August 3, 2011
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.
Temperature effects on a printed-circuit board (PCB) can make it difficult to achieve target performance goals, even with the best PCB substrate materials. Modeling these effects takes imagination—to visualize different sources of heat, for example, and thermal paths where the heat might travel. It also requires an understanding of both thermal-mechanical and electromagnetic (EM) relationships to account for the assortment of variables that can influence PCB performance with changing temperatures. As a result, modeling thermal effects on PCB performance combines predictions provided by the heat diffusion equation as much as from Maxwell’s equations for EM fields.
In a high-frequency circuit, heat can come from the environment (the ambient temperature) or from DC and RF sources: from the flow of supply current, for example, or from RF input power to the circuit or power generated by an active device, such as a power transistor. Ideally, heat will flow away from the circuit, without creating any “hotspots.” As seen in the last blog, excessive temperatures can cause damage to a PCB’s dielectric material, and even cause copper transmission lines to delaminate from the dielectric material. Creating models that predict thermal effects on PCBs can help prevent creating inadvertent hotspots and potential damage to a circuit.
Commercial software tools traditionally have provided either thermal-mechanical modeling or EM modeling, but not both. Due to acquisitions and partnerships, the capabilities of some of these software tools can be linked, enabling EM modeling tools, for example, to exchange files for analysis with thermal-mechanical modeling tools.
Even when using such powerful software tools, it is helpful to have a clear understanding of the potential thermal paths in a PCB and the variables that should be included when modeling a PCB’s thermal behavior. A common assumption is that the majority of heat from a heat source such as a DC current or RF power transistor will flow through the PCB’s dielectric material (through its thickness or z-direction) for dissipation within the ground plane and package. But there are a number of other thermal paths in a typical PCB. For example, in a microstrip circuit, the copper in the transmission lines and bias (power-supply) lines has far greater thermal conductivity than the PCB’s dielectric material and will more readily conduct heat than the dielectric material. Heat from a source will flow along these lines sooner than through the dielectric, eventually dissipating in other components connected to the lines or through radiation into the air above the lines.
Dielectric PCB materials are characterized by a number of parameters related to temperature, including coefficient of thermal expansion (CTE) in the z-direction (through the dielectric material to the ground plane), the glass transition temperature (Tg), and the decomposition temperature (Td). In estimating thermal effects, any model should also consider the amount of current and/or RF power applied, the expected current density, the thermal conductivity of the transmission lines (usually copper), the thermal conductivity of the substrate material, the dissipation loss of the substrate, the insertion loss of the transmission lines (where high losses at high power levels equates to generated heat), reflection loss, radiation loss, the thicknesses of the conductors and dielectric material, even the geometry of the circuitry.
What are desirable PCB material characteristics when operating with higher levels of current or RF power? Simply put, dielectric materials rated for higher temperatures can generally handle higher DC or RF power levels than materials characterized for lower-temperature operation. For example, FR-4 is a popular PCB substrate material, but it is characterized by low thermal conductivity of 0.2 W/m/K—a warning sign against its use in circuits with high DC or RF power. (Some PCB material data sheets may also provide a material’s thermal resistance, in °C/W.) FR-4 has a CTE in the z-axis of typically 175 ppm/°C and its Tgcan range from +110 to +135°C, The loss tangent is typically 0.020 at 1 MHz and 0016 at 1 GHz. Typical values of FR-4 relative dielectric constant are 4.70 at 1 MHz and 4.34 at 1 GHz. Dielectric constant can play a role in thermal modeling since it will determine the dimensions of a transmission line, such as microstrip, required to achieve a given impedance, such as the 50 Ω common to high-frequency circuits.
In contrast, a material engineered for high-power, high-temperature use, RT/duroid® 6035HTC high-frequency laminate from Rogers Corporation, exhibits high thermal conductivity (1.44 W/m/K) and low loss tangent (0.0013 at 10 GHz) to minimize the effects of heat. The laminate is characterized by a CTE in the z-axis of typically 39 ppm/°C, a Tgof greater than +280°C, and has a relative dielectric constant of 3.5 at 10 GHz.
RO4350B™ and RO4360™ materials feature a Tgof greater than +280°C. For applications requiring a lower dielectric constant, RO4350B laminate has a value of 3.48 at 10 GHz, with thermal conductivity of 0.62 W/m/K, z-axis CTE of 35 ppm/°C, and loss tangent of 0.0037 at 10 GHz. For applications requiring a higher dielectric constant, RO4360 laminate features a typical value of 6.15 at 10 GHz, with thermal conductivity of 0.80 W/m/K, z-axis CTE of 30 ppm/°C, and loss tangent of 0.0038 at 10 GHz.
Commercial modeling tools can provide fairly accurate estimations of PCB performance when fueled with the right input parameters, such as many of the characteristics mentioned above. But such program require a great deal of computing power, and often computing time, to calculate PCB thermal effectives for such a large number of variables. In estimating the effects of high power in high-frequency circuits, it is usually better to be conservative, and some modeling approaches allow some quick “ball-park” calculations of thermal effects based on the key thermal parameters. One such approach can be found in a two-page design note (note 3.3.2) from Rogers Corporation, “Temperature Rise Estimations in Rogers High Frequency Circuit Boards Carrying Direct or RF Current.” It is available for free download from the Rogers’ website at
The note provides calculations to estimate the heating of a microstrip circuit for both DC and RF power, and can be used in conjunction with Rogers’ free Microwave Impedance (MWI) Calculator to predict the thermal effects on a wide range of microstrip circuits.
For those planning to attend the upcoming PCB West show (www.pcbwest.com) scheduled for the Santa Clara (CA) Convention Center on September 27-29, 2011, the author will delve deeper in PCB thermal management as part of a one-hour presentation, “Using High Frequency PCB Laminates for Improving Thermal Management Issues,” scheduled for Wednesday, September 28th.