ROG Blog
RSS

The Rog Blog is contributed by John Coonrod and various other experts from Rogers Corporation, providing technical advice and information about RF/microwave materials.

FR-4 Versus High Frequency Laminates

August 3, 2010

Microwave Journal

4 Comments

August 9, 2010

John Coonrod is a Market Development Engineer for Rogers Corporation, Advanced Circuit Materials Division. John has 23 years of experience in the Printed Circuit Board industry. About half of this time was spent in the Flexible Printed Circuit Board industry doing circuit design, applications, processing and materials engineering. The past ten years have been spent supporting circuit fabrication, providing application support and conducting electrical characterization studies of High Frequency Rigid Printed Circuit Board materials made by Rogers. John has a Bachelor of Science, Electrical Engineering degree from Arizona State University.

When does a high frequency laminate make more sense than FR-4?

Epoxy-based FR-4 circuit-board materials are popular for a wide range of electronics applications, and for good reason. These glass-reinforced substrates are reliable, low in cost, with well understood mechanical and electrical characteristics. They are used in everything from audio circuits through microwave designs. But they are not the answer for every circuit, especially for a wide range of high frequency designs, and for several reasons.

When trying to decide between FR-4 or a high frequency laminate, the lower cost of FR-4 may be important, but it may not be the only consideration. For example, how critical is signal loss to your design? In most passive circuits, such as filters, loss increases as frequency increases. FR-4 has a higher dissipation factor (Df) than laminates engineered for high frequency use, circuits fabricated on FR-4 will suffer higher losses than similar circuits formed on a high frequency circuit. Typical values are about 0.020 for FR-4 and about 0.004 for a high frequency laminate, or a dissipation factor that is about one-fourth that of FR-4. This translates into a meaningful difference in insertion loss between the two circuit-board materials. Some designs can tolerate a certain amount of loss, but it will depend upon the loss budget for a given project.

Another factor to consider when trying to choose between FR-4 and a high-frequency laminate is how those losses occur. Are they evenly distributed across the frequency band of operation? FR-4 is characterized by a Df that increases steadily with frequency, so as frequency rises, insertion loss increases. High-frequency laminates typically exhibit a more stable Df characteristic with frequency, with considerably less loss at higher frequencies.

Stable impedance is important to many designs, and this is another area where FR-4 and high frequency laminates can deliver much different results. Stable impedance depends on maintaining a stable dielectric constant (Dk) across the length and width of a substrate and also as temperature changes. While they are low in cost, FR-4 materials can suffer relatively wide variations in Dk across the length and width of a circuit board and with temperature. This can make it impossible, for example, to achieve flat amplitude response with frequency in an amplifier. High frequency laminates, on the other hand, provide a Dk that remains constant across the length and width of a board. This means more predictable performance in large circuits and more repeatable performance in multiple similar circuits formed from the same board.

For designs that must perform over temperature, FR-4 and high-frequency laminates can also provide much different results. To evaluate performance variations with temperature, circuit-board materials are characterized by the thermal coefficient of dielectric constant (TCDk). It is simply a measure of how much the dielectric constant changes over a defined temperature range. For FR-4, it is typically 200 parts per million (ppm) per degree Celsius (°C) change in temperature. While that might seem like a small number, it can result in large variations over a wide temperature range. Compare it to about 40 ppm/°C for high frequency laminates, a much smaller number resulting in much less variations with temperature. If a circuit must work with little variation over a wide temperature range, a high frequency laminate is probably a better choice than FR-4.

Sometimes, the very value of a circuit board’s Dk can play an important part in choosing a particular material. In a wavelength-dependent circuit, such as an RF or microwave design, the value of the dielectric constant will impact the size of the circuit’s transmission lines, and thus the size of the circuit. Simply put, circuit-board materials with higher Dk values can yield smaller circuits. If the size of a circuit is important, then the value of the Dk should be part of the decision process in choosing a circuit-board material. FR-4 typically has a Dk of about 4.5. That’s higher than most PTFE high frequency materials, but lower than a number of high frequency laminates which offer Dk values of 6.15 and even as high as 11.0. Using materials with these higher values can shave 25% or more off the size of a circuit built on FR-4.

There are several other factors to consider when deciding between FR-4 and a high frequency laminate, including the type of operating environment and the power/thermal requirements. For circuits where moisture might be a problem, for example, most high frequency laminates have much lower moisture absorption than FR-4, which can be critical in outdoor applications. Also, high frequency thermoset-based laminates may be more thermally robust than FR-4, and better able to handle the high temperatures of lead-free soldering processes.

