1.  Polar Instruments 2002 Differential Impedance Effect of Etch Taper, Prepreg, and Resin Flow on the Value of the Differential Impedance Ken Taylor Polar Instruments Dr.Alan Staniforth, Martyn Gaudion

2.  Polar Instruments 2002 Discrepancies Software accuracy Closed form equations Empirical data tables Field Solver Structure Cross Section Track profiles Dimensions –Predicted –Microsection measurements Dielectric Value Non-homogenous –Bulk specification

3.  Polar Instruments 2002 Closed Form Equations Used very successfully until 3 ~ 5 years ago Valuable Approximations –Various degrees of complexity –See IPC-2141, Wadell, Cohn, Hilberg, others –Good for larger track dimensions > 20 mils –Relatively few, simpler structures Unsuited for differential structures Can be performed on a typical scientific calculator

4.  Polar Instruments 2002 Field Solver Highly complex mathematical modeling Accurately predicts E-field vectors Accounts for track cross sectional profile –Charge distribution Method of Moments –Green’s Function  < 0.5%  < 2% No practical lower dimension limit... yet Requires PC

5.  Polar Instruments 2002 Field Solver Assumes regular trapezoidal track cross section Equal etch tapers Assumes track identical pair Requires good fab process Reduces calculation resource

6.  Polar Instruments 2002 Track Assumptions

7.  Polar Instruments 2002 Charge Distribution and Track Profile

8.  Polar Instruments 2002 Field varies with Track Profile

9.  Polar Instruments 2002 Impedance Varies with Etch Taper

10.  Polar Instruments 2002 FR4 Dielectric Core and Prepreg have different  r Both are Mixtures of Resin and Glass Fibers Materials are non-homogenous –  r specified for laminate is the bulk value –  r for glass ~ 6.1  r for epoxy ~ 3.2 So significant local variations occur for  r

11.  Polar Instruments 2002 Typical E-field distribution Embedded Microstrip rr

12.  Polar Instruments 2002 FR4 structure 5 mil P P P C C

13.  Polar Instruments 2002 FR4 structure

14.  Polar Instruments 2002 FR4 structure 3.1 4.2

15.  Polar Instruments 2002 Field distribution Impedance value increases

16.  Polar Instruments 2002 FR4 structure 3.1 4.2 Impedance Value Increases

17.  Polar Instruments 2002 Resin Layer in Differential Microstrip

18.  Polar Instruments 2002 Resin Layer in Differential Stripline

19.  Polar Instruments 2002 Conclusions Simplified modeling of differential structures leads to impedance discrepancies of several ohms A field solver is required which takes into account –Structure –Track dimensions Thickness Widths Spacing Etch taper Symmetry –Dielectric layer composition and  r values

20.  Polar Instruments 2002 Conclusions Resin flow into region coplanar with tracks increases the impedance of typical –embedded microstrip by 3.5 ~ 4 ohms –embedded stripline by 3 ~ 3.5 ohms

21.  Polar Instruments 2002 Conclusions Accurate implementation of predicted impedance requires –Accurate production of predicted dimensions Track width Track spacing Track etch taper –Increased etch taper leads to enlarged resin region Track symmetry –Loss of symmetry also affects differential performance –Accurate knowledge of dielectric composition Dielectric layers Resin region and component  r value

22.  Polar Instruments 2002 Questions now?…. Questions later? –Americas: 800 328-0817 ken.taylor@polarinstruments.com richard.smith@polarinstruments.com –Europe: martyn.gaudion@polarinstruments.com –Asia: amit.bhardwaj@polarinstruments.com Thank You