UBM Tech

New power regulations bring power-factor correction to lower-power supplies

Margery Conner, Technical Editor - August 26, 2010

 View as PDF

Until recently, power factor and PFC (power- factor-correction) circuits were of concern only for utilities and manufacturers of motors. Utilities have for years specified the power-factor performance of large inductive motors. The utilities also generally can charge industrial customers for reactive-power consumption. However, residential customers are starting to introduce more reactive-power loads into the mix as energy-efficient lights, such as CFLs (compact fluorescent lights) and LED-based lights containing their own ac/dc lighting ballasts, begin to emerge. As a result, power factor has moved from the realm of large-scale industrial motors down to that of consumer electronics.

Power factor is the ratio of real power to apparent power. When both current and voltage are sinusoidal and in-phase, the power factor is one. If they are sinusoidal but not in-phase, then the power factor is the cosine of the phase angle. Purely sinusoidal current and voltage waveforms occur when the load comprises resistive, capacitive, and inductive elements that are all linear.

In a purely resistive load, real power is the same as apparent power, and the power factor is one. When the load has inductance or capacitance, however, the apparent power is greater than the real power because the capacitance and inductance introduce a phase lag between the current and the voltage. Although utilities currently charge residential customers only for the real power they consume, the utilities must add power to support the out-of-phase current and voltage. The additional power is wasted in the form of resistive losses on the grid’s transmission lines. Because these losses increase as the square of the current increases, losses due to low power factor can quickly add up.

The ac line sees an SMPS (switched-mode power supply) as a nonsinusoidal, nonlinear impedance (Figure 1). Figure 2 shows the voltage and current for the circuit of Figure 1. To more closely follow the input voltage and avoid sharp current spikes, the capacitor must charge over the entire positive portion of the cycle. In addition to upping the power factor closer to one, this shaping of the current allows for the use of a smaller capacitor and avoids the creation of harmonic noise, thus reducing THD (total harmonic distortion). This compensating additional circuitry is the PFC circuit. Energy-efficiency specifications regulate both THD and power factor as part of the power-factor specification.

Note that PFCs decrease rather than increase power-supply efficiency. “Power-supply designers are only moving to incorporating improved power factor in their supplies because of government mandates,” says Steve Mappus, a systems engineer in the High Power Solutions group at Fairchild. The utility companies are the immediate beneficiaries of improved power factor, but consumers benefit downstream because utilities need not build additional power plants, holding pollution and carbon emissions in check.

Many topologies and approaches exist for enabling PFC (see sidebar “Utilities and PFC”). In general, PFC is either passive or active. Government and industry regulations specify only the power factor and the THD, leaving the decision about whether to use a passive or an active circuit to the design engineer (tables 1 through 3).

Passive PFC is a simple, relatively inexpensive approach, but it has drawbacks. Chief among them is that it’s difficult, although not impossible, to get a power factor of more than 0.7, and the trend in global regulations is toward power factors of 0.9 and higher. Another difficulty with passive PFC is that the capacitors go directly on the ac line, necessitating capacitor ratings of 400V or higher. This requirement makes the use of electrolytic capacitors the most common approach. Electrolytic capacitors’ lifetime drops with higher temperatures, so you must derate the capacitor if your system will need to work in a hot environment. To derate it, you must either choose a higher temperature, more-expensive capacitor or allow for a shorter capacitor lifetime. Another drawback with passive PFC is that its voltage output is unregulated as it feeds into the dc/dc-conversion stage. For these reasons, the trend is toward active PFC, usually in the form of a boost-converter circuit between the bridge rectifier and the storage capacitor.

The most common configurations for an ac/dc power supply with PFC are two-stage and single-stage designs. In a two-stage design, the ac line feeds into an ac/dc converter, usually comprising a bridge rectifier feeding into a capacitor whose output is usually full of second-harmonic ripple. A dc/dc converter follows the ac/dc converter to provide electrical isolation and voltage regulation. This approach keeps the two stages separate, is easy to troubleshoot, and is simple. However, the double conversion is less efficient, and costs are higher because of the need for two stages. As PFC becomes more prevalent, singlestage PFC is becoming more common.

Kishore Manghnani, vice president of Green Technology for Marvell, argues that single-stage power-converter/PFC ICs are the best design approaches for LED lighting. “With two-stage you end up using two separate chips—one for PFC and the other for the LEDdriver circuit, which includes dimming and the TRIAC [triode-alternatingcurrent-switch] interface,” he says. “In a single-stage converter/PFC chip, you need no additional components: You put the LED driver and the PFC all in one chip.” You might wonder whether you can use cheaper passive PFC, given that cost is a pacing item in LED lighting, but Manghnani advises against this approach. “The biggest advantage of active PFC is that you can use a lowvoltage capacitor,” he says. “In passive PFC, the capacitor must support 400V. In the active single-stage [approach], the capacitor is only 40V. Plus, the lifetime of the capacitor [for the same cost] could be four to five times longer.”

