PCB Tech - Analysis on the Thermal Design of High-Frequency board Switching Power Supply

1. Introduction

Electronic products generally have strict requirements on operating temperature. Excessive temperature rise inside the power supply equipment will cause the failure of temperature-sensitive semiconductor devices, electrolytic capacitors and other components. When the temperature exceeds a certain value, the failure rate increases exponentially. Statistics show that the reliability of electronic components decreases by 10% every time the temperature of electronic components rises by 2°C; the lifespan at a temperature rise of 50°C is only 1/6 of that at a temperature rise of 25°C. Therefore, electronic equipment will meet the requirements of controlling the temperature rise of the entire chassis and internal components. This is the thermal design of electronic equipment. For high-frequency board switching power supplies, which have high-power heating devices, temperature is the most important factor affecting their reliability. For this reason, there are strict requirements on the overall thermal design. The complete thermal design includes two aspects: how to control the heat generated by the heat source; how to dissipate the heat generated by the heat source. The ultimate goal is how to control the temperature of the electronic equipment after the thermal equilibrium is reached within the allowable range.

2. Heating control design

The main heating components in the switching power supply are semiconductor switching tubes (such as MOSFET, IGBT, GTR, SCR, etc.), high-power diodes (such as ultra-fast recovery diodes, Schottky diodes, etc.), high-frequency transformers, filter inductors and other magnetic components And fake load, etc. There are different methods of controlling heat generation for each kind of heating element.

2.1 Reduce the heat generation of the power switch

The switch tube is one of the components that generate more heat in the high-frequency switching power supply. Reducing its heat can not only improve the reliability of the switch tube itself, but also reduce the temperature of the whole machine, improve the efficiency of the whole machine and the mean time between failures. (MTBF). When the switch tube is in normal operation, it is in two states of on and off, and the loss generated can be subdivided into the loss caused by the two critical states and the loss caused by the on state. Among them, the on-state loss is determined by the on-state resistance of the switch tube itself. This loss can be reduced by choosing a low on-resistance switch tube. The on-state resistance of MOSFET is larger than that of IGBT, but its operating frequency is higher, so it is still the preferred device for switching power supply design. Now IR's new IRL3713 series HEXFET (hexagonal field effect transistor) power MOSFET has achieved 3mΩ on-state resistance, so that these devices have lower conduction loss, gate charge and switching loss. The American APT company also has similar products. The losses in the two critical states of turn-on and turn-off can also be reduced by selecting devices with faster switching speeds and shorter recovery times. But the more important thing is to reduce losses by designing better control methods and buffering techniques. This method can show advantages when the switching frequency is higher. For example, various soft-switching technologies can enable the switching tube to be turned on or off in the zero voltage and zero current state, thereby greatly reducing the loss caused by these two states. However, some manufacturers still use hard-switching technology from the perspective of cost, and they can reduce the loss of the switching tube and improve its reliability through various types of buffering technologies.

2.2 Reduce the heat generation of the power diode

In the high-frequency switching power supply, there are many applications of power diodes, and the selected types are also different. For the power diodes that rectify the input 50Hz alternating current into direct current and the fast recovery diodes in the snubber circuit, under normal circumstances, there will be no better control technology to reduce losses, and only high-quality devices can be selected, such as the use of conduction voltage. Lower Schottky diodes or ultra-fast recovery diodes with faster turn-off speed and soft recovery to reduce losses and heat. The rectifier circuit on the secondary side of the high-frequency transformer can also adopt a synchronous rectification method to further reduce the rectification voltage drop loss and heat generation, but both of them will increase the cost. Therefore, how the manufacturer grasps the balance between performance and cost and achieves the highest cost performance is a question worthy of study.

2.3 Reduce the heating of magnetic components such as high-frequency transformers and filter inductors

Various magnetic components are indispensably used in high-frequency switching power supplies, such as chokes in filters, energy storage filter inductors, isolated power supplies, and high-frequency transformers. They will produce more or less copper loss and iron loss during work, and these losses are emitted in the form of heat. Especially for inductors and transformers, the high-frequency current flowing in the coils will double the copper loss due to the skin effect, so the loss caused by the inductors and transformers becomes a non-negligible part. Therefore, in the design, multiple thin enameled wires should be used for parallel winding, or wide and thin copper sheets should be used for winding to reduce the influence of skin effect. The magnetic core is generally made of high-quality ferrite material, such as TDK magnetic material produced in Japan. A certain margin should be left in the model selection to prevent magnetic saturation.

