What changes above 30 GHz, why PTFE laminates are the working assumption there, and the design and process practices that keep a millimeter-wave link stable.
As 5G, automotive radar and satellite links move into higher bands, millimeter-wave circuit design has turned into a specialty of its own. Above 30 GHz the available bandwidth is enormous, but so is the penalty for every imperfection in the material and the layout. Path loss climbs quickly with frequency, and variations that would be harmless at 3 GHz start to decide whether a link closes. Standard epoxy laminates like FR-4 are not candidates at these frequencies. Polytetrafluoroethylene-based substrates are, and this article explains what they contribute and how to design with them.
Operating above 30 GHz turns circuit design into a precision discipline. Four constraints dominate, and each one traces back to the substrate rather than to the schematic:
Polytetrafluoroethylene is a high-performance thermoplastic used as the base of high-frequency laminates, usually reinforced with glass fibre or ceramic filler. Four of its properties are the reason it holds this position in the millimeter-wave world.
Choosing the material is half the problem. The other half is the design and the process built on top of it, and both have to be planned with the same care.
At millimeter-wave frequencies microstrip is no longer the automatic answer. Grounded coplanar waveguide shields better, radiates less and makes it easier to place series and shunt components on the same layer. It also confines the field around the signal line more tightly, which matters in densely packed modules where leakage between neighbouring structures would otherwise be a problem. Our notes on processing considerations for high-frequency PCBs cover the practical constraints that come with these structures.
Full-wave electromagnetic simulation should come before layout release. The model is only as good as its inputs, so feed it the measured Dk and Df of the actual laminate, the finished dielectric thickness and the copper foil roughness. Because PTFE grades are consistent, simulation usually tracks measurement closely, and that close agreement removes most of the cost of iterating on hardware.
PTFE is softer and chemically more inert than FR-4, so the same process window does not apply. Three steps carry most of the risk:
Above 30 GHz, PTFE is not an optional upgrade; it is the working assumption. What decides whether a design succeeds is whether the material data, the simulation model and the fabrication process agree with one another. Designers who understand high-frequency substrate materials can predict behaviour, but the prediction only holds if the shop actually delivers the stack-up that was simulated. That is why the five technologies behind 5G antenna array impedance consistency read as one continuous argument rather than five separate techniques, and why Dk and Df are the parameters worth getting right before anything else is decided.
Frequently Asked Questions
Its loss is too high and its dielectric constant is too variable. At millimeter-wave frequencies both effects compound: the signal is attenuated before it arrives, and the phase and impedance move with temperature in ways the design cannot absorb.
No. Uniformity across the panel matters as much as the nominal figure, because local Dk variation breaks impedance continuity. A slightly higher but consistent Dk is easier to design with than a low average with wide swings.
It shields better, radiates less, and confines the field closer to the signal line, which reduces leakage between neighbouring structures in a dense module. It also makes series and shunt components easier to place.
At 30 GHz and above, copper surface roughness can push conductor loss above dielectric loss as the limiting factor. That is why PTFE laminates are normally specified with low-profile copper foil rather than standard foil.
The process, not the material. Drilling soft PTFE without tearing hole walls, activating the inert surface before plating, and holding tight line widths are the steps where most designs lose the performance the simulation promised.
Send us your stack-up, frequency plan and impedance targets and we will confirm the material and the process window before you release the layout.
Get a Fast Quote Fast Reply on WhatsApp