*Reliable, low component count, low-cost solution*

*Switch-mode power supplies are now used in almost all electronic systems due to their good line and load regulation, high conversion efficiency, and compact size. Unfortunately, switching power supplies are nonlinear components and can generate narrow, high amplitude pulses out of phase with the line voltage. The significant harmonic content of the current pulses, coupled with the reactive input of the power supply, tends to degrade the AC source, causing EMI problems and energy loss.*

*By Edward Ong, Product Marketing Manager, Power Integrations, Inc., San Jose, CA*

The relationship between input voltage and current is expressed as power factor (PF). In systems that generate nonlinear load currents, power factor correction (PFC) circuits are employed to make the input of the power supply appear as a linear load to the power system. An effective PFC circuit reduces both peak and RMS currents, and optimizes the efficiency of power delivery from the AC source.

Because of the impact of power factor on the supply infrastructure, government agencies have introduced and progressively tightened their requirements for PF and harmonic distortion. IEC/EN62019-3-2 is widely used as the PF standard for electronic equipment, both for domestic and commercial applications.

PFC can be achieved through various topologies, such as buck, boost, flyback, Cuk, and single-ended primary-inductor converter (SEPIC). The boost topology has become popular because it is simple, and the continuous input inductor current makes it an ideal candidate for PFC. A number of control strategies have been introduced by power supply IC manufacturers including peak, average, and hysteresis continuous conduction current (CCM) mode control, and critical mode control discontinuous conduction mode (DCM). PFC ICs have enabled power supply manufacturers to achieve significant improvements in PF performance in recent years, although there have been reliability issues associated with the complexity of some implementations.

*Hiper*PFS highlights

*Hiper*PFTM, a new PFC IC introduced by Power Integrations, differentiates itself by employing a unique control strategy, constant ampseconds on-time control, and constant volt-seconds off-time control. The single chip solution provides integrated loss-free current sensing and eliminates the current control loop external compensation components, thus minimizing design complexity. The innovative variablefrequency continuous conduction mode operation (VF-CCM) suppresses EMI and minimizes switching losses by operating at low average switching frequency.

### Amp-second and volt-second control

The core of HiperPFS is a constant amp-second on-time and constant volt-second off-time control algorithm. In Figure 1, a boost PFC is used as an example to illustrate the control mechanism. Integrating the switch current and controlling it to have a constant amp-second product over the on-time of the switch allows the average input current to follow the input voltage. Integrating the difference between the output and input voltage maintains a constant volt-second balance dictated by the electromagnetic properties of the boost inductor, thus regulating the output voltage and power.

The controller sets a constant value for the integration of the switch current during each on-cycle of the power MOSFET. The integrated current per cycle is considered constant over a half line cycle due to the very low bandwidth of the boost converter output voltage control loop, far below the 120 Hz half line cycle frequency, in fact. In order to regulate the output voltage, control voltage VC varies steadily over many cycles in response to load or line changes. With this constant amp-seconds control, we can first assume:

For the off-time control, a current source proportional to the difference between the output and input voltage is employed. The current is integrated and compared to a fixed voltage reference (V_{OFF}) to determine the cycle off-time. The volt-seconds for the off-time (t_{OFF}) can be expressed as:

Since the volt-seconds during the on-time must equal the volt-seconds during the off-time, to maintain flux equilibrium in the PFC inductor, the on-time (t_{ON}) is controlled such that:

Substituting t_{ON} from (3) into (1) gives:

The relationship of (4) demonstrates that by controlling a constant amp-second on-time and constant volt-second off-time, the input current iin is proportional to the input voltage V_{in}, therefore providing inherent power factor correction with very simple control circuitry.

### Variable frequency continuous conduction mode (VF-CCM)

The plots shown in Figure 2 illustrate the resulting variation in frequencies with input line voltage and output load. As the line voltage increases, the voltage difference across the PFC inductor becomes smaller, and a longer time is required for the off-time integrator to reach the V_{OFF} threshold. As the input voltage decreases, the off-time integrator requires less time to satisfy the volt-second balance.

The switch-on time varies with the load. As the load increases, the PFC switch current increases to meet the load demand. With the increased switch current, the on-time integrator requires less time to satisfy the amp-second balance. Consequently, the switching frequency increases.

The variable switching feature of VF-CCM operation minimizes switching losses by maintaining a low average switching frequency and maximizes efficiency over the entire load range of the converter.

At light load, the off-time integrator control reference (V_{OFF}) is modified by an internal error signal (V_{E}), which is directly proportional to the output power. The modified V_{OFF} slope reduces the average frequency further to minimize switching losses. High efficiency at light load is a challenge for traditional PFC CCM approaches in which fixed MOSFET switching frequencies cause fixed switching losses on each cycle, even at light loads. Fixed frequency CCM operation is shown in Figure 3.

