This post authored by John Coonrod, Technical Marketing Manager, and team originally appeared on the ROG Blog hosted by Microwave Journal.

Much of the buzz on the show floor at the 2017 IMS in Honolulu was about millimeter-wave devices and circuits. At one time, frequencies above 30 GHz were considered “exotic” and only for military or scientific applications. But times have changed, and available spectrum is scarce. Millimeter-wave frequencies are now used in commercial vehicular radars, and big plans are being made for these small wavelengths in Fifth Generation (5G) wireless communications networks, in support of moving massive amounts of data quickly. More and more design engineers are faced with developing practical millimeter-wave circuits to 77 GHz and beyond. But first, they must decide upon the best transmission-line technology for those high frequencies as well as the circuit material that can support those circuits with quality, low-loss signal propagation. Drivers and cell-phone users everywhere will be counting on them!

At microwave frequencies, microstrip is by far the most popular transmission-line technology, compared to stripline and coplanar waveguide (CPW). It has a signal plane on the top copper layer and bottom ground plane. It is relatively simple and cost-effective, and allows surface mounting of components for ease of construction.

Unfortunately, as signal frequencies move into the millimeter-wave range, microstrip circuits can behave like antennas, radiating electromagnetic (EM) energy away from a desired signal propagation path and resulting in much higher radiation losses than at lower frequencies. Microstrip radiation losses are also dependent upon the thickness and dielectric constant (Dk) of the circuit substrate material. Thinner substrates suffer less radiation loss than thicker substrates. Also, circuit materials with higher Dk values have less radiation loss than circuit materials with lower Dk values.

In microstrip, the effective Dk is a combination of the Dk of the substrate material and air, since EM waves in a microstrip transmission line propagate in part through the dielectric and in part through the air above it. In contrast to microstrip, stripline is like a flattened coaxial cable. It consists of a conductor surrounded by top and bottom dielectric layers which in turn are covered by ground planes. The Dk of stripline is the same as that of the dielectric material, since air is not involved in the propagation process.

CPW circuits are fabricated with a number of variations, including as standard, grounded coplanar waveguide (GCPW), and conductor-backed coplanar waveguide. Standard CPW metallizes parallel conductors (in the form of a flat waveguide) on the top of a dielectric layer, with ground metal areas just beyond the conductors. GCPW adds a bottom ground-plane layer but requires plated-through-hole (PTH) viaholes through the dielectric substrate material to connect the top and bottom ground planes. The extra ground planes on the top copper layer helps GCPW achieve high isolation between signal lines and can be designed to minimize spurious wave propagating modes. Placement of the PTH viaholes is critical, and can impact transmission-line impedance and loss.

Like microstrip, GCPW has an effective dielectric constant as the result of EM waves propagating through the dielectric material as well as through the air around the conductors. GCPW, like microstrip, also allows surface-mounting of components for ease of fabrication, in contrast to stripline where PTH vias need to connect the components on the outer circuit layers to the inner signal layer. In terms of millimeter-wave frequencies, GCPW has lower dispersion than microstrip, with less radiation loss, and is capable of supporting higher-frequency propagation than microstrip circuits. GCPW also achieves more effective suppression of spurious propagation than microstrip, and is more amenable to practical signal-launch configurations (such as from waveguide, cables, and connectors) at millimeter-wave frequencies than microstrip.

Finding the Right Circuit Material

If GCPW is the optimum transmission line for millimeter-wave circuits, it should then be fabricated on a circuit material with optimum characteristics for millimeter-wave frequencies. Since signal power tends to decrease with increasing frequency, an optimum circuit material for millimeter-wave circuits should have low loss at those high frequencies. The insertion loss of millimeter-wave transmission lines is due mainly to the aforementioned radiation losses, conductor losses, and dielectric losses. Radiation losses tend to be design-specific, whereas conductor and dielectric losses will depend upon the choice of circuit material.

Dielectric losses are a function of the type of dielectric material, and usually well defined by a material’s dissipation factor (Df), with lower values indicating lower dielectric losses. A circuit material capable of consistent performance at millimeter-wave frequencies will also exhibit minimal variations in Dk, so that dielectric losses do not change dramatically with frequency.

In considering circuit materials for millimeter-wave circuit applications, the thermal coefficient of dielectric constant (TCDk) parameter provides reliable insight into the stability of a material’s Dk with temperature. The TCDk parameter provides an understanding of what to expect of a particular circuit material’s performance at millimeter-wave frequencies, with lower TCDk values indicating less change of Dk with temperature and less variations in frequency phase response resulting from variations in Dk with temperature.

Conductor losses can be traced to a number of variables at millimeter-wave frequencies, including the surface roughness of the copper conductors and the choice of plated finish for the conductors. Copper is an excellent conductor, but increasing surface roughness results in increasing conductor loss and greater propagation phase delays. The main area of concern for copper surface roughness is at the copper-substrate interface, with conductor loss due to the copper surface roughness increasing as a function of increasing frequency. The small wavelengths of millimeter-wave signals result in less skin depth in the circuit material as part of EM propagation, and circuit materials with greater copper surface roughness will more severely impact the insertion loss and phase response at millimeter-wave frequencies. The effect of copper surface roughness on insertion loss is also dependent upon the thickness of the circuit material, with thinner circuits affected more by copper surface roughness than thicker circuits.

At millimeter-wave frequencies, circuit materials with excessive copper surface roughness will have more impact on the conductor loss of microstrip circuits than on the conductor loss of GCPW circuits. Switching to a circuit material with smoother copper finish will bring less of an improvement in conductor-loss performance for a GCPW circuit than for a microstrip circuit, especially at millimeter-wave frequencies. In particular, tightly coupled GCPW circuits, which feature closely spaced conductors and ground areas, are less subject to the effects of copper surface roughness than loosely coupled GCPW circuits (with greater spacing between conductors and ground).

An optimum circuit material for millimeter-wave circuits should cause minimal phase angle variations, since such behavior can be critical to many millimeter-wave applications, such as 77-GHz vehicular radar systems. By minimizing variations in certain material-based attributes, such as copper thickness, Dk, conductor width, and substrate thickness, variations in phase angles can be minimized at millimeter-wave frequencies.

Additional details on finding the right combination of circuit material and transmission-line technology are available as part of the Microwave Journal webinar, “Design Considerations and Tradeoffs for Microstrip, Coplanar and Stripline Structures at Millimeter-wave Frequencies,” presented by John Coonrod. 

ROG Mobile App

Download the ROG Mobile app to access Rogers’ calculators, including the popular Microwave Impedance simulation tool, literature, technical papers, and the ability to order samples of the company’s high performance printed circuit board materials.

Ask an Engineer

Do you have a design or fabrication question? Rogers Corporation’s experts are available to help. Log in to the Rogers Technology Support Hub and “Ask an Engineer today.

Tagged with:  

This post authored by John Coonrod, Technical Marketing Manager, and team originally appeared on the ROG Blog hosted by Microwave Journal.

High-frequency filters are increasingly essential components within wireless products, especially as those wireless products continue to compete for limited frequency spectrum. Various types of RF/microwave filters help wireless radio transmitters and receivers operate with their proper signals while shielding them interference caused by out-of-band signals. Printed-circuit filters can be designed with various responses, including bandpass, bandstop, lowpass, or highpass filters, and from a number of different transmission-line technologies, including microstrip, stripline, or coplanar-waveguide (CPW) transmission lines. For the best results, filter designers should start with a printed-circuit-board (PCB) material having optimum characteristics for RF/microwave filters. The choice of circuit material can not only impact a filter’s performance, but even the size of a printed circuit filter.

The job of a filter is to shape part of the frequency spectrum, ideally stopping unwanted signals while passing desired signals with minimal loss or attenuation. Each filter type performs these functions by means of different spectral regions: stopbands, passbands, and transitions between a stopband and a passband. For example, a lowpass filter has one passband in the lower-frequency portion of its frequency range and one stopband in the upper-frequency part of its frequency range, with one transition region between them. A highpass filter is the opposite, with one passband in the upper-frequency part of its range and one stopband in the lower-frequency part of its range, and one transition region between them. A bandpass filter has a passband, lower and upper stopbands, and two transition regions. A band-reject filter can be thought of as the opposite, with a stopband with transition regions linking upper and lower passbands.

