Issue link: https://iconnect007.uberflip.com/i/241303
article influence of via stub length and antipad size on the insertion loss profile continues Figure 10: Screen shot of single-ended insertion loss testing. A screenshot of typical measurement data is shown in Figure 10. The orange trace in the upper portion of the display is the magnitude of insertion loss over the full frequency range for a coupon with a very short stub, where the yellow trace is for a long stub. The lower part of the screenshot shows magnitude and phase for all four single ended S-parameters. For single-ended structures, four different antipad sizes and 10 different stub lengths were measured on five panels with two identical coupons each. This resulted in 400 full 2-port S-parameter matrices, spanning the frequency range from 10 MHz up to 40 GHz with 2048 points. To exclude odd readings in the data, the magnitude of the insertion loss was plotted for each of the 400 measurements in one chart, (Figure 11). To get a less noisy picture of the influence of the stub length and the antipad sizes, the data of the five panels and two identical coupons for each stub length/antipad size combination were averaged and plotted (Figure 12). The via stubs cause a large resonant dip, with the longest stubs creating the notches in the insertion loss curve at lower frequencies than the shorter stubs. The antipad size generates some small changes, but with less of a clear effect than the stub length. To evaluate the influence of the antipad size, an ANOVA was performed, with results presented in Figures 17, 19, 20. To answer the question of the maximum acceptable stub length, the additional insertion loss caused by the via stubs is extracted from January 2014 • The PCB Design Magazine 51