Four engineering advantages that explain the shift from FR-4 to high-frequency laminates — loss, Dk stability, impedance control and moisture resistance — and the six industries driving the change.
FR-4 epoxy glass laminate has been the default substrate for electronics for decades, and the reasons are sound: it is cheap, its process is mature and for most circuits it is more than good enough. The problem is that most circuits are not what the industry is now building. As designs move into high frequency and high speed, the weaknesses of FR-4 stop being a footnote and start being the limiting factor. High-frequency laminates were developed to remove that ceiling, and the demand for them follows directly from the advantages described below.
The table gives the electrical comparison in the form it usually appears in a design review.
| Material class | Df @ 10 GHz | Dk behaviour | Where it belongs |
|---|---|---|---|
| FR-4 (epoxy glass) | ~0.02 | Drifts upward with frequency; varies by batch and with humidity | Consumer products, low-frequency digital circuits, power supplies, standard control boards |
| PTFE with ceramic filler (RO3003 class) | ~0.001 | Flat across frequency and temperature | Radar, satellite links, low-loss microwave circuits |
| Ceramic-filled hydrocarbon (4350B class) | ~0.0037 | Stable, Dk 3.48, tight tolerance across the panel | 5G equipment, base stations, automotive radar, RF and microwave boards |
As a signal travels through a substrate it loses energy in two ways — in the dielectric and in the conductor — and high-frequency laminates are engineered to minimise both. This is the single clearest reason for the difference in demand.
The dissipation factor measures how much energy the dielectric absorbs and turns into heat. FR-4 sits around 0.02; a laminate such as Rogers RO3003 reaches roughly 0.001. At 10 GHz the dielectric loss of the high-frequency board can therefore be tens of times lower than that of the FR-4 board built to the same dimensions. On a radar front end or a base station amplifier, that difference is what allows a longer reach and a lower power bill from the same design.
High-frequency current concentrates near the conductor surface, so the texture of the copper becomes part of the loss budget. High-frequency laminates are normally clad with very low-profile or reverse-treated foil, which removes most of the roughness penalty that FR-4 boards carry by default. How the material platform and the foil choice fit together is the subject of our overview of high-frequency microwave PCB materials.
Lower loss converts into three practical benefits. Less energy is wasted as heat, so cooling requirements fall. At the receiver, weaker signals remain detectable, which extends range and improves communication quality. And because the loss is predictable across the band, the design margins do not have to absorb an uncertainty that grows with distance.
The dielectric constant determines impedance, phase and propagation speed, which makes its behaviour over frequency and environment as important as its nominal value. FR-4 has a specific set of weaknesses in this area.
Its dielectric constant drifts noticeably as frequency rises. It varies between production batches. And it responds to changes in temperature and humidity. Individually each is manageable; together they make the material unsuitable for a design where impedance is a specification rather than an outcome.
Two failures follow from it. Impedance moves away from the design value — a line intended for 50 ohm may sit well outside that — which produces reflections and a distorted eye diagram. Phase also drifts, which is fatal in power dividers and phase shifters where the circuit depends on a precise phase relationship; in a phased array the visible result is inaccurate beam steering. High-frequency laminates hold both values steady across frequency, batch and environment, so the design behaves the way it was simulated. The same reasoning underpins the material comparison in our article on how RO3003 and RO4003 differ in performance.
Even high-speed digital circuits such as PCIe and DDR benefit, and not because their clock rates are high: their extremely fast edges contain very high-frequency harmonics. A stable dielectric constant and a uniform dielectric structure let the fabricator build transmission lines with tight impedance tolerance, which is what keeps differential pairs matched, holds bit error rates down and preserves timing margin. On FR-4 the inherent variability puts a ceiling on this somewhere in the low gigahertz range.
FR-4 takes up moisture relatively readily. Water changes its dielectric behaviour, and during high-temperature assembly it can cause delamination or popcorning that ruins the board outright. High-frequency laminates absorb very little, which keeps performance stable in humid environments and keeps the assembly process predictable at volume. For equipment that has to work outdoors for a decade, this is as important as the loss figure.
