Losses related to power switching power supplies are worth collecting

Usually, wanting to understand the nature of the parasitic parameters on each component that makes up a typical converter will help determine the magnetic component parameters, design the PCB, design the EMI filter, etc. This is the hardest part of any switching power supply design.

To improve the efficiency of switching power supplies, it is necessary to identify and roughly estimate various losses. The internal losses of the switching power supply can be roughly divided into four aspects: switching loss, conduction loss, additional loss and resistance loss. These losses usually occur together in lossy components and are discussed separately below.

01

Losses associated with power switches

The power switch is one of the two main loss sources inside a typical switching power supply. Losses can basically be divided into two parts: conduction losses and switching losses. Conduction loss is the loss when the power switch is in the conductive state after the power device has been turned on and the driving and switching waveforms have stabilized; switching loss occurs when the power switch is driven and enters a new working state, driving and switching The loss when the waveform is in the transition process. These phases and their waveforms are shown in Figure 1.

The conduction loss can be measured as the product of the voltage and current waveforms across the switch. These waveforms are approximately linear, and the power loss during conduction is given by Equation (1).

A typical way to control this loss is to minimize the voltage drop during the conduction period of the power switch. To achieve this goal, the designer must operate the switch in saturation. These conditions are given by equations (2a) and (2b), driven by base or gate overcurrent, ensuring that the collector or drain current is controlled by external components rather than the power switch itself.

The switching loss during power switch conversion is more complex, including both its own factors and the influence of related components. Waveforms related to losses can only be observed with an oscilloscope with a voltage probe connected to the drain-source (collector-emitter) end. An AC current probe can measure drain or collector current. When measuring the loss at each switching instant, a shielded short-lead probe must be used, because any length of unshielded wire may introduce noise from other power supplies, thus failing to accurately display the true waveform. Once you have a good waveform, you can roughly calculate the area enclosed by the two curves using a simple piecewise summation of triangles and rectangles. For example, the turn-on loss in Figure 1 can be calculated by equation (3).

This result is only the loss value during the turn-on period of the power switch. Adding the turn-off and conduction losses can get the total loss value during the switching period.

02

Losses associated with the output rectifier

Among the total losses within a typical non-synchronous rectifier switching power supply, the losses of the output rectifier account for 40%-65% of the total losses. So it is very important to understand this section. The waveforms associated with the output rectifier can be seen in Figure 2.

Rectifier losses can also be divided into three parts: turn-on loss, conduction loss, and turn-off loss.

The conduction loss of the rectifier is the loss when the rectifier is turned on and the current and voltage waveforms are stable. The suppression of this loss is achieved by selecting the rectifier with the lowest forward voltage drop when passing a certain current. PN diodes have flatter forward VI characteristics, but the voltage drop is relatively high (0.7~1.1V); Schottky diodes have lower turning voltage (O.3~0.6V), but the voltage-current characteristics are not too steep. This means that as the current increases, its forward voltage increases faster than that of a PN diode. The transition process in the waveform is segmented into rectangular and triangular areas, and this loss can be calculated using equation (3).

Analyzing the switching losses of the output rectifier is much more complex. The inherent characteristics of the rectifier itself can cause many problems within the local circuit.

During turn-on, the transition process is determined by the forward recovery characteristics of the rectifier tube. The forward recovery time tfrr is the time it takes for the forward voltage to start flowing through the diode. For PN type fast recovery diodes, this time is 5~15ns. Schottky diodes sometimes exhibit longer forward recovery time characteristics due to their inherently higher junction capacitance. Although this loss is not significant, it can cause other problems within the power supply. During forward recovery, the inductor and transformer do not have a large load impedance and the power switch or rectifier is still off, which allows the stored energy to oscillate until the rectifier eventually begins to flow forward current and clamps the power signal.

At the moment of turn-off, the reverse recovery characteristic plays a major role. When a reverse voltage is applied to both ends of the diode, the reverse recovery characteristics of the PN diode are determined by the carriers in the junction. These carriers with limited mobility need to go out in the opposite direction from the original entry into the junction, thus forming a Reverse current flowing through the diode. The losses associated with this may be large, because the reverse voltage will quickly rise very high before the junction charge is depleted, and the reverse current will be reflected through the transformer to the primary-side power switch, increasing the losses of the power tube. Taking Figure 1 as an example, you can see the current peak value during turn-on.

A similar reverse recovery characteristic occurs in high-voltage Schottky rectifiers, which is not caused by carriers but by the high junction capacitance of this type of Schottky diode. The so-called high voltage Schottky diode means that its reverse breakdown voltage is greater than 60V.

03

Losses related to filter capacitors

Input and output filter capacitors are not the main source of loss in switching power supplies, although they have a great impact on the working life of the power supply. If the input capacitor is not selected correctly, the power supply will not operate as efficiently as it should.

Each capacitor has a small resistance and inductance in series with the capacitance. Equivalent series resistance (ESR) and equivalent series inductance (ESL) are parasitic components caused by the structure of the capacitor. They will prevent external signals from being applied to the internal capacitance. Therefore, capacitors perform best when operating at DC, but perform much worse at the switching frequency of the power supply.

The input and output capacitors are the only source (or storage) of high-frequency currents generated by the power switch or output rectifier, so the current flowing through these capacitors ESR can be reasonably determined by observing these current waveforms. This current inevitably generates heat within the capacitor. The main task of designing filter capacitors is to ensure that the internal heat of the capacitor is low enough to ensure the life of the product. Equation (4) gives the calculation formula for the power loss caused by the ESR of the capacitor.

Not only can the resistive part of the capacitor model cause problems, but if the parallel capacitor leads are asymmetrical, the lead inductance will cause uneven heating within the capacitor, thus shortening the life of the hottest capacitor.