CircuiTree - July 2008 - (Page 24)

Toward a PCB Production Floor Metric for Go/No Go Testing of Lossy High-Speed Transmission Lines the two parallel lines of the switchback was 0.432 mm (0.017 in). This ﬁgure was calculated from a conventional rule of thumb to space aggressor traces (>5x height above the ground plane for microstrip, >3x height for stripline). The following TDR trace (Figure 6) shows clearly the unwanted effects of that folding, with the two halves of the traces running parallel. On the surface microstrip TDR trace shown in Figure 6, the switchback can clearly be seen as a large capacitive dip at 3200 ps with the following portion of the trace raised by additional reﬂections and crosstalk. Switchback effects could still be seen with a spacing of 1.42 mm (0.056 in). From these practical results it would seem that any surface microstrip trace, laid out to include a switchback, would have to have the trace separation in excess of this value. As expected, the embedded stripline trace with a separation of 0.889 mm (0.035 in) shows no sign of coupling or erroneous reﬂections because FEXT effects are minimal. Corners will always have a subtle effect, but at these frequencies, the stripline trace could be laid out in either a linear or switchback fashion. and via at both ends allowing investigation using TDT and VNA techniques. Figure 7 and 8 show that the vias as manufactured on this board have no signiﬁcant effect on the RIE result at these frequencies. Probes/R&R Using repeated measurements taken by three appraisers, the R&R of two probing scenarios were investigated. In the ﬁrst case the short reference trace and the longer trace were measured in two separate probing actions. In the second case a specially conﬁgured probe was used to capture the data from both of the traces in one probing action. The ﬁgures in Table 1 show the calculated variations in RIE due to different contributing factors. It was found that a better R&R was achieved by using a single probing action over addressing the coupon twice. There are also obvious test time and error beneﬁts to be made in the test/production environment. In addition, stripline results show far less part variation than those of surface microstrip. Figure 4 RIE Values for 12 in and 8 in Surface Microstrip Traces Figure 5 RIE Values for 12 in and 8 in Stripline Traces ilar in both data sets. In previous studies using these same systems, a much closer correlation was achieved. This magnitude difference needs further work to identify its cause. From Figure 4 and 5, it can be concluded that the data from the 200 mm still allow a distinction between the higher speciﬁed dielectrics and the more lossy 410 material, but the difference between RIE values is smaller (as expected). A 200 mm trace could be used for gross material differentiation, but a 300 mm trace would be necessary for more subtle distinctions. The choice between the two may be based on availability of board space versus desired precision. VNA Correlation The loss of the 300 mm (12 in) traces (Figure 9 to 12) was measured by VNA. In both the stripline and microstrip through (S21) loss, two distinct groups of curves are seen that represent the different material types. The Nelco 4000-13-SI and the IS408 materials are characterized with a lower loss relative to the IS410 (1080) and the IS410 (2116). An exception to this grouping is seen for layer 8 of the IS408 Via Effects Some traces were manufactured with an entry pad and via at one end of the trace only, mimicking expected production coupons. Other traces included a contact pad Switchback Results To save on board space it would be convenient if the 200 mm trace could be laid out in a manner such that it folded back on itself, thus occupying less linear space. The closest spacing, included on this board, between Figure 7 RIE Values for Single- and DualVia Ended Surface Microstrip Traces Figure 9 12 in S21 Microstrip Loss Figure 6 TDR Trace Showing the Effects of Switchback on the Microstrip Test Trace 24 July 2008 • circuitree.com Figure 8 RIE Values for Single- and DualVia Ended Stripline Traces Figure 10 12 in S21 Microstrip Loss (Zoomed) http://circuitree.com

Table of Contents Feed for the Digital Edition of CircuiTree - July 2008

CircuiTree - July 2008 Contents My Line Industry Review Tech Talk Flexible Thinking Toward a PCB Production Floor Metric for Go/No Go Testing of Lossy High-Speed Transmission Lines Intelligent Design 20-Year Retrospective Ask the Flexperts Environmentally Speaking BPA Growth Curves Considering Design Variants to Maximize Process Efficiency Market Outlook Technical Product Spotlights Classified Ads Upcoming Events Ad Index

CircuiTree - July 2008

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.