Different transfer functions describe a filter’s transition regions. A Chebyshev filter, for example, is characterized as having an abrupt transition from the passband to the stopband; i.e., very little spectrum is required to make the change from the lowest signal loss to the highest signal attenuation. A filter with a Butterworth or binomial function, on the other hand, makes a more gradual transition from the passband to the stopband. It requires a greater amount of frequency spectrum to make the transition from filter regions, but it can also achieve a passband with low loss and very little ripple compared to a Chebyshev filter with its shorter transitions.

A filter’s frequency response is really a composite of the responses of its different spectral regions, with the transfer function having a major influence on the loss characteristics of the passband and stopband regions. A Chebyshev filter is capable of a quick, clean transition from a passband to a stopband, but at the cost of some amplitude variations or ripple in the passband insertion-loss response. A Butterworth filter can achieve a much flatter passband insertion-loss response, but less attenuation of signals at frequencies closer to the passband than a Chebyshev filter.

A printed circuit filter designer is faced with achieving a set of responses for a desired frequency range but also with trying to minimize transmission and reflection losses at the filter’s input and output ports by means of impedance matched junctions. The input and output ports are often coaxial connectors and most typically at a characteristic impedance of 50 Ω. What difference can the choice of circuit material have on a particular filter design and why use one type of circuit material rather than another?

When sorting through PCB material options prior to a design, a filter designer usually starts with dielectric constant (Dk) as a key parameter. PCB filters are typically formed of resonant circuit structures, such as the quarter-wave or half-wavelength resonators used in edge-coupled microstrip bandpass filters. The Dk of the dielectric material will determine the dimensions of the transmission lines required for specific resonator characteristics and frequencies. Circuit materials with higher Dk values will yield smaller filter resonator structures for a given wavelength and frequency, when miniaturization of a filter design is an important goal. In any case, for predictable, repeatable filter and resonator performance, the Dk of a circuit material choice should be as consistent as possible, held to the tightest tolerance possible.

What many filter designers may not realize when choosing a circuit material, however, is the anisotropy of the material—that is, the Dk value is different in the x-y plane of the material than in the z-axis (the thickness) which is the material Dk value often used as a starting point for filter computer simulations. Due to such anisotropic behavior, for proper modeling and design of a microstrip edge-coupled bandpass filter, the coupled fields in the x-y plane should be calculated as a function of the x-y Dk value. Alternatively, a filter designer may select a circuit material with more isotropic behavior to simplify the design process.

In general, circuit materials with lower Dk values are more isotropic than circuit materials with higher Dk values. To compare two commercial circuit materials, RO3003™and RO3010™circuit materials from Rogers Corp. exhibit low and high Dk values, respectively, with different degrees of isotropy. RO3003 laminate has a z-axis Dk value of 3.00 (with a tolerance of ±0.04 in the z-axis) and is nearly a true isotropic material, with similarly low Dk value in the x-y plane. Designing filters with coupled resonant structures, such as microstrip edge-coupled bandpass filters, is straightforward often with first-pass design success when using commercial computer-aided-engineering (CAE) circuit simulators.

However, for designing much smaller filter circuits for a given frequency, RO3010 circuit material has a much higher z-axis Dk value of 10.2 (with tolerance of ±0.30 in the z-axis). It is much more anisotropic than RO3003 material, with Dk value in the x-y plane that is much closer to the 3.0 range of the RO3003 material. This means that filter design strategies and computer simulations must account for the significant difference of Dk values in the x-y plane and the z-axis of RO3010 material. But the higher Dk value of this material significantly increases the coupling between resonant structures, which can help improve the overall performance of a filter design while miniaturizing its dimensions.

rog-mobileNote: Those interested in learning more about how circuit material anisotropy can impact filter design see the ROG Blog, “Substrate Anisotropy Affects Filter Designs,” which also examines the effects of moisture absorption on circuit material Dk.

