[SI-LIST] Re: Couplin capacitance

Jeremy Plunkett Wrote:
>
>
In your results, the ratio of the coupling capacitance to the capacitance to
ground becomes slightly smaller as the trace width is increased while the
spacing is held constant, indicating that total capacitive crosstalk is also
decreasing.  This is what I would expect based on similar simulations I've
done using the Hspice 2D field solver.  However I would check your geometry
and units, because the amount of coupling capacitance is very high compared
to what I would expect.
regards,
Jeremy
>
>
Dear Jeremy:

Thanks for your reply.

One critical part of the geometry and unit conversion involves maintaining
the separation of the two traces fixed (at 6 mils) while varying the trace
width (4, 7 and 10 mils).  Let us examine it closely.

My original post shows that the two trace files for the Min case (W= 4 mils)
were generated utilizing the MIT cubegen program:

>cubegen -xo-0.000127 -xh0.00011 -yh0.0254 -zo0.00014 -zh0.000018 -nx3 -ny3
-nz3 > trace1_min.in

>cubegen -xo0.000127 -xh0.00011 -yh0.0254 -zo0.00014 -zh0.000018 -nx3 -ny3 -
nz3 > trace2_min.in

xo-0.000127 and xo0.000127 imply center-to-center spacing of 0.000254 m = 10
mils.  Furthermore, xh0.00011 means trace width of 0.00011 meter ~ 4 mils.
Therefore, edge-to-edge separation is 10 - 4 = 6 mils.

Unit conversion formulas:
1 mil = 0.0000254 meter.
1 inch = 0.0254 meter.
1 meter = 39,370.08 mils = 39.37 inches

For the typ case (W = 7 mils) the desired trace files were created by
commands:

>cubegen -xo-0.000165 -xh0.00018 -yh0.0254 -zo0.00014 -zh0.000018 -nx3 -ny3
-nz3 > trace1_typ.in

>cubegen -xo0.000165 -xh0.00018 -yh0.0254 -zo0.00014 -zh0.000018 -nx3 -ny3 -
nz3 > trace2_typ.in

xo-0.000165 and xo0.000165 mean that center-to-center spacing between these
two traces is 0.0003 meters = 13 mils.
Since the trace width as indicated by xh0.00018 is 0.00018 m = 7 mils,
then edge-to-edge spacing equals 6 mils

Similarly, for the max case (W=10 mils) the two required files
trace1_max.in and trace2_max.in, were obtained via:

>cubegen -xo-0.0002 -xh0.000254 -yh0.0254 -zo0.00014 -zh0.000018 -nx3 -ny3 -
nz3 > trace1_max.in

>cubegen -xo0.0002 -xh0.000254 -yh0.0254 -zo0.00014 -zh0.000018 -nx3 -ny3 -n
z3 > trace2_max.in

Here the Center-to-center spacing is 0.0004 m  ~ 16 mils as dictated by
xo-0.0002 and xo0.0002.
The trace width, implied by xh0.000254, is 0.000254 m = 10 mils
The edge-to-edge spacing is again ~ 6 mils.

The height of each trace relative to ground plane is determined by
zo0.00014. Its value of 0.00014 m ~ 5.5 mils is larger than the desired 4
mil dielectric substrate thickness because the plane layer is 1.44 mil
thick.

It would be interesting to compare FastCap results
with those generated by Hspice 2D field solver, XFX, or other field
solver programs.  I believe that FastCap is a 3D field solver capable of
producing accurate data.

An approximation was used in my analyses that needs to be mentioned.
When generating the ground plane file:

>cubegen -xo-0.019 -xh0.0381 -yh0.051 -zh0.0000366 -p -t -nx10 -ny10 -nz10 -
e1 > plane.in

As you can see the " -p " and  " -t  " options of cubegen are applied in
order to reduce memory requirements and speed up FastCap computations.
Perhaps, someone likes to comment regarding  such approximations, fringing
field effects, optimum discretization, etc.

Best Regards,

Abe





------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field

or to administer your membership from a web page, go to:
http://www.freelists.org/webpage/si-list

For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field

List archives are viewable at:     
                http://www.freelists.org/archives/si-list
or at our remote archives:
                http://groups.yahoo.com/group/si-list/messages 
Old (prior to June 6, 2001) list archives are viewable at:
                http://www.qsl.net/wb6tpu

Other related posts: