Hi All,

This is probably going to sound a bit crazy but I'm after a DIY method to come up with a pcb stack that would work at 40G and above. Since aluminum has a low Dk why cant I glue a one-sided FR4 copper clad board to either side of a 1/8 inch aluminum plate and create vias between (isolating via tunnels from aluminum obviously). Would this even remotely work?

Thanks,

Joe

 

I would recommend, if you were to try this, that you use alumina, not aluminum. One is a ceramic and the other is a metal. The Dk of metals is infinite. The Dk of alumina is about 9.5-11.5, depending on a variety of factors (excluding its frequency dependence).

Ignoring the material choice, the idea you have is the basis for a laminated RF PCB. However, actually achieving this is...not easy. You need some expensive equipment to be able to make good laminations with no voids (which can greatly affect RF performance) while also retaining mechanical strength, and then doing plated vias with it will also be challenging.

Additionally, "40G and above" is VERY ambitious, considering the wavelengths at these frequencies is very short (30 Ghz is 10 mm, and with the permittivity of alumina or FR-4, is even shorter - down to about 3-4 mm!). On top of that, 1/8" (125 mil) is VERY thick for these frequencies. Using a calculator from Microwaves101, a 50-ohm, 90-degree line would be 311 mils wide and 24 mils long; not exactly a..."common" aspect ratio.

In the long run, while your idea is on the right track, "DIY" methods for boards at these frequencies just really isn't all that viable unless you've got a good-sized budget, space for equipment, and a lot of experience with RF. It would be significantly easier to just design into a Rogers-Corp.-material-based stackup (such as RO-4350b laminate on FR-4 core). Frequently, from my experience, stackups at these frequencies are barely a few 10s of mils (with the core being the thickest).