Joel Turchi, Member of Technical Staff, ON Semiconductor
Almost everyone is aware of the need for optimal efficiency, from consumers and business operators seeking to constrain costs in times of high energy prices, through to designers looking to meet the requirements of an increasingly complex and numerous set of standards. If not already motivated by the high cost of wasting energy, then the environmental impact of generating energy that will end up as heat is an increasingly significant issue.
Recognizing the need to improve, governments and industry associations have written standards that, in some cases, must be met before a product is placed on the market. Cost conscious or environmentally aware customers are relying on these standards when making purchasing decisions to know they are buying efficient products.
One of the key areas that needs to be addressed is the power factor correction (PFC) stage including the electro-magnetic interference (EMI) filter.
Efficiency is not just a single point
For any power-related application, efficiency has always been an issue and a parameter that manufacturers have stated in their specifications. However, in the past, efficiency was quoted as the best possible number – a single point, usually at about 75% of full load.
As a result, manufacturers focused on this load level to improve the perceived efficiency of their products. However, in practice, devices only operate at this power level for a small proportion of the time. In real applications, especially those with dynamic loads, this means that the actual efficiency is far lower than may be expected.
To address this situation, modern energy standards consider performance across the whole of the efficiency curve, not just the best point on the curve. As a result, designers are looking at how to design key elements of power conversion systems to perform better at low and medium load levels. One of the most critical areas is the PFC stage and EMI filter that, together, can consume up to 8% of the output power.
PFC: A brief overview
The voltage supplied by utility companies is always sinusoidal, but the waveform shape and phase of the line current depend on the load that is being powered. For the simplest resistive load, the load current is also sinusoidal and in phase making the power easy to calculate.
If there is a reactive element in the load, such as an inductor or capacitor, then the load current remains sinusoidal but it is phase shifted with respect to the voltage. In this case, the active power (also called ‘real’ or ‘average’ power) is calculated as before but multiplied by the cosine of the phase angle (‘displacement factor’). The more reactive the load, the lower the active power.
Things become more complex with non-linear loads, such as the input stage of a typical switch mode power supply with a diode bridge and bulk input capacitor.
Here, the current is a series of inrush spikes and power is calculated using a Fourier transformation.
Averaging the product of two sinusoids requires a complex calculation and only gives a non-zero result if the two waveforms have the same frequency. However, from this it can be derived that only the fundamental component provides real power and the harmonics just generate useless circulating currents.
Similar to the displacement factor, a distortion factor models the effect that a distorted (non-sinusoidal) waveform has on the real power, defining the real power as the product of the rms voltage, rms current and both factors. Further analysis would show that the total harmonic distortion (THD).
In fact, the power factor of a system is simply the product of the displacement and distortion factors – and, therefore, the real power is the product of the rms voltage, rms current and power factor.
Practical approaches to correcting the power factor
The main standard relating to PFC is EN 61000-3-2 - this was written with the goal of minimizing the THD of any current that is supplied from the grid. It achieves this by defining the maximum magnitude for all harmonics from the 2nd to the 40th. PFC requirements are also addressed in other documents such as the Energy Star specification and it is this that many believe has led to the prevalence of PFC technology in so many applications.
By far the most common (and effective) type of PFC used to meet these standards is active PFC. A typical approach is to add a PFC pre-regulator between the input bridge rectifier and the bulk capacitor to deliver a constant voltage while ensuring that the current drawn is sinusoidal.
There are many benefits to this approach other than the obvious improvement in power factor. The output from the PFC stage is generally a fairly well regulated 400 V, which makes the design of the downstream converter easier and less costly. Also, the non-pulsating current reduces EMI filtering requirements saving bulk and cost.
However, this type of PFC pre-converter cannot be 100% efficient and does, therefore, contribute to system losses. In any power system there are two primary types of loss, switching and conduction. Conduction losses are the sum of two types of loss; those that are proportional to the power of the system due to factors such as the forward voltage of the diodes in the bridge and those that are proportional to the square of the system power which comprise resistive losses, such as the on-resistance of MOSFETs. The latter tend to have the most effect on efficiency at higher power levels.
Click image to enlarge
Figure 2: PFC sits between the diode bridge and the bulk capacitor
Switching losses, on the other hand, are for a large part, proportional to the current and hence to the power which is conveyed. For another part, they are constant and not related to the system power. They are caused by parasitic capacitances and charge currents and are generally proportional to the switching frequency of a system. As designers increase operating frequencies to reduce system size, so switching losses become more of a challenge, especially at lower power levels where they can represent a significant efficiency loss.
PFC control schemes
Various control schemes for PFC have been developed to suit the needs of different systems although, in general, the goal is to reduce switching losses at light loads and conduction losses at heavier loads.
Click image to enlarge
Figure 3: Switching and conduction losses contribute to the overall losses in a power system
There are three basic control schemes as shown. Continuous conduction mode (CCM) operates at a fixed frequency and limits the inductor current ripple while allowing increased losses. It is typically used for higher power systems (> 300 W).
Critical conduction mode (CrM) starts a new switching cycle as the inductor current falls to zero saving the need for a fast-recovery diode. This leads to a variable switching frequency with a relatively large ripple current. This simple and low-cost scheme is popular in low power applications, including lighting. As low on-resistance MOSFETs become more common, CrM is being used in higher power applications.
Click image to enlarge
Figure 4: Primary single-channel PFC operating modes
Frequency-clamped critical conduction mode (FCCrM) was introduced by ON Semiconductor several years ago to limit the frequency spread seen with CrM. In light load conditions where frequency is highest, the operating mode is changed to discontinuous conduction mode (DCM) to reduce switching losses. Additional circuitry addresses the ‘deadtimes’ that are typical of DCM, ensuring that the current waveform is the correct shape.
ON Semiconductor offers a wide range of component solutions including power factor controllers and power switches along with significant design resources allowing designers to develop PFC solutions with confidence.