Currently, Energy Star’s draft proposal for energy-efficient luminaires calls for a power factor of 0.7 in residential lighting and 0.9 in commercial, but Marvell argues that the power factor for LED lighting should be 0.9 for all lights with wattage higher than 5W, which is currently the specification in Europe and Korea. “The United States is a bit behind in this area,” Manghnani says. “Europe and Korea require lighting power to be more than 0.9. It doesn’t cost anything extra, so why not add it?”

The mandate for minimum power-factor requirements comes at the same time as the industry is imposing increasingly tight efficiency standards on power supplies, causing a double whammy on designers, who must strive for more stringent efficiencies as power factor creeps up. Thus, research is ongoing on the most efficient ac/dc-converter/PFC circuits. It’s important to understand the general categories for control and which power-supply types these methods can work with. A brief overview of active-PFC methods follows. For more information and circuit diagrams, see references 1 and 2.

The main control methods for active- PFC circuits are DCM (discontinuous-conduction mode), CCM (continuous- conduction mode, and CRM (critical-conduction mode). Various chip manufacturers have their own versions of CRM control, such as BCM (boundary-conduction mode) and TM (transition mode). “Current conduction” in these terms refers to the inductor current.

Low-power supplies typically use DCM. CCM works for all power levels but involves a hard reverse recovery of the output diode when the MOSFET switch turns on. This recovery can cause high losses in a standard inexpensive diode. For high efficiency, therefore, you must use a more expensive diode, such as one made from silicon carbide.

In contrast, BCM circuits switch on the MOSFET with no current in the diode, allowing you to use inexpensive output silicon rectifiers. The trade-offs are that BCM uses a more complex variable-switching-frequency scheme, and its higher peak currents result in higher losses at higher power, limiting most BCM designs to less than 300W. At higher power levels, the CCM boost is more effective due to its lower ripple currents that result in lower peak currents and lower differential-mode EMI (electromagnetic interference).

However, recent innovations in converter/PFC design interleave multiple-phase BCM-controller ICs, such as those from Fairchild, On Semiconductor, and Texas Instruments. Interleaved designs parallel two or more BCM power stages, allowing your design to reach 1 kW or more and reducing the ripple current in the output, which allows for smaller inductors. “If you look at the costs of the controller in comparison to the costs of the PFC, which includes the inductor and all the power components, the controller is not a significant cost,” says Jim Aliberti, product-marketing engineer at Texas Instruments. “It’s the magnetics. People are looking for a way to reduce the costs, and interleaved design has enabled lower system costs because you don’t have to process as much ripple current, allowing you to use smaller magnetics.”

As with most power-conversion applications, digital power is making inroads into PFC, as well. For example, Cirrus Logic offers an IC for digitally controlled PFC. The active-PFC CS1500 and CS1600 DCM ICs target use in power supplies requiring as much as 300W. The CS1500 addresses power supplies for applications such as laptops, digital TVs, and PCs, and the CS1600 targets electronic-lighting ballasts. At approximately 30 cents (high volumes), the chips compare in price with analog PFC ICs but use 30% fewer additional components and fewer parts for EMI filtering. The power factor, which varies with the input line voltage and the load, is greater than 0.95. The iW3620 LED driver from iWatt is also a digital single-stage, active-PFC device. Texas Instruments offers a development kit for high-voltage PFC, which includes hardware and software to implement two-phase interleaved digital PFC for regulation compliance. The kit can work with the company’s application-development kits for the C2000 Piccolo microcontrollers, such as the C2000 ac/dc developer’s kit, as well as end-product kits, such as motorcontrol and LED-lighting-control kits.

Providing efficiency of as much as 30% in applications such as air conditioners and refrigerators, dc motors are replacing inductive motors in both residential and industrial applications. Because of their complex control algorithms, most dc-motor controllers use a DSP. Designers can add digital PFC to designs that already have a DSP—often with no additional components. The cost in engineering learning time can be considerable, though, which is why TI offers its DSP-developer kits for several applications. The digital-PFC kit can work to provide a PFC block for a DSP-motor controller. Another likely application for digital PFC is LED lighting. For example, a DSP such as the C2000 can run PFC in addition to powering an LED array (Figure 3).

Do you remember the advertising push several years ago for the smart refrigerator that would track when you were low on milk and autonomously order more? That idea didn’t catch on. Perhaps the more likely intelligence will be a DSP that controls all of the power-efficient function of home appliances.

You can reach Technical Editor Margery Conner at 1-805-461-8242 and margery.conner@cancom.com.

 References “Power Factor Correction (PFC) Basics,” Fairchild Semiconductor, 2004.Power Factor Correction Handbook, On Semiconductor, September 2007.

 For More Information Cirrus Fairchild Semiconductor National Semiconductor STMicro Climate Savers iWatt On Semiconductor Texas Instruments Energy Star Marvell PGE