As fascinating and challenging as the PCB design process can be, it is important to take all necessary precautions to ensure proper circuit operation, especially when dealing with high power PCBs. As the size of electronic devices is continuously and progressively reducing, design aspects such as power supply and thermal management must be taken into due consideration. This article will present some guidelines that the designer can follow to design a PCB suitable for supporting high power applications.
Trace width and thickness
In principle, the longer the track, the greater its resistance and the amount of heat to be dissipated. Since the goal is to minimize power losses, in order to ensure high reliability and durability of the circuit, the recommendation is to keep the traces that conduct high currents as short as possible. To correctly calculate the width of a track, knowing the maximum current that can pass through it, designers can rely on the formulas contained in the IPC-2221 standard, or use an online calculator.
As for the trace thickness, the typical value for a standard PCB is about 17.5 µm (1/2 oz/ft2) for the internal layers, and 35 µm (1 oz/ft2) for the external layers and for the ground planes. High-power PCBs typically use thicker copper to reduce track width for the same current. This reduces the space occupied by the traces on the PCB. Thicker copper thicknesses range from 35 to 105 µm (1 to 3 oz/ft2), typically used for currents greater than 10 A. Thicker copper inevitably comes at an additional cost, but can help save space on the cards since, with higher viscosity, the required track width is much smaller.
The board layout should be considered from the very early stages of PCB development. An important rule that applies to any high-power PCB is to determine the path followed by the power. The location and amount of power flowing through a circuit is an important factor in evaluating the amount of heat the PCB needs to dissipate. The main factors affecting the layout of a printed circuit board include:
- power level flowing through the circuit;
- ambient temperature in which the board operates;
- amount of air flow affecting the board;
- material used for manufacturing the PCB;
- density of components which populate the board.
Even though with modern machinery this need is less urgent, in the changes of direction it is advisable to avoid right angles, but to use 45° angles, or curved lines, as indicated in Figure 1.
It is of fundamental importance to first establish the position on the PCB of high-power components, such as voltage converters or power transistors, which are responsible for generating a large amount of heat. High-power components should not be mounted near the edges of the board, as this causes heat build-up and a significant rise in temperature. Highly integrated digital components, such as microcontrollers, processors and FPGAs, should be positioned in the center of the PCB, allowing for uniform heat diffusion across the board and consequently a decrease in temperature. In any case, the power components should never be concentrated in the same area to avoid the formation of hot spots; rather, a linear type arrangement is preferable. Figure 2 shows the thermal analysis of an electronic circuit, with the areas with the highest concentration of heat highlighted in red.
Placement should start from the power devices, whose traces should be kept as short as possible and wide enough to eliminate noise generation and unnecessary ground loops. In general, the following rules apply:
- identify and reduce current loops, in particular high current paths.
- minimize the resistive voltage drops and other parasitic phenomena between components.
- place high power circuits away from sensitive circuits.
- take good grounding measures.
In some cases, it may also be preferable to place components on several different boards, as long as the form factor of the device allows to do it.
Proper thermal management is necessary to keep each component within safe temperature limits. The junction temperature should never exceed the limit indicated in the manufacturer’s datasheet (generally between +125 °C and +175 °C for silicon-based devices). The heat generated by each component is transferred to the outside through the package and the connection pins. In recent years, electronic component manufacturers have built increasingly thermo-compatible packages. Even with these package advances, heat dissipation becomes increasingly complicated as the size of integrated circuits continues to shrink.
The two main techniques used to improve PCB thermal management consist in the creation of large ground planes and in the insertion of thermal vias. The first technique allows you to increase the area available on the PCB for heat dissipation. Very often, these planes are connected to the upper or lower layer of the board to maximize the heat exchange with the surrounding environment; however, inner layers can also be used to extract part of the power dissipated by the devices on the PCB. Thermal vias are instead used to transfer heat from one layer to another layer on the same board. Their function is to direct heat from the hottest spots on the board to other layers.
Many of the components used in electronic circuits, such as regulators, amplifiers, and converters, are extremely sensitive to fluctuations in the surrounding environment. If they detect significant thermal variations, they can alter the signal they are producing, generating errors, and reducing the reliability of the device. It is therefore important to thermally insulate these sensitive components, so that they are not affected by the heat produced on the board.
Another technique used to allow a trace to carry larger amounts of current is to remove the solder mask from the PCB. This exposes the underlying copper material which can then be supplemented with additional solders to increase the thickness of the copper and decrease the overall resistance in the current-carrying components of the PCB. While it may be considered more of a workaround than a design rule, this technique allows PCB traces to withstand more power without requiring an increase in trace width.
When a power rail is distributed and shared between multiple board components, it is possible that the active components generate dangerous phenomena, such as ground bounce and ringing. This can cause voltage drops close to the component’s power pins. To overcome this issue, decoupling capacitors are used: one terminal of the capacitor must be placed as close as possible to the pin of the component receiving the power supply, while the other terminal must be connected directly to a low impedance ground plane. The goal is to reduce the impedance between the power supply rail and ground. Decoupling capacitors act as a secondary power source, providing components with the current they need during each transient (voltage ripple or noise). There are several aspects to consider when selecting a decoupling capacitor. These factors include selecting the correct capacitor value, dielectric material, geometry, and location of the capacitor relative to the electronic component. A typical value for decoupling capacitors is 0.1μF ceramic.
The design of high-power PCBs requires the use of materials with particular characteristics, first of all thermal conductivity (TC). Traditional materials, such as low-cost FR-4, have a TC of about 0.20 W/m/K. For high-power applications, where heat increases need to be minimized, it is preferable to use specific materials, such as Rogers RT laminate. With a TC value up to 1.44 W/m/K, this material handles high power levels with minimal temperature rises.
In addition to using materials that can handle power and heat with low losses, PCBs must be fabricated using conductive and thermal materials with very similar coefficient of thermal expansion (CTE), so that any expansion or contraction of the materials due to high power or temperatures occur at the same rate, minimizing mechanical stress on the material.
This article was originally published on EEWeb.