ROG Mobile App

Download the ROG Mobile app to access Rogers’ calculators, including the popular Microwave Impedance simulation tool, literature, technical papers, and the ability to order samples of the company’s high performance printed circuit board materials.

Ask an Engineer

Do you have a design or fabrication question? Rogers Corporation’s experts are available to help. Log in to the Rogers Technology Support Hub and “Ask an Engineer” today.

 

 

This post authored by John Coonrod, Technical Marketing Manager, and team originally appeared on the ROG Blog hosted by Microwave Journal.

VIDEO: High Frequency Circuits Which Bend and Flex

Flexibility can be an important feature for printed-circuit boards (PCBs). Not all circuits are planar; some may need to be bent once to fit a particular product design while some might need to undergo continuous flexing as part of an application. Not all circuit materials are created equal, and some are more mechanically flexibile than others and can survive a certain amount of bending and flexing without damage. Understanding what makes a circuit material capable of bending and flexing, and what happens to it when it is bent or flexed, helps when specifying circuit materials for such uses.

Circuit boards are composites of different materials, such as conductive metals and dielectric materials, each with its own mechanical properties. The material stackup will depend on the type of circuit and the number of circuit layers. As more different materials are combined to form a PCB, especially in multilayer PCBs, the task of predicting the effects of bending and flexing becomes more complex. A key material parameter in determining how well a particular material will bend and flex is the modulus or stiffness of the material, with some of the composite materials of a PCB significantly stiffer, or with much higher modulus values, than others.

For example, the metallization in an RF/microwave PCB, primarily copper, will essentially determine the limits of flexibility in a circuit board since it has the highest modulus value of the material stackup, at 17,000 kpsi. Compare this to the much lower modulus values of dielectric materials, such as polytetrafluoroethylene (PTFE) with ceramic filler, at 300 kpsi, or PTFE with microfiber glass filling, at 175 kpsi. In a typical microstrip circuit, with conductor layer, dielectric, and ground-plane layer, the dielectric layer offers great flexibility but the top and bottom metal layers will set the limits of bending and flexibility for the composite structure.

Since high-frequency circuit boards are composite structures, the differences in flexibility of the component materials must be considered to determine how much bending and flexing a circuit board can withstand without damage to the stiffest of its material components, the metallization layers. This can be done by treating a PCB as if it were a beam being bent, with a certain bend radius depending upon the stiffness of the beam. A rubber beam will bend much more easily than a higher modulus metal beam, and be capable of enduring a much smaller bend radius without cracking. A PCB considered as a beam will also have a certain bend radius depending upon the overall stiffness of the composite group of materials, with the metallization layers setting the limits on the flexibility and minimum bend radius of the circuit board.

As with a beam, when a PCB is bent into a section of an imaginary circle, with a bend radius for that circuit, strain is placed on different parts of the beam and the PCB, with tension on the outer side and compression on the inner side of the bend radius. Between the areas of tension and compression lies an almost infinitely thin transition area or neutral axis with no strain. The strain increases as the distance from the neutral axis to the tension or compression plane increases. In a balanced circuit board, the neutral axis would lie at the geometrically center of the circuit board.

Stress from tension and compression works in different ways on a PCB’s materials, with tension pulling materials apart and compression squeezing them together. For a PCB with microstrip circuitry and copper conductors on the outer bend radius, this means that the stiffest or highest-modulus material in the composite PCB is being subjected to a certain amount of tension that will increase as the bend radius is made smaller. At the same time, the bottom ground plane is also being stressed and subjected to compression. Both forms of stress, if excessive, can lead to cracks in a microstrip circuit’s metallization layers. In addition, stress occurs at the interfaces of materials with different modulus values, such as the intersection of the copper conductor layer and the dielectric layer. Cracks from stress can start at the interface and work through the copper layer. To minimize damage to the metallization layers and ensure reliability in bent and flexed circuit boards, the key is to determine the amount of stress that a particular PCB can endure without cracking the metal layers.

