. Please refer to the original FastHenry's User's Manual(s) for details. You can access the manuals from the first page of the embedded on-line help.
This manuals will also clarify what is the 'correct discretization', that actually depends on your frequency of interest vs. the dimensions of your structure and the properties of the material (conductivity).
I may also add that since you are "trying to get 2D values from a 3D model", i.e. your structure will be uniform along the length, and you are probably not interested in the fringing field effects at the end, you need a long enough structure. You also might be better off using, for the ground plane, a segment and not a full ground plane, as your current will flow along the length for the most part.
You could also use FasterCap in 2D mode, and calculate the L from the C, if you are only interested in the limit of 'high frequency' of the microstrip (skin effect already confined the current to the surface only, but the quasistatic assumption is valid on the section)
Best Regards,
Enrico
|
|
|
