Considerations For Thermal Management When Using Open Frame Power Supplies In OEM Applications

Abstract

As a system-level electronics designer, you have defined the functionality of your product, i.e., the inputs and outputs, the logic and operations as well as user interface. You have considered your electromagnetic interference (EMI) requirements and even determined what your power requirements will be. It is important at this stage that you begin to consider the thermal environment that your device will be operating in and the resulting internal temperatures that the components of your product will be subject to. All the real-world electrical components in your product, as a consequence of being non-ideal, will generate heat. Excessive heat can have a dramatic affect on the functionality, reliability and safety of all electrical products. This paper discusses thermal concepts and management techniques to help the system-level designer successfully integrate multiple heat-generating electronic components into a system. In particular, the paper addresses the system level integration of an open-frame switching power supply.

Introduction

All electronic devices and components generate heat as current passes through them. As the materials within these devices and components undergo changes in temperature, not only is their electrical performance affected, but other less obvious changes take place that can have negative affects on the system. When designing a system using an open frame power supply, particular care must be taken. Not only are power supplies sensitive to high temperature operation, but they are also a large source of thermal energy.

Thermal Circuit Theory

Heat (Q) is a flow of energy transferred from a system to the environment or vice versa. This thermal energy is the “collective kinetic and potential energies associated with the random motions of the atoms, molecules, and other microscopic bodies within an object.” (Halliday 1997, p.461). If the system and environment are at equal temperatures, they are said to be at thermal equilibrium, and heat transfer will not take place. If the temperatures are different, thermal energy will transfer from the higher temperature system to the lower temperature system in order to reach equilibrium.

Heat can be transferred via three different mechanisms: Conduction, Convection, and Radiation. In electric circuits there is a linear relationship given by Ohm’s Law, V = IR, for electric conductors. Resistance (R) in electric circuits acts to impede the flow of current (I). Thermal resistance (θ) in thermal circuits acts to impede the flow of thermal energy (Q). θ is a function of the component’s geometry and thermal conductivity. The following table summarizes the circuit analogy.

Each component in the system has different material properties and thus a different thermal resistance. Like a current in an electric circuit, heat flows in the path of least resistance. Like resistors, thermal resistances in series sum up:

The parallel relationship is also analogous:

This theory is useful in understanding heat removal, which we will discuss later.

Thermal Planning

To begin thermal planning, there are several issues that are important to consider; these are:

Subsystem Boundaries
Functionality (over applicable temperature range)
Safety
Reliability
Consequences of component failure(s)

In the early stages of developing a product specification, the operating temperature range was most likely determined. It is then critical to understand what the maximum desired internal ambient temperature of the product will be. This will most likely vary among the subsystems and other components, and so defining the subsystem boundaries is critical. For thermal considerations, a subsystem boundary should be drawn around components with similar specified temperature ranges in close proximity, around total subsystems with a collectively specified operating temperature, or between subsystems or components with significant differences in their thermal energy dissipation.

Within a subsystem, one should know what the maximum operating temperatures of that subsystem or other components are, and by applying some margin of safety to those temperatures, find the desired maximum internal temperature. For example, if we define a subsystem as the total volume envelope around an open frame power supply with a maximum operating temperature of 50 oC, and apply 10% safety margin, one should target a maximum internal ambient temperature of 45 oC.

Problems may arise however, when one discovers that the target temperature for another subsystem is 70 oC, and that system also dissipates a lot of thermal energy. It may become difficult to keep the air temperature around the open frame power supply below 45 oC. The easy solution would be to physically separate these subsystems, however, that may not always be possible, or completely effective. In this case, one must consider how to remove excess thermal energy, which will be discussed later.

It is also important to note that a common but not necessarily valid practice is often employed in these circumstances. Many designers, knowing that their open frame power supply will be subjected to temperatures beyond the designed limits, will attempt to derate the power supply, and choose a supply designed to output more than the required power. For example, if the power requirements are 50 Watts, within a 70 oC environment, some designers may choose a power supply rated for 80 Watts at 50 oC operation. As will be discussed later this can cause problems down the road.

Safety Considerations

Beyond the functionality of the product, it is important to understand the safety considerations of the product’s temperature. The subsystems and components may be well within the desired temperature range, but surface temperatures of the product may still exceed those considered safe for use by whatever industry application the product is intended. These acceptable temperatures are defined by the safety standards applied to the device. Most common are the following:

Comercial IEC 60950-1
Industrial
Militrary
Medical IEC 60601-1

The reliability and lifetime of a product is greatly affected by excessive temperatures. Each electrical component is rated for use within a particular temperature range. To exceed those temperatures is to greatly increase the likelihood of component failure. For some components, electrolytic capacitors in particular, high temperature operation, even within the suggested temperature range, can have dramatic effects on the useful life of the component. An electrolytic capacitor may be rated for 10,000 hours of continuous use at a particular temperature, but an increase of only 10 oC will drop that by half to 5000 hours. This is particularly critical when used in power supplies, because these capacitors are also sensitive to high ripple currents, which switching power supplies commonly generate. So, to take an open frame power supply that is designed to last 5 years in the field at 50 oC, and subject it to 60 oC ambient temperatures, even if derated, can still cause early field failures.