A loose comparison makes the distinction memorable. FR-4 is a gravel road: cheap, easy to build, perfectly adequate for slow traffic, but rough and unpredictable once something fast uses it, and its condition changes with the weather. A high-frequency laminate is the paved surface of a racing circuit: expensive to lay down, but flat, consistent and built for the one thing it has to do. The phrase “far greater demand” describes requirements rather than volume, and keeping that distinction in mind prevents the wrong material from being specified in either direction.
Demand tracks the growth of the sectors that cannot be built on FR-4. Six of them account for most of it.
5G and 6G base stations use massive MIMO antennas, power amplifiers, filters, power dividers and phase shifters, all of which need low-loss stable laminates to work as designed. Satellite links operating in the Ka and Ku bands need the same properties in both ground stations and payload hardware, and RF front ends in general depend on them. Our article on RO4003 boards in communication equipment shows how those requirements translate into a material choice.
77 GHz anti-collision radar is one of the largest applications. The antenna board has to work at millimeter-wave frequencies, where loss directly determines detection range and resolution. V2X communication modules add a second high-frequency circuit per vehicle.
Airborne, naval and ground radar systems combine extreme requirements for power, accuracy and reliability. Electronic warfare equipment such as jammers and surveillance receivers adds wideband performance to the list, and missile guidance systems need RF circuits that keep working in environments no consumer product ever sees.
High-frequency probe cards sit between the tester and the wafer, and their performance sets the accuracy of the measurement. The RF modules inside vector network analyzers and spectrum analyzers do the same job for the instrument itself, so the laminate is part of the specification.
Inside 100G, 400G and 800G optical transceivers, driver and RF circuits increasingly use high-frequency laminates to cope with the data rate. Core processing boards and interface cards in high-end routers and switches follow the same path. Our notes on FR-4 against Rogers materials across cost, supply chain and market reality cover what that shift means commercially.
Advanced imaging systems rely on the material in two places: the RF coils in MRI equipment and the microwave sources in linear accelerators. Both need stability that consumer-grade substrates cannot provide.
None of this makes FR-4 obsolete. In cost-sensitive consumer electronics with modest performance requirements it remains the right answer, and it will stay the majority material by volume for the foreseeable future. The point is that the boundary has moved: above a few gigahertz, in radar and satellite hardware, and anywhere impedance is a specification, high-frequency laminates are no longer an upgrade but the working assumption. Choosing between them is a question of where the design actually sits relative to that boundary, which is the comparison set out in our article on the RO4000 series and, from the opposite direction, in our discussion of Rogers 4350B against FR-4.
Frequently Asked Questions
No. It remains the most economical choice for consumer electronics, low-frequency digital circuits, power supplies and standard control boards, and it will stay the majority material by volume. What has changed is the frequency and precision range over which it is the right answer.
FR-4 sits near a dissipation factor of 0.02, while a laminate such as RO3003 is around 0.001. On an identical geometry that makes the dielectric loss of the high-frequency board tens of times lower, which shows up as longer reach and less wasted energy.
Because impedance, phase and propagation speed are all calculated from it. A value that drifts with frequency, batch or humidity produces a circuit that no longer matches its simulation, and in phase-critical circuits the drift appears as inaccurate beam steering.
Often, yes. Fast edges contain high-frequency harmonics even when the clock rate is modest, and a variable dielectric constant makes tight impedance control impossible. Above a few gigahertz FR-4 stops being reliable for that purpose.
Because absorbed water changes the dielectric constant and the loss, and during reflow it can cause delamination or popcorning. A laminate that absorbs very little water keeps both its electrical performance and its assembly yield under control.
Send us your frequency plan, loss budget and production volume and we will confirm whether you need a high-frequency laminate, and which grade fits.
Get a Fast Quote Fast Reply on WhatsApp