With the fixed frequency CCM design, the sub-harmonic noise focuses on a few fixed frequencies, making filtering EMI noise a challenge. In variable frequency control, the energy delivered in the switching pulses is spread across a range of frequencies over the half AC line cycle. This means that HiperPFS typically reduces the total X and Y capacitance requirements of the converter and the inductance of both the boost choke and EMI noise suppression chokes, reducing overall system size and cost.

### Simplification of PFC design

Figure 4 shows a typical HiperPFS-based PFC application circuit. The VF-CCM operation has eliminated the requirement for external compensation networks, producing a very simple solution.

The voltage monitor pin (V) current is used internally to detect the peak of the input line voltage. This drives the line feed forward function, maintaining a constant voltage feedback loop gain over the operating input line range to improve line regulation and transient response. *Hiper*PFS combines other functions, such as power limit and brown-in/brown-out protection.

By contrast, Figure 5 shows an example of a traditional CCM average current mode control. It needs a current amplifier and a compensation network. The current sense resistor needs to be positioned in series with the inductor current. In addition to the resistor power loss, the noise sensitivity issue also becomes challenge, especially when the inductor ripple current is low.

### VF-CCM vs. critical conduction mode operation (CRM)

The CRM boost power factor converter operates at the boundary of continuous conduction mode and discontinuous conduction mode. Normally, the switch on-time is fixed, achieved by comparing the voltage loop error amplifier output voltage to a saw-tooth reference waveform. When the levels match, the switch is turned off. The switch is turned on when the inductor current falls to zero. As the inductor value is fixed, the input current automatically tracks the input voltage, therefore attaining power factor correction. The inductor current is depicted in Figure 6.

CRM control shares some benefits with HiperPFS, such as simple design with no current control compensation and variable switching frequency. Selection of the freewheeling diode is not critical, as the diode is turned off when the switch current is zero. However, there are significant drawbacks that prevent CRM from being used in higher power PFC designs.

CRM operation results in high peak current in the MOSFET and freewheeling diode, therefore it requires higher current ratings for the devices.

• There are higher switching and conduction power losses in the MOSFET.

• CRM operation requires a larger core. Because it generates higher peak-to-peak inductor current, this results in higher hysteresis losses in the inductor and higher copper losses.

• CRM either needs a current sense resistor to sense zero inductor current, or a zero current detection winding to turn on the MOSFET.

• Compared to a similar VF-CCM design, CRM operation generates close to twice the peak current. This increases noise issues incurring more cost in the EMI filter elements.

CRM-based PFC ICs have been popular in low-power PFC designs, as they are simple and allow the use of inexpensive freewheeling diodes. However, *Hiper*PFS is even simpler and offers many design advantages, such as lower MOSFET conduction and switching losses, lower diode conduction losses, lower inductor core and copper losses, higher efficiency over load range, lower EMI and smaller EMI filters, lower component count, and integrated protection features. With the easy availability of ultrafast recovery rectifiers with soft recovery characteristics, the HiperPFS VF-CCM mode of operation presents the ideal solution for wide range of low-, medium-, and highpower PFC applications.

### Design example

A 347-watt PFC front-end converter, shown in Figure 7, has been designed using a HiperPFS PFS714EG integrated PFC controller and is the subject of a comprehensive qualification report (RDR-236). The design example demonstrated is available to developers and can be used as a reference for prototyping new designs.

The design delivers greater than 95% efficiency from 10% to full load (see Figure 8). The high efficiency enables the design to meet 80+ PC specification requirements.

The power supply operates at a high power factor of 0.998 at 115 VAC input full load and 0.984 at 230 VAC input full load (Figure 9). It easily meets EN62019-3-2 Class C and D compliance with low harmonic input current components (Figure 10).

**Conclusion**

The innovative constant amp-seconds and volt-seconds control concept introduced with HiperPFS brings a brand new high-performance PFC solution for the boost PFC converter. When compared with traditional CCM and CRM operations, HiperPFS offers power supply designers a better choice with a simple, reliable, low component count, and low-cost solution.

**References:**

1) Power Integrations PFS704-729EG HiperPFS Family Datasheet.

2) Power Integrations Application Note AN-52, Application Note AN-53.

3) Reference Design Report (RDR-236) for a High Performance 347 W PFC Stage Using HiperPFS PFS714EG.

4) L. Rossetto, G. Spiazzi, and P. Tenti “Control Techniques for Power Factor Correction Converters”.

5) Lloyd H. Dixon, Jr. “High Power Preregulators for Off-Line Power Supplies,” TI-Unitrode slup087.