The amount of stress on a PCB from bending and flexing is not simply a matter of knowing the modulus of the stiffest material component but in knowing how the PCB is constructed. For example, in a multilayer circuit board, differences in the thicknesses of the dielectric layers can cause increased amounts of strain when the circuit is bent. Each layer of a multilayer circuit structure will have its own modulus, and the structure will have a modulus as a whole. Since copper is the stiffest material component of most microwave circuits, the thickness of the copper and the percentage of copper in the entire PCB material stackup will contribute a great deal to the overall modulus and flexibility of the PCB as a whole.

Even the type of copper can determine the flexibility of a microwave circuit. Due to differences in the grain structures of rolled copper and electrodeposited (ED) copper, rolled copper is typically better than ED copper for PCBs that must be bent or flexed. For applications that may call for ED copper, some special types of ED copper are available for better bending and flexing than standard ED copper.  In addition, finishes added to copper conductors, such as electroless nickel/immersion gold (ENIG) plating, can add a high modulus to the overall PCB modulus, limiting the amount of bending and flexing that a PCB can safely endure.

Different microwave circuit constructions will present different bending and flexing capabilities. Stripline, with copper conductors sandwiched between upper and lower dielectric layers, is inherently better equipped for bending and flexing than microstrip. The signal conductor layer in a typical stripline construction is at or close to the neutral axis for minimum stress; however, the outer ground planes will typically have high stress.

General guidelines to avoid damage when bending or flexing circuit materials pertain to single-bend and dynamic flexing situations. When a single bend is required, the bend radius should be at least 10 times the thickness of the circuit so that the strain on the circuit layer is 2% or less. For dynamic flexing, strain should be held to less than 0.2% for more than 1 million flex cycles and less than 0.4% for 1 million or less flex cycles.

Readers wishing to learn more about how to model stresses placed on PCBs from bending and flexing are invited to view John Coonrod’s MicroApps presentation, “High Frequency Circuits Which Bend and Flex,” from the 2016 IEEE International Microwave Symposium (IMS). The presentation provides circuit bending prediction models and includes a microstrip example using ½-oz. rolled copper on 5-mil-thick RO3003™ laminate material from Rogers Corp.

ROG Mobile App

Download the ROG Mobile app to access Rogers’ calculators, including the popular Microwave Impedance simulation tool, literature, technical papers, and the ability to oScreen shot 2014-08-08 at 1.33.54 PMrder samples of the company’s high performance printed circuit board materials.

Ask an Engineer

Do you have a design or fabrication question? Rogers Corporation’s experts are available to help. Log in to the Rogers Technology Support Hub and “Ask an Engineer” today.

This post authored by John Coonrod originally appeared on the ROG Blog hosted by Microwave Journal.

Millimeter-wave circuits were once considered exotic and only used for specialized applications, typically in the military space. For one thing, frequencies with such small wavelengths, from about 30 to 300 GHz, required special components and circuits scaled to those diminutive wavelengths. But lower-frequency bands are being consumed by a growing number of wireless applications, and millimeter-wave frequency bands are looking more and more attractive for communications systems of the future. Such high frequencies have even been proposed as part of an emerging fifth-generation (5G) wireless standard that will be challenged to connect billions of global Internet of Things (IoT) devices by means of available wireless bandwidths. Millimeter-wave bandwidths have long been employed for military radar systems and are increasingly being used in commercial automotive collision-avoidance radar systems. Achieving millimeter-wave circuit designs on reliable printed-circuit-board (PCB) materials in a practical manner will be the challenge in making these higher frequencies affordable. Substrate-Integrated-Waveguide (SIW) circuit technology may just be the solution.

As noted in an earlier blog (“Make Waveguide in Planar PCB Form”), SIW structures are essentially waveguide in planar form, with the capability to support millimeter-wave signals with relatively low loss even at those higher frequencies. SIW technology offers improved performance at millimeter-wave frequencies compared to traditional transmission-line technologies, such as microstrip, stripline, and even grounded coplanar-waveguide (GCPW) approaches, with limitations at millimeter-wave frequencies.

