#11 questionable resistance output from calcRL

[Dan McMahill]

I'm not sure if this is a bug or a pilot error, but the
resistance output from the RL wavelet simulator seems a
bit off. For kicks I set up a 75 ohm microstrip line
which is made of a 50 mil wide trace over 62 mils of
er=4.8, loss tangent=0.025, dielectric and a ground
plane before that. The metal is set up to be 1 oz
copper which is 1.4 mils thick with a conductivity of
5e7 S/m (not exactly copper, but close enough).

I then stepped the frequency from 500 MHz to 2.5 GHz in
100 MHz steps and got the following RL vs frequency output:

5e+08 458.515 3.93703
6e+08 458.437 4.05789
7e+08 458.384 4.09798
8e+08 458.348 4.11396
9e+08 458.323 4.104
1e+09 458.306 4.07355
1.1e+09 458.295 4.03161
1.2e+09 458.287 3.98189
1.3e+09 458.282 3.92802
1.4e+09 458.28 3.87208
1.5e+09 458.278 3.82871
1.6e+09 458.278 3.76147
1.7e+09 458.279 3.70898
1.8e+09 458.28 3.65937
1.9e+09 458.281 3.61287
2e+09 458.282 3.56983
2.1e+09 458.284 3.50034
2.2e+09 458.286 3.49441
2.3e+09 458.288 3.46202
2.4e+09 458.289 3.43308
2.5e+09 458.291 3.40745

The problem is that at 500 MHz, the skin depth I
believe is about 0.13 mils or 1/10th of the line
thickness. Therefore, the entire 500 MHz to 2.5 GHz
range should be in the skin depth limited region and
the resistance should go up more or less with the
square root of frequency. The data instead shows a
small increase up to about 800 MHz and then resistance
goes down.

By my calculations, I should have seen more like 9
ohms/meter at 500 MHz and 9 * sqrt(f/500e6) ohms/m for
f>500MHz.

Here is the input file I had:

#----------------------------------
# File: /u2/disk1/dan/wcalc/mmtl/microstrip.xsctn
# Mon Sep 27 01:23:37 UTC 2004
#----------------------------------

package require csdl

set _title ""
set ::Stackup::couplingLength "0.0254"
set ::Stackup::riseTime "10"
set ::Stackup::frequency "1e9"
set ::Stackup::defaultLengthUnits "mils"
set CSEG 100
set DSEG 100

GroundPlane GrouPlan1 \
-thickness 1.4 \
-yOffset 0.0 \
-xOffset 0.0
DielectricLayer Substrate \
-thickness 62 \
-lossTangent 0.025 \
-permittivity 4.8 \
-permeability 1.0 \
-yOffset 0.0 \
-xOffset 0.0
RectangleConductors Trace \
-width 50 \
-pitch 1.0 \
-conductivity 5.0e7S/m \
-height 1.4 \
-number 1 \
-yOffset 0.0 \
-xOffset 0.0

So, am I doing something obviously wrong or is this a bug?

Thanks
-Dan

[Dan McMahill]

Logged In: YES
user_id=306927

I should have mentioned, this is tnt-1.2.2