January 9, 2012

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. www.rogerscorp.com/acm. This blog is part of Microwave Journal's guest blog series.

Choosing a substrate material for use as the printed-circuit board (PCB) in a new design can be a stressful experience, unless it is a matter of sticking with that “tried-and-true” material that has worked so well in the past. But with ever-evolving and improving dielectric and laminate materials, and increasing demands to achieve high performance levels at reduced costs, most design engineers are wise to consider the cost-versus-performance benefits of different types of commercial PCB materials. Previous blogs have detailed how selecting a PCB material can be influenced by different performance requirements. This blog will explore the role that a computer-aided-design software tool plays in choosing the most suitable PCB material.

The technical design literature is filled with articles that document differences between simulated and measured results, with authors often blaming the unforeseen electrical effects of test fixtures or problems in fabricating the PCB prototype as the reasons for the discrepancies. The computer models are assumed to be accurate because they have been developed by experienced RF/microwave engineers and are based, as was seen in the last blog, on industry-standard test methods to characterize different parameters of the PCB substrate. But a CAD simulation is only as accurate as its models, and models are rarely perfect representations of the original.

One basic question to ask when considering different PCB materials for a new design is “How well is a material of interest represented in terms of models in circuit and electromagnetic (EM) simulation software tools?” For example, RT/duroid® 5880 from Rogers Corporation has often been cited in technical articles as a basis for modeled-versus-measured comparisons of new design circuits. Models for this substrate material have been developed by various sources, and simulated performance levels achieved with models of the material have been well documented as being in close correlation with measured results. Such a background tends to build confidence in this PCB material in terms of both simulation and prototype fabrication stages of a design project.

But what about other PCB materials? Most suppliers of CAD software tools can provide lists of the PCB materials that they support in terms of models. And if their models are not exact, a software user can typically create one to their liking by modifying the key defining characteristics of a PCB material model, such as relative dielectric constant, dissipation factor, coefficient of thermal expansion (CTE), and coefficient of dielectric constant. All software models of PCB material are based on one or more values of relative dielectric constant (ε_r), derived from one of the measurement approaches covered in the last blog for one or more test frequencies. These values may or may not agree exactly with the data-sheet values for ε_r provided by a material manufacturer. Rogers Corporation, for example, provides two sets of values for ε_r on its data sheets: one is a value meant for use with substrate manufacturing processes and one is for use with CAD simulators for circuit design. Experience has shown that having design values yields closer agreement between simulated and measured prototype results.

Quite simply, the accuracy of a PCB model is improved by accounting for as many electrical effects as possible. In the case of predicting insertion loss for a circuit design based on a particular PCB material, for example, many different factors are involved and one or two can easily be overlooked in a simulation. Insertion loss is affected not only by conductor losses, dielectric material loss, radiation losses, and leakage losses, but also by such factors as copper surface roughness, solder mask, plated finishes, and circuit configurations.

Solder mask is often added to a PCB to prevent unwanted connections between conductors. But it tends to be lossy at microwave frequencies. It is also high in moisture absorption and can impact PCB performance in high-humidity environments. Different types of solder mask can even have an effect on the ε_r and dissipation factor of a PCB. In short, a circuit with solder mask will tend to exhibit higher insertion loss than the same PCB material without solder mask, and this is an effect that must be included in an accurate PCB material model.

The type of surface finish used on a PCB, such as an electroless nickel immersion gold (ENIG) finish, can also play a role in the insertion-loss performance of circuits fabricated on that PCB. While an ENIG finish can add a level of protection to a circuit, it can also add to the insertion loss of that circuit, although the amount of additional loss is also a function of the type of circuit. A broadband circuit will tend to suffer higher losses due to surface finish than narrowband circuit.

Yet another factor that can impact the accuracy of a model used to predict loss in a circuit design is copper surface roughness. The effects of copper surface roughness can be difficult to predict, since they depend on the surface roughness profile and the operating frequency. Yet, as with the effects of solder mask and surface finish, copper surface roughness must be included in any PCB material model in hopes of achieving accurate simulations.

As a last word on selecting a PCB substrate, consider first whether a material meets all the electrical requirements for a design, how it rates mechanically, and where it fits within a budget. Then review how well the material is represented in terms of software model in the same commercial CAD simulator that will be used to help design a new circuit. Do these models account for as many electrical effects as possible? If not, is it possible to modify the model for a PCB material of interest to include such effects as solder mask, surface finish, and copper surface roughness? An accurate PCB material model should include as many variables as possible, to better ensure that those CAD simulations closely match measurements on a fabricated circuit prototype.

Do you have a design or fabrication question? John Coonrod and Joe Davis are available to help. Log in to the Rogers Technology Support Hub and “Ask an Engineer” today.