Printed Circuit Design & Fab - October 2008 - (Page 34) ModElinG a h a h d h d h w=d d r w=d d r d d Figure 2 –Indentation profiles considered for the surface roughness. FiGurE 2a FiGurE 2b FiGurE 2c FiGurE 2. Indentation profiles considered for the surface roughness. Figure 2 –Indentation profiles considered for the surface roughness. FiGurE 4. Insertion loss comparison for the different indentaFigure 4 –Insertion loss: comparison for the different indentation profiles tion profiles. Conclusion The effect of the conductor surface profile (surface roughness) on a standard interconnect stripline structure was analyzed. The electrical performance of the interconnect was studied in frequency FiGurE 3. Insertion loss comparison between Hammenstad Figure 3– Insertion loss: comparison between Hammenstad and for FIGuRE 2A. and full wave (S-parameters), and a consistent decrease in the transmisdomain and Jensen formulation and full wave results Jensen formulation results. sion properties of the line have been highlighted. Hammenstad and Jensen’s analytical formulation is demonstrated to be effective them to model the frequency dependent loss by means of an only for a specific profile of the surface roughness; therefore, a additional loss term defined as: unique closed formulation which allows for the modeling of the ' K sr (2) real copper surface profile is not yet available. Future research will EQuation (2) c c be in the direction of probabilistic approaches that appear to be 2 well suitable for surface roughness modelling9. pCd&f 2 (3) K sr 1 arct g 1.4 EQuation (3) Figure 3– Insertion loss: comparison between Hammenstad and Jensen formulation and full wave results. s rEFErEncES where Rs surface resistivity, Z0 impedance of the transmis1. X. Shi, Jian-Guo Ma, Manh Anh Do, Er-Ping Li, “Sensitivity Analysis of Coupled Interconnects for RFIC Applications, IEEE Transaction on Elec” sion line and w, trace width. Surface resistance is a material tromagnetic Compatibility, vol. 48, November 2006, pgs.607-613. property, partially governed by surface roughness, while Z0 2. t. Liang, S. Hall, H. Heck and G. brist, “PCb transmission Line Modeling and w are both design parameters. for Multi-Gb/s Link Analysis, DesignCon East 2005. ” FiGurE 3 compares the insertion loss results due to the full 3. R. kollipara et Al., “Practical Design Considerations for 10 to 25 Gbps wave numerical simulation for the stripline model represented Coppewr backplane Serial Links, DesignCon 2006. ” 4. G. brist, S. Clouser, S. Hall, and t. Liang, “Non-Classical Conductor in FiGurE 2a with the results due to the Hammerstad and a Losses due to Copper Foil Roughness and treatment, IPC Electronic ” Jensen analytical formulation. A good agreement between the h h Circuits World Convention, Feb. 2005. two sets of data is observed. r 5. X.Chen, “EM Modeling of Microstrip Conductor Losses Including SurAt this point, alldthe profiles represented in FIGURE 2 d” d w=d face Roughness Effect, IEEE Microwave and Wireless components Letare modelled and simulated by means of CST MWS and CST ters, vol.17 February 2007 pgs.94-96. , , 6. M.V.Lukic, D. S. Filipovic, “Modeling of 3-D Surface Roughness Effects Design Studio is used to cascade four blocks of the 3D model with Application to in order to get the Figure 2 –Indentation profileslong stripline. the surface roughness. µ-coaxial Lines,” IEEE Trans. on Microwave Theory insertion loss of a 1600 μm considered for and Techniques, vol.55, March 2007 pgs.518-525. , The simulations have typically used 4 GB of random access 7 CSt Studio Suite 2008 www.cst.com. . memory taking on average, 50 minutes of running time on a 8. A.Ciccomancini Scogna, M.Schauer, “Stripline Simulation Model with 3.6 GHz Intel Pentium D processor. The results are reported tapered Cross Section and Conductor Surface Profile, Proc. of IEEE ” International Symposium on EMC, July 4-12, 2007 Hawaii, uSA. , in FiGurE 4. 9. P G. Huray, S. Hall, S. Pytel, F oluwafemi, R. Mellitz, D., Hua, P Ye, “Fun. . . It is noted that the level of increase in conductor losses is damentals of a 3-D “Snowball, Model for Surface Roughness Power dependent of the shape and distribution of indentation and Losses, on Proc. of SPI, May 13-16, 2007 Genova, Italy. ” , that the rectangular profile produces different results from the 10.J. Carroll, k. Chang, “Metallization Effects on GaAs Microstrip Line other two cases; therefore, Hammerstad and Jensen analytical Attenuation, IEEE Trans on Microwave Theory and Tech. vol.41, July ” 2003, pgs.1227-1229. formulation is only valid for a specific profile, but for more complex geometries a 3D model needs to be realized and a full dr. antonio ciccoMancini ScoGna is a senior application wave analysis needs to be performed. engineer at CSt of America Inc. He can be reached at antonio.ciccomancini@cst.com. 34 printEd CirCuit dESign & fAB OCTOBER 2008 http://www.cst.com
For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.