PCB007 Magazine

PCB007-Jan2024

Issue link: https://iconnect007.uberflip.com/i/1514628

Contents of this Issue

Navigation

Page 75 of 95

76 PCB007 MAGAZINE I JANUARY 2024 fiberglass styles. e disadvantage of the spread weave fiberglass is that today's high viscosity ceramic-filled PPO resins do not flow, or have at best limited flow, from one side of the spread weave to the other. e spread weave behaves more like a flat plane of glass. Whatever voids/ spaces there may be within the fiberglass X-Y plane do not allow the ceramic or flame-retar- dant particles to pass through the glass. e high viscosity of the resin system used to main- tain a uniform distribution of the ceramic and flame retardant is one more inhibitor of flow from one side of the flattened weave through to the other side. e result is that flow and fill of difficult circuitry can only be achieved by resin on one side of the fiberglass. ese mate- rials are challenged to completely flow and fill plated-up HDI or two-ounce copper layers. Multiple plies of prepregs can be used but the trend is toward thinner dielectrics, not thicker layers of prepreg. Second, because it is well known that the flat- tened weave-based materials are worse for flow and fill, the tendency is to use prepregs with higher (75–80%) resin contents. While these high resin content prepregs are better suited for heavy copper and plated-up layers (the high resin content can result in higher CTE values) they are more likely to crack because the resin- rich regions lack reinforcement, and will yield worse reliability in HDI stackups. As designers require thinner dielectric mate- rials, the choice for the fabricator becomes limited and the fabricator is forced to use fiber glass styles like 1067 or 1086 as shown in Figure 18. ese glass styles are very tightly woven materials and limit the amount of resin that can flow from one side to the other. is causes issues when heavier copper like two- ounce is used and the fabricator observes voiding issues in low pressure areas. is is especially a problem in very high layer count boards (40+ layers). One solution is to use non-reinforced pre- pregs with flow rates greater than 65% and with no barriers to hinder the flow. e test vehicle used was a planar magnetic, four-ounce, 20-layer design, which provides the worst-case scenario for high and low pres- sure areas due to the stacking of copper in the coil areas of the design. Due to the very high flow rates, the initial testing containing 70+ sheets of prepreg failed because of the exten- sive flow and lamination pressure. Figure 18: Images of various spread weave fiberglass styles. Figure 19: Cross-sectional image showing the pre-filled and final lamination layers.

Articles in this issue

Archives of this issue

view archives of PCB007 Magazine - PCB007-Jan2024