SIW has often been described as a form of transition between microstrip and dielectric filled waveguide (DFW). SIW can be fabricated with many of the same methods as microstrip. At millimeter-wave frequencies, however, microstrip circuits require small features and extremely tight machined tolerances to support the transmission of such high frequencies. In addition, at millimeter-wave frequencies, SIW circuits do not exhibit radiation losses suffered by microstrip. In fact, SIW circuits in general do not have the potential problems with electromagnetic interference (EMI) of the other transmission-line formats. SIW technology provides the means to realize extremely compact components; it is suitable for passive components, such as filters, but has also been used as active components, such as oscillators, at microwave through millimeter-wave frequencies. Commercial EM simulation software is most often used to aid in the design, simulation, and optimization of SIW circuitry, and such software programs can effectively model the effects of the dielectric substrates used as the foundations for SIW circuitry.

In forming SIW transmission lines, a rectangular waveguide is created within a substrate, usually on circuit-board material such as RO4350B™ LoPro® laminates from Rogers Corp. which has a low relative dielectric constant of 3.48 in the z-axis (through the thickness) measured at 10 GHz. This low-loss circuit material, which is widely used as the foundation for wireless base-station power amplifiers, features properties well suited to SIW circuits. It can be fabricated with the methods used for FR-4 circuit materials, to maintain low production costs.

SIW circuits and their dielectric-filled waveguide transmission lines are formed on a circuit material such as RO4350B LoPro laminate by adding a top metal plane over a laminate with a ground plane, then fabricating rows of conductive plated viaholes on both sides along the length of the substrate material. These plated-through-hole (PTH) viaholes are used to make the sidewalls of the rectangular waveguide structure formed on the PCB material. In forming the SIW embedded waveguide structure, more conductive metal is actually used than in stripline or microstrip transmission lines for similar wavelengths, resulting in less conduction loss at microwave and millimeter-wave frequencies.

What is critical in the fabrication of SIW circuits is the formation and spacing of the viaholes. Close spacing yields less conduction loss through the use of more conductive metal to form the waveguide structures, but also results in longer and more complex production times in fabricating the SIW PCBs. Wider spacing can save production time, but can also raise conduction losses and can result in higher EM leakage losses because of the wide spacing. SIW circuits will also suffer dielectric losses (as will all high-frequency circuit formats), which are dependent upon the choice of circuit laminate material. For SIW circuits, whether at microwave or millimeter-wave frequencies, and really with all high-frequency circuits, PCB materials should be chosen wisely for optimum balance between performance and cost.

One PCB material parameter that is critical for SIW reliability is coefficient of thermal expansion (CTE) which gauges the expansion of a circuit material with elevated temperatures. The SIW viaholes are plated through holes (PTHs) through the dielectric PCB material, and high values of CTE, which denote excessive expansion with temperature, will result in undue stress on the sidewalls of the PTHs. Circuit materials, such as RO4350B LoPro noted previously, and RO4835™ LoPro material from Rogers Corp., with stable CTE characteristics, are ideal candidates for high-reliability SIW circuits. RO4835 LoPro circuit materials have been used for years in the fabrication of multilayer circuits with high layer counts, relying on PTHs for interconnection of those many layers. As an added benefit for creating cost-effective SIW circuits, both RO4835B LoPro and RO4835 LoPro materials can be fabricated with standard FR-4 epoxy/glass processes to help minimize production costs.

At millimeter-wave frequencies, SIW circuits exhibit low loss similar to their larger mechanical waveguide descendants, and considerably less than the other, more conventional transmission-line formats, such as microstrip, stripline, and GCPW. But SIW circuits also share other traits of larger mechanical waveguide, including a lower-frequency cutoff point. As with mechanical waveguide, SIW circuits are designed for particular operating frequencies and bandwidths depending upon the circuit dimensions, and designers must be aware that they will be working with a lower-frequency (and upper-frequency) cutoff point and a target low-loss passband. But SIW circuits can also work quite well with traditional transmission-line formats, with microstrip and GCPW transmission lines serving as fairly simple and effective feedlines from other parts of a circuit to the SIW circuitry.