One last point to consider in thermal planning, is the consequences of the failure of any particular component or subsystem, that is being stressed at or near its operating temperature limit. If the device is being stressed it will likely be one of the first components to fail in operation, and so, for safety and other considerations, the consequences of that failure should be known, and perhaps even designed for.

Heat Removal

For most applications, designers make use of conduction and convection to remove heat. The main tools for heat removal are heat sinks using the principle of conduction, and air to remove thermal energy from the heat sink surface through convection. The goal is to keep all temperatures in the desired range.

Heat sinks are made in many application-specific shapes and sizes, usually of a highly thermally conductive (high k) material. They are often attached to the heat-generating device, with a thermal compound in between to eliminate air pockets (air is a poor conductor of heat), with the heat flowing into the heat sink. From there heat is absorbed at the heat sink surface into the air by either natural or forced convection. The most common heat sink is the finned type consisting of a flat plate with a number of fins extending from the surface. The heat sink thus increases the surface area exposed to the air. This increases convective heat transfer defined by the Newtonian cooling equation.

Newtonian Cooling Equation:

Qc is the convective heat transfer; As is the surface area (increased with the heat sink); ∆T is the difference in temperature between the heat sink surface and neighboring air; and hc is the convection heat transfer coefficient (Harper 2005, pp. 3.21-3.22). Please see the Harper reference for more details and analysis.

Fans can provide a stream of air for forced convection (if natural is insufficient). If the heat sink is purchased as a standard component, it will typically have a datasheet containing the heat sink’s natural and forced convection parameters used to calculate hc. Air flow inside the product enclosure depends on several factors, including fan size, type, and location. The air can flow in a smooth continuous path (laminar) or in an irregular motion with eddies (turbulent). More heat transfers in turbulent flow (Harper 2005, pp. 3.23).

There are three simple configurations for this type of heat transfer:

A fan forces outside air into the enclosure (Pressure).
A fan forces air out of the enclosure (Vacuum).
At one part of the enclosure, a fan forces air in, while another fan forces air out elsewhere in the enclosure (Combined Pressure and Vacuum).

There is debate about which is the superior cooling configuration. For example, one reference for this paper concludes configuration 1 is best (Grimes 2002, p.683), and another concludes 2 is more favorable (Rezek 1966, p.42). In reality, it depends most on the application. If the air outside the enclosure were warm or corrosive, then configuration 2 may be preferred over 1. It also depends on the component layout inside the enclosure as well as on the fan size and speed. An in-depth discussion air flow would be found in a fluid mechanics textbook.

Conclusion

Thermal management is a critical process influencing the safety, reliability, and the performance of all electronic products. When utilizing an open frame power supply in an OEM product, it is crucial that the thermal environement be carefully evaluated and controlled to ensure proper and reliable operation of the product over its intended lifetime. This process requires the determination of thermal resistances, air flows, and other geometric calculations, as well as careful layout and heat removal. With the proper planning, the reliability and lifetime of the end product can be greatly improved.

List of References

Elpac Technical Staff. Personal Interview, Topic: “Thermal Management.” Elpac Electronics Inc., Irvine, CA, June 8, 2005.

Grimes, Ronan and Davies, Mark. “The Effect of Fan Operating Point and Location on Temperature Distribution in Electronic Systems,” 2002 Inter Society Conference on Thermal Phenomena (IEEE 2002).

Halliday, D., et al, Fundamentals of Physics, Fifth Edition, New York: John Wiley & Sons, 1997, pp. 461, 471-472.

Harper, Charles, Electronic packaging and interconnection handbook, Fourth Edition, New York: Mc Graw-Hill, 2005, Chapter 3 on thermal management.

Luiten, Wendy. “Sense and Nonsense Thermal Requirements,” 19 th IEEE SEMI-THERM Symposium, 0-7803-7793-1/03, (IEEE 2003), pp. 332-333.

Rezek, Gerard. “Suction vs. Pressure Forced Air Cooling – Part II,” IEEE Transactions on Parts, Materials and Packaging (IEEE 1966).

Zhang, Wei, et al. “Integrated EMI/Thermal Design for Switching Power Supplies,” 0-7803-5692-6/00, (IEEE 2000).

About the Authors

Ethan Matthes is a graduate student at the University of California Irvine, studying Power Electronics and RF Design. He recieved his BSEE from Oregon State University. He is also a certified Engineer in Training.

Ryan Benhard is a Product Engineer at Elpac Electronics, Inc. in Irvine, California. He holds a BSEE/ME from Rose-Hulman Institute of Technology.