What is the best way to decide if your project might work better with a high frequency laminate than with FR-4? Size up the electrical and mechanical requirements for the job, and see how they might be handled, first by FR-4 and then using a high frequency laminate. Data sheets are readily available for both types of circuit-board materials and provide the key facts on electrical and mechanical parameters and tolerances for ease of comparison. If the mechanical and electrical variations of FR-4 are too wide for your application, consider paying a bit more for a high frequency laminate, but with much improved performance and reliability. In the end, the higher yields from using a dielectric material with tighter manufacturing tolerances, such as RO4000® thermoset substrates might even mean lower production costs than when using FR-4.

This article doesn't really answer my question, which is, what is the (generally considered) cut-off frequency for FR-4 vs. the Roger's type of laminates. In other words, how high in frequency can one use FR-4 material and still be following good practices?

The question is important in cost sensitive designs, and especially if one is expecting to go through a series of design iterations because using FR-4 materials are cheaper and delivery times are quicker than using expensive microwave grade materials.

The article deals with losses, DK, and water absorption, which aren't really important factors in many applications. If one designs using the correct DK for the material being used then, theoretically, the design should function properly. While FR-4 is probably more lossy, the difference isn't hardly measurable, especially for a compact design (i.e., short line lengths) and at below 1 Ghz. Water absorption also isn't important except for exposed boards, like antennas.

In my experience there are other factors that creep in that cannot be easily controlled with FR-4 material, such as the strand orientation of the fibers which, at higher frequencies can have an effect that will change the performance from batch to batch, since most PCB fab houses don't pay attention to that. The Roger's material doesn't have that sort anisotropy.

From my own experience I decided, years ago, to use 500 MHz for a cutoff in my designs. But recently I've had a design at 450 MHz, which is right at that that threshold and so I'm wondering if the FR-4 material has improved any over the years or should I use the Roger's material, just to be safe? After all, I see computer boards boasting microprocessor speeds of a few GHz, yet appear to be made on FR-4 material. So have FR-4 materials gotten better, and hy should I re-visit that 500 MHz frequency design rule? What design rule are other people using in that regard?

In any case, this article did not give me any insight that I didn't already have. It also did not answer my question.

Edward Teyssier: You mention that "After all, I see computer boards boasting microprocessor speeds of a few GHz, yet appear to be made on FR-4 material." - but there is a very common misconception when it comes to computer microprocessor operation.

A "P.C. motherboard" may indeed be quoted to have a speed of 2GHz or GHz - however if one 'scopes any of the pins on the microprocessor chip itself (in practice that's nearly impossible of course) one will not find any GHz-range clock signals at all.

The reason that the board (and processor) is quoted to run at these frequencies is that they can perform many operations in one clock cycle. Some of these operations may be simultaneous (e.g. to make various data available quickly even if not reqired later in the cycle) and some may be sequential, but only take a few nanoseconds to perform.

In the "old days" microprocessors had a clock of perhaps 1MHz or 4MHz, and they could perform operations that would take up 2, 3 or 4 clock cycles, such was the speed of the silicon back then; so a 4MHz 68010 32-bit processsor chip actually had a 4MHz clock.

Microprocessor chips, like most integrated circuits, eveolved rapidly and we got to the time when a processor could do (e.g.) 10 operations in one external clock cycle - based on the use of a "clock multiplier" in the chip itself.

Important to note here - that 'external clock' - that is the clock signal fed along the board to the processor chip. A chip of today might receive a clock signal at, say, 100MHz - but on-chip there may be a 36-times multiplier which would allow the chip to perform operations at 3.6GHz. Note that the only real high-frequency (GHz-range) signals to be found are in the microprocessor chip itself. They are not on the board, although we often see 'lengthened' traces which are designed in to ensure that certain signals arrive a few pico or nano seconds later than others.

It's all down to the wonderful advances of science, engineering and terminology!

You must login or register in order to post a comment.

Edward Teyssier

May 4, 2021

Soundarya G

July 1, 2022

Nice article. Thanks for the clarification

F. C. Trevor Gale

January 31, 2024

Chandre Gowda

November 14, 2024

This article cleared my confusion on selecting laminate for our UWB application. Thank you for the details.