2.4 Reduce the calorific value of the fake load

In order to avoid the voltage increase caused by no-load state, high-power switching power supplies are often equipped with dummy loads-high-power resistors. This is especially true for power supplies with source PFC units. When the switching power supply is working, the dummy load has to pass a small amount of current, which will not only reduce the efficiency of the switching power supply, but also its heat generation is a factor that affects the thermal stability of the whole machine. The position of the dummy load on the printed board (PCB) is often very close to the electrolytic capacitor used for output filtering, and the electrolytic capacitor is extremely sensitive to temperature. Therefore, it is necessary to reduce the calorific value of the dummy load. A more feasible way is to design the dummy load as a variable impedance method. The size of the dummy load impedance is controlled by detecting the output current of the switching power supply. When the power supply is in a normal load, the dummy load exits the current consumption state; when there is no load, the dummy load consumes the largest current. This will not affect the stability of the power supply at no load, nor will it reduce the efficiency of the power supply and generate a large amount of unnecessary heat.

3. Heat dissipation design

3.1 The basic method of heat dissipation and its calculation method

There are three basic ways of heat dissipation: heat conduction, convection heat transfer and heat radiation.

1) Heat conduction The heat transfer that occurs between the various parts of the object in direct contact or within the object is heat conduction. The mechanism is the mutual transfer of molecular kinetic energy between objects at different temperatures or parts of objects at different temperatures. The concept of heat conduction is very similar to that of current. Heat is always conducted from a place with a high temperature to a place with a low temperature. There is thermal resistance in the process of heat conduction, just as there is resistance in the flow of current. Its heat flow Φ=[W], where Rt is the thermal resistance, and τ is the temperature difference. The thermal resistance Rt=[K/W], where δ is the thickness of the conductor, λ is the thermal conductivity, and A is the cross-sectional area of the conductor. In this way, in the design of the switching power supply, the temperature rise τ=ΦRt can be obtained from the power dissipation of the heating source. In practical applications, the heat flow from the heat source to the radiator often has to pass through thermal conductors of several different materials, that is, there are series of different thermal resistances. In the calculation, the total thermal resistance is the sum of multiple thermal resistances.

2) Convective heat transfer heat is transferred to the fluid layer close to it through heat conduction. After this layer of fluid is heated, its volume expands, its density becomes smaller, and it flows upward, and the surrounding dense fluid flows over to fill it. The fluid absorbs heat and expands and flows upwards, and circulates in this way, continuously taking heat away from the surface of the heating element. This process is called convective heat transfer. The calculation of convective heat transfer generally adopts the formula proposed by Newton: Φ=αA(θ1－θ2)[W], where A is the area of the wall in contact with the fluid [m2], α is the convective heat transfer coefficient, and θ1 is the wall temperature [ K], θ2 is the average temperature of the fluid [K]. It can be seen that the heat flux Φ is proportional to the product of the convective heat transfer coefficient α, the cross-sectional area A, and the temperature difference between the solid surface and the fluid (θ1-θ2). Convection heat transfer is a complex heat transfer process, which is not only determined by the heat process, but also by the dynamic process of the gas. Simply put, there are two factors that affect convective heat transfer: (1) The physical properties of the fluid, such as density, viscosity, expansion coefficient, thermal conductivity, specific heat, etc.; (2) The flow of the fluid is natural convection Still forced convection, laminar flow or turbulent flow. Because in laminar flow, heat transfer mainly relies on heat conduction between unrelated flow layers; while in turbulent flow, the fluid generates vortices outside of the laminar flow bottom layer close to the wall to enhance heat transfer. Generally speaking, under the same other conditions, the heat transfer coefficient of turbulent flow is several times larger than that of laminar flow, or even more.

3) Thermal radiation The propagation of electromagnetic waves caused by temperature differences is called thermal radiation. Its process is much more complicated than heat conduction and convection heat transfer. It is the energy that converts part of the heat energy of an object into electromagnetic waves. It spreads around through the medium that can transmit electromagnetic waves such as air and vacuum. When it encounters other objects, part of it is absorbed and converted into heat energy, and the rest is Be reflected back. The infrared radiation emitted by various objects is a kind of heat radiation. In vacuum or air, the radiation ability Φ radiated by an object depends on the nature of the object, surface condition (such as color, roughness, etc.), surface area, and surface temperature. Φ=εσbA(T14－T24) where σb is Boltzmann's constant with a value of 5.67*10－8, A is the radiation surface area [m2], T is the absolute temperature of the surfaces of the two objects [K], ε is the surface blackness . The darker and rougher the surface of the object, the stronger the radiation ability.