Power Factor Correction Increases the Efficiency of Industry

Energy usage and efficiency levels are being placed under closer scrutiny than ever before, as consumers become more concerned about the size of their utility bills and business owners look to curb their spiraling operational expenses. All this is being further exacerbated by the increasingly important environmental aspect, with the widespread recognition that inefficient devices generate waste heat that will ultimately have an unwanted ecological impact.

In the world of power electronics, conversion efficiency has always been a critical topic for discussion and one of the principal parameters on any datasheet. Seeking to present their products in the most favorable light, power supply manufacturers will generally quote the ‘best possible’ number, which was often a single value at around 80% load. However, it should be noted that in practical applications the power drawn may not be at this level at all. 

Loads may fluctuate significantly – depending on how they are being operated and, in redundant configurations, the power drawn will always be much lower (unless a fault condition occurs). This means that the actual efficiency of the system can be far less than the quoted efficiency value would imply.

Recognizing the gravity of this situation, new energy guidelines were put in place by standards bodies, industry groups, and government agencies. These guidelines have generally taken the form of citing efficiency curves that mandated minimum acceptable efficiency levels across all operating loads, from 20% to full load. Consequently, design engineers have been able to evaluate fundamental building blocks within power systems to identify where losses would occur, and then act to eliminate them, thereby ensuring that the new efficiency guidelines are met. Power factor correction (PFC) is critical in addressing the sources of potential loss and should be implemented accordingly.

Understanding losses in power systems

A no power system will be perfectly efficient and, while modern switching semiconductor devices now offer unprecedented levels of performance, there will always be some losses during operation, thereby contributing to reduced efficiency levels. In power systems, there are broadly two types of loss to be aware of: switching and conduction.

Conduction losses include losses due to the forward voltage of the bridge diodes, which are proportional to system power and the on-resistance of switching devices such as MOSFETs and IGBTs. These are directly proportional to the square of the overall system power. As they increase with delivered power, they tend to have more effect in scenarios that are closer to full load. Traditionally, the most focus used to be applied here.

The second type of loss is switching loss. As design engineers strive to heighten power density levels and reduce system size, switching frequencies continue to increase, allowing for a reduction in the size of the bulky magnetic components incorporated into the system. Switching losses relate to the constant recharging of parasitic capacitances (such as those found in switching device gates). These are proportional to switching frequency and consistent throughout the operating power range. These losses tend to be most prevalent at lower power levels, where they can have a significant effect on system efficiency.

So why is PFC of such importance to efficiency?

All grid power supplied by utility companies is AC, and the voltage waveform is always sinusoidal. However, the shape and phase of the current waveform are not necessarily sinusoidal and are determined by the load being powered. For the simplest purely resistive load, such as a heating element the load current is in phase with the voltage and remains sinusoidal. Calculating the power delivered in this case is merely a matter of multiplying the voltage and current together.

Other types of loads, such as motors, may include a reactive component (inductive or capacitive). In this case, while the current waveform remains sinusoidal, it will be phase-shifted with respect to the voltage waveform, with the amount of reactance in the load determining the amount of phase shift. The power calculation needs to take the phasing into account, and so the real power is determined by the equation:

Real power = V * I * cos(Φ)

Here f represents the phase angle between the voltage and the current waveforms and cos(Φ) is known as the ‘displacement factor.’ In resistive loads, with the current and voltage in phase cos(Φ) will have a value of 1 – meaning that real power remains the product of voltage and current, as normal. However, real-world loads are often not anything like that simple, especially where the load is a switch-mode power supply (SMPS), for instance. These units typically have a diode bridge rectifier and inrush capacitor that will cause the current waveform to lose its sinusoidal shape and become a series of spikes.

As the waveform is distorted and no longer sinusoidal, the real power is calculated using a ‘distortion factor’ (cos(Θ)) which is linked to the total harmonic distortion (THD) of the waveform. So, in systems where the current and voltage are in phase, but the current waveform is non-sinusoidal, the following equation applies:

Real power = V * I * cos(Θ)

In circumstances where the current waveform is both phase-shifted and distorted, things will get a little more complicated. Here both the displacement factor and the distortion factor must be applied:

Real power = V * I * cos(Θ) * cos(Φ)

The power factor of any system is simply the product of the two factors:

Power factor = cos (Θ) * cos(Φ)

In practical terms, this means that the greater the phase difference between voltage and current, or the more distorted the current waveform, the lower the power factor and, therefore, the lower the real power. As the power factor also impacts on efficiency, this is now a key area for power designers to tackle.

The need to correct the power factor

Relatively complex mathematics will show that multiplying two sinusoidal waveforms together can only give a value that is greater than zero if the frequencies are the same. As a result of this, it can be deduced that harmonic currents make no contribution to the useful output power of a system and should be reduced or eliminated.

This is precisely the approach taken in what is considered by most to be the main PFC standard, EN 61000-3-2. In common with many of the modern efficiency specifications, including Energy Star from the U.S. Environmental Protection Agency (EPA), EN61000-3-2 seeks to reduce the THD of current waveforms by defining strict limits for harmonic currents, right up to the 40th harmonic.

The most common approach for implementing PFC is to insert an active stage between the bridge rectifier and bulk capacitor, using one of several common control schemes found in commercially-available PFC controllers. Probably the most widely-used control scheme is continuous conduction mode (CCM), which operates with a fixed frequency and is often found in higher power (>300W) systems. A popular alternative is the critical conduction mode (CrM) control. This dispenses with the need for a fast-recovery diode by switching only when the inductor current falls to zero. This reduces system costs but leads to a variable switching frequency. CrM is particularly prevalent in low power systems, such as ones for lighting.

Figure 1: ON Semiconductor’s FL7921R QR current mode lighting controller (Source: ON Semiconductor)

Please visit Power Electronics News for the complete article