| Author |
 Topic  |
|
|
indr29
India
4 Posts |
 Posted - Feb 22 2013 : 11:52:11
|
I am going through the FastHenry_nonuniform_gndplane_manual, at page no 26 of that, an example of 3D inductance is given(3d_exaqmple2.inp) in input files. that calculating inductance of M5 over Si substrate. M5 line is at a distance of 8 um above the Silicon substrate. But while defining the segment, nodes are defined at distance of z = 223.5 ?
NS1 x=0 y=-1000 z=223.5
NS2 x=0 y=1000 z=223.5
EM5 NS1 NS2 w=26 h=1 sigma=26 nwinc=4 nhinc=1
1. My first query is why z=223.5 instead of 8 um, as separation between M5 and substrate is 8 um ?
2. Secondly, while describing the coplanar GSG system,
**FAST HENRY CODE**
* GSG Analysis using fasthenry
* * Set the units for all following dimensions
.units um
*
.default z=0 sigma = 3.658e1
.default nwinc=8 nhinc=1
*
* 12 as gnd plane segment
g1 x1 = -12.4 y1 = 0 z1 = 0
+ x2 = -14.4 y2 = 0 z2 = 0
+ x3 = -14.4 y3= 1000 z3 = 00
+ thick = 2
* discretization:
+ seg1 = 20 seg2 = 20
* nodes to reference later:
+ nin1 (-13.4,0,00)
+ nout1 (-13.4,1000,00)
*
* g1 as gnd plane segment
g2 x1 = 32.4 y1 = 0 z1 = 00
+ x2 = 34.4 y2 = 0 z2 = 00
+ x3 = 34.4 y3 = 1000 z3 = 00
+ thick = 2
* discretization:
+ seg1 = 20 seg2 = 20
* nodes to reference later:
+ nin2 (33.4,0,0)
+ nout2 (33.4,1000,00)
* L shaped trace over ground plane
*
* The nodes
N1 x=0 y=0 z=00
N2 x=0 y=1000 z=00
* The elements connecting the nodes
E1 N1 N2 w=0.8 h=2
* Short together the end of the g2, g3 and its corresponding
* point on the ground plane directly beneath nin1, nin2, nout1, nout2,nin,nout
.equiv nin1 N1
.equiv nin2 N1
* compute loop inductance from (N2) to its corresponding
* point directly underneath (nout) on the groundplane
.external N2 nout1
.external N2 nout2
* Compute impedance matrix for low frequency (essentially DC)
.freq fmin=5e8 fmax=5e8 ndec=10
* mark end of file
.end
**********************
I have used two external commands for inductance calculation. But dont know whether i am right or wrong. Is this the correct method for describing a GSG system ?
Also if it is correct then another doubt is that after executing the FH code 2 X 2 impedance matrix obtained
Computed matrices (R+jL)
Row 0: n2 to nout1
Row 1: n2 to nout2
Freq = 5e+008
Row 0: 23.5958+1.14434e-009j 17.086+5.4546e-010j
Row 1: 17.086+5.45446e-010j 23.5957+1.50142e-009j
Now question is which one is Self inductance and which one is mutual? How to obtain the total inductance of the GSG system ?
Sorry if i asked trivial questions.
|
|
|
Enrico
550 Posts |
Posted - Feb 26 2013 : 18:23:44
|
Let's go one by one
quote:
1. My first query is why z=223.5 instead of 8 um, as separation between M5 and substrate is 8 um ?
The thickness of the substrate is 430 um:
g1 x1 = -1500 y1 = -1500 z1 = 0
+ x2 = 1500 y2 = -1500 z2 = 0
+ x3 = 1500 y3 = 1500 z3 = 0
+ thick = 430 sigma= 0.015
The z coordinate is zero, so the plane thickness extends 215 um above zero and 215 um below zero.
Now the line is 8um above the plane, so this bring us to 215+8 = 223 um. We are almost there, and you should have understood now the 'trick': the element connecting the nodes is 1 um thick
EM5 NS1 NS2 w=26 h=1 sigma=26 nwinc=4 nhinc=1
and the node position is in the center, so this is 0.5 um above and 0.5 um below, giving us the magic number 223.5
quote:
2. I have used two external commands for inductance calculation. But dont know whether i am right or wrong. Is this the correct method for describing a GSG system ?
It is correct but I understand you wanted the result in a different format. In this case you are treating the two grounds (left and right) as if they were independent conductors, so you get two inductances (and their mutual couplings). In reality your assumption is that they are BOTH ground, so you want a single inductance value, considering both the 'left' loop and the 'right' loop. To get directly this value you need to make the two grounds the same conductor, i.e. you have to short them also (at least) on the output port:
.equiv nin1 N1
.equiv nin2 N1
.equiv nout1 nout2
* compute loop inductance from (N2) to its corresponding
* point directly underneath (nout) on the groundplane
.external N2 nout1
Then the external declaration is only one, since nout1 is shorted with nout2 so it is the same.
quote:
3. Now question is which one is Self inductance and which one is mutual
This question is not applicable any more but, for the sake of completeness, the self inductances are on the diagonal and the mutual are off-diagonal.
Best Regards,
Enrico
|
 |
|
| |
Topic |
|
|
|
GSG Coplanar Waveguide
