Issue link: https://iconnect007.uberflip.com/i/1500943
54 DESIGN007 MAGAZINE I JUNE 2023 the construction loss. By conducting this test, the panel-to-panel variation is mitigated. Insertion Loss Data To simplify data analysis, a stacked bar chart was constructed (Figure 3), focusing on three frequencies commonly used in real-world applications. • 6 GHz: Standard operation • 40 GHz: 5G • 77 GHz: ADAS e stacked bar includes two components. First, the blue bar represents loss due to the construction of the test vehicle (the data obtained using the copper-only half of the test vehicle). Second, the red bar shows the addi- tional loss due to surface finish application to the second half of the same test vehicle. us, the total height of the bar represents the total loss for construction and finish. Insertion Loss Data Analysis Minor differences in the height of the blue bars (construction loss) can be observed at each frequency in Figure 3. In a perfect world, we would hope to observe the same signal loss from each of the copper- only test vehicles. However, experimental errors are possible, reducing repeatability between tests. e variance observed between the construction-only losses is believed to be due to either signal loss measurement error or, more likely, minor differences in the con- struction (etched definition, trace unifor- mity, surface roughness, etc.) between the individual test vehicles. ese differences are small enough to allow effective judgment and comparison of signal losses for each surface finish. Once taking into consideration the minor experimental variance, it is possible to catego- rize the surface finish test candidates into three categories: Category 1 OSP and immersion silver are shown to add no additional signal loss over the construction sample. Silver is a better conductor than cop- per while OSP is an ultra-thin organic coating; both finishes have no detrimental effect on the signal loss. Figure 3: Signal loss due to construction (blue) and surface finishes (red) at representative frequencies.