Issue link: https://iconnect007.uberflip.com/i/1355791
104 I-CONNECT007 I REAL TIME WITH... IPC APEX EXPO 2021 SHOW & TELL MAGAZINE Czaplewski: We're working on the next gen- eration of products and we need a little bit more signal integrity improvement. One of the levers to pull is to make some adjustments within the PCB design. So that's why we evalu- ated at reduced layer-to-layer misregistration and reduced antipad diameter around backdrill PTHs. is research is preparing us for the future product genera- tions. Johnson: Walk us through the paper, which is titled, "Signal Integrity, Reliabil- ity, and Cost Evaluation of PCB Interlayer Crosstalk Reduction." Czaplewski: e push for higher data rates and more functionality is lead- ing to higher density in PCBs, which increases the opportunity for crosstalk. We're looking to mitigate that. We modeled the signal integ- rity impacts of reducing layer-to-layer mis- registration from five mils, down to four and three mils. We also looked at reducing the antipad diameter in the backdrilled regions of PTHs from 30 mils on a 10-mil finished PTH to 28 mils. As one reduces the antipad diameter on the backdrilled holes, there is an increased risk of exposing the planes dur- ing the backdrill operation, which can cause a wide range of reliability issues. In the reli- ability portion of the paper, we sought to understand that by intentionally exposing the planes. We had some of the backdrilled holes filled with resin to evaluate if that would have an impact on reliability. ese reduced layer misregistration and reduced antipad diam- eter are going to impact the yield at the PCB manufacturer. So we also conducted manufac- turer polling to get a relative cost impact of our two proposed changes to create a cost benefit analysis. Johnson: I found it quite interesting that the two different solutions—misregistration as well as reducing the diameter of the antipads— are both intended to get more precise and more condensed board fabrication. You're pushing tolerances closer to zero in both cases. Czaplewski: Correct. Johnson: Walk us through what you found. Czaplewski: I'll start with the reliability part because that's my expertise area. I expected to have electrochemical migra- tion or some sort of corrosion forming between biased planes within the backdrilled holes, but surprisingly we did not detect any fails in the unfilled holes. We actually had some fails in the filled holes, which was opposite of what we were expecting. We're still investi- gating this, but we're thinking it's related to the hole fill process and potentially the hole clean process before resin fill. at was an interesting finding and something that needs to be inves- tigated more before implementing reduced antipad diameters, which, with manufacturing tolerances, can result in plane exposure. Johnson: Looking at your data, it seems that reducing misregistration did bring improve- ments, but the big winner was, in fact, the anti- pad diameter. Czaplewski: Yes. From the crosstalk reduction standpoint, reducing the antipad diameter had more benefit than reducing the layer-to-layer misregistration. However, you must take into consideration that reduced antipad diameter, in this case, is only useful on backdrilled layers. If there are not many signal traces running near backdrilled holes or on backdrilled layers, the benefits will be limited. e misregistration