With the expected growth in the demand for millimeter-wave circuits, especially with the expansion of IoT and higher-frequency automotive applications, SIW technology appears quite attractive as an effective and practical solution for designing and fabricating affordable millimeter-wave circuits. In particular, SIW technology provides the means to realize miniaturized planar antennas as millimeter-wScreen shot 2014-08-08 at 1.33.54 PMave frequencies, perhaps in volumes reaching millions of components as needed for these many emerging wireless applications.

ROG Mobile App

Download the ROG Mobile appto access Rogers’ calculators, including the popular Microwave Impedance simulation tool, literature, technical papers, and the ability to order samples of the company’s high performance printed circuit board materials.

Ask an Engineer

Do you have a design or fabrication question? Rogers Corporation’s experts are available to help. Log in to the Rogers Technology Support Hub and “Ask an Engineer” today.

This post authored by John Coonrod originally appeared on the ROG Blog hosted by Microwave Journal.

Plated finishes are necessary additions to printed circuit boards (PCBs). Not only do PCB finishes provide smooth, solderable surfaces for attaching components, they also provide protection for a PCB’s copper conductors. Without such protection, a PCB’s conductive copper would quickly oxidize and deteriorate when exposed to the environment, resulting in degraded circuit performance. The added protection provided by a PCB’s plated finish means added loss, however. The choice of plated finish can make a real difference in a PCB’s conductive loss, especially for broadband, high frequency circuits. To better understand the loss performance of different plated finishes, various transmission lines were fabricated on different circuit laminates and different plated finishes applied.

PCBs, with or without a plated finish, suffer losses that typically increase with increasing frequency. The losses that circuit designers measure on a microwave transmission line, such as microstrip, stripline, or grounded coplanar waveguide (GCPW), stem from a combination of signal losses from the PCB, including conductor loss, dielectric loss, radiation loss, and leakage loss adding up to form insertion loss. Circuit design also contributes to the loss performance: Achieving good impedance matching along transmission lines, at circuit junctions, and at component mounting points helps to minimize signal reflections and losses from those reflections, often measured in a circuit’s transmission lines as return loss.

Copper is an excellent conductor, with low insertion loss for transmission lines and cables formed from copper. But the copper on the dielectric materials of PCBs does not always offer the smoothest, most level surface for mounting miniature circuit components, such as those in ball-grid-array (BGA) housings or tiny surface-mount-technology (SMT) packages. A plated finish can provide that smooth mounting surface for miniature components, and it can deliver long-term protection against copper deterioration. Some finishes also protect the plated through holes (PTHs) that serve as the electrical connections between different circuit layers in multilayer PCBs. Unfortunately, most PCB finishes come with a price, since most increase insertion loss to some degree, depending upon frequency and other factors, including the thickness of the substrate, choice of transmission-line technology, and layout of the circuit and how it is affected by the finish.

Most plated PCB finishes are less conductive than the copper conductors formed on the PCB’s dielectric material, and will suffer more loss than copper, especially at higher frequencies. The exception is silver, an excellent conductor, which is also expensive and usually applied in a very thin layer as a finish. PCB conductor losses are frequency dependent mostly due to the manner in which RF current uses a conductor. At lower frequencies, the RF current will use more of the conductor. At higher frequencies, the RF current tends to flow along the surface of the conductor, using only the outside skin of the conductor. Conductor loss rises as the RF current uses less of the conductor, and because of these skin effects at higher frequencies, plated finishes can have greater impact on PCB insertion loss at higher frequencies.

The impact of a plated finish on PCB insertion loss can also depend on the transmission-line technology. For example, for microstrip, with high current density along the edges of the conductor, the plated finish can have significant impact on conductor loss. For GCPW with current density distributed along the four edges of the ground-signal-ground conductor, the plated finish will have more impact on conductor loss.

Finding a Finish

A number of different plated finishes are available for high-frequency circuit boards, including electroless-nickel-immersion-gold (ENIG) finish, organic surface protectant (OSP), electroless nickel, electroless palladium, immersion gold (ENIPIG) finish, and soldermask finish. For example, for an ENIG finish, nickel is plated onto a PCB’s conductive copper, serving as a barrier between the copper and a thin layer of gold that is applied thereafter. The thin layer of gold, an excellent conductor, typically disappears and is absorbed into soldered connections as components are soldered onto the PCB’s transmission lines and conductive traces. As might be apparent from the materials used in this finish, it is expensive, but it is RoHS compliant and provides excellent protection for PTHs in multilayer circuit assemblies.

PCB finishes using OSP are popular as environmentally sound, “green” PCB treatments that are lead free and provide extremely flat mounting surfaces for components. This low-cost finish is applied by means of a chemical bath process and it is a very low-cost finish, but it is not well suited for PTH protection and there is no way to measure the thickness of the finish when trying to evaluate the reliability of the finishing process. Additionally, OSP is typically considered a temporary finish and not a permanent, final finish, although in an optimum environment it may have extended life. Soldermask is a polymer material that provides a protective coating for copper traces and prevents solder from making unwanted connections and short circuits.

Comparing Insertion Loss

How do the different plated finishes compare in terms of PCB conductor loss and insertion loss? By fabricating some circuits with different types of transmission lines on some standard PCB laminates and using different plated finishes, it was possible to compare the impacts of different finishes on insertion loss by means of measurements and computer simulations. For example, with microstrip and GCPW transmission lines on  RO4003C™ laminates from Rogers Corp., measurements revealed significantly less loss for microstrip with bare copper than for microstrip with an ENIG finish. However, measurements also revealed that more difference in loss existed for GCPW with bare copper than for GCPW with an ENIG finish.

When circuits were fabricated on different thicknesses (6.6, 10.0, and 30.0 mil thick) of RO4350B™ laminates from Rogers Corp., the total insertion loss tended to be less for the thicker materials. Thinner circuits are dominated more by conductor losses than other losses and, for each plated finish evaluated, it added to the PCB’s conductor losses.

With yet another circuit material evaluated during these plated finish tests and simulations, 5-mil-thick RT/duroid® 6002 circuit laminates using rolled copper from Rogers Corp., significantly higher conductor losses were found for microstrip circuits with ENIG plated copper conductors than for microstrip circuits with bare copper conductors, when tested at frequencies through 40 GHz. However, when the copper conductors for the same material were plated with immersion silver, little difference in conductor loss was found between microstrip circuits with bare copper and those with immersion silver plating, even for frequencies through 100 GHz (using a differential measurement method). For the same circuit material, little difference was found for microstrip circuits with bare copper conductors and with OSP copper conductors, even through 100 GHz. When soldermask was evaluated for this circuit material, microstrip circuits with bare copper conductors exhibited considerably less loss than copper conductors with soldermask.

In short, the lowest conductor losses, with microstrip and GCPW, are achieved using bare copper conductors. But it is not realistic to fabricate reliable PCBs with bare copper conductors, and plated PCB finishes provide much-needed long-term protection. As was discovered from measurements and simulations, all PCB plated finishes are not the same, with some suffering less loss than others. For measurements on high-frequency PCB materials through 110 GHz, circuits with bare copper conductors have the least conductor loss, followed by circuits with immersion tin, ENIPIG finish, and then ENIG finish.

Screen shot 2014-08-08 at 1.33.54 PMAuthor’s Note: This ROG blog is based on a presentation at PCB West 2015 Conference & Exhibition, “Wideband Insertion Loss Testing of Multiple PCB Final Plated Finishes.” In that presentation, considerable broadband, high-frequency test data and computer simulations were compared for different finishes on several commercial circuit laminates from Rogers Corp.

ROG Mobile App

Download the ROG Mobile app to access Rogers’ calculators, including the popular Microwave Impedance simulation tool, literature, technical papers, and the ability to order samples of the company’s high performance printed circuit board materials.

Ask an Engineer

Do you have a design or fabrication question? Rogers Corporation’s experts are available to help. Log in to the Rogers Technology Support Hub and “Ask an Engineer” today.

 

Page 1 of 3123