The need for high efficiency power conversion is permeating through a wide range of applications. Many of the regulatory initiatives are combining the high efficiency requirement with a power factor correction (PFC) requirement. As will be shown in this article, there are innovative PFC solutions for all power ranges which yield >95% efficiency at the required line and load conditions that enable compliance with the regulatory standards.
Power Factor Correction Topologies
A boost topology offers the simplest option for PFC because of the presence of an inductor at the input. In addition, there is no isolation requirement, so the simple boost works well. However, different schemes are available to control the instantaneous value of the inductor current in order to achieve the power factor correction. A quick overview of these schemes is provided in Figure 1.
Critical Conduction Mode (CRM) PFC
This is a traditionally preferred method for lower power applications since it requires a simpler control scheme. Each cycle starts as soon as the core is reset completely. The trr of the boost diode doesn’t contribute to switching losses, hence a boost diode with lower speed and additionally lower forward voltage drop is a good and cost-effective choice to achieve good efficiency (such as MUR450 and MUR550 from ON Semiconductor). Meanwhile, thanks to the natural oscillation between the inductor and the parasitic capacitor seen from the power switch, voltage on the power switch will go down naturally, which will reduce the turn-on losses further.
However, CRM PFC suffers from some limitations such as the higher rms current and wide operating frequency over a line cycle. Hence, the CRM configuration is more suitable under 250 W applications. The drawback caused by wide operating frequency is the risk of generating interference, higher switching losses, highly dissipative snubbering networks, and lower power factor ratio at light load.
Discontinuous Conduction Mode (DCM) PFC
An alternative way to solve the pain caused by wide operating frequency is to operate at DCM. But the pure DCM operation will need a higher inductor peak current, and hence increase the rms current. The conduction losses and power switch turn off losses increase too.
Discontinuous Conduction Mode / Critical Conduction Mode (DCM / CRM)
An innovative solution that captures the benefits of CRM and eliminates the drawback of wide switching losses is offered through the NCP1601 or NCP1605 DCM/CRM controllers from ON Semiconductor. Depending on the PFC inductance and maximum frequency set, one can choose the operating mode, DCM only or DCM plus CRM. To optimize the efficiency, it is recommended to make it operate at CRM at full load, peak of sinusoidal low line conditions, and operate at DCM at the other conditions. Hence, the peak currents at low line are maintained at the same level as CRM, but the maximum frequency is significantly reduced, easing the filtering burden. Another crucial benefit of reducing the switching frequency is that it helps in reducing light load or no load power consumption to meet various regulatory standards.
The NCP1601 and NCP1605 [3] have a patented control architecture that maintains PFC through the mode transitions and delivers superior performance compared to alternative approaches. This simple approach leads to effective results – a power factor of above 0.99 and efficiency of 95% at 115 Vac and full load.
Continuous Conduction Mode (CCM) PFC
This scheme is the preferred method for higher power (>300 W) applications due to its lower peak currents and constant frequency. However, the switching losses caused by trr of the boost diode will influence the efficiency in evidence. When designing the CCM PFC, one has to use the boost diode with lower trr to reduce the switching losses. It could be a Silicon Carbide Schottky Diode, which has zero reverse recovery. However, this decision is a trade-off between efficiency and cost. Another solution is to choose a boost diode with soft reverse recovery, such as MSR860 from ON Semiconductor.
The method of controlling the CCM PFC is rather complicated compared with CRM PFC. The traditional control solutions have been complex, involving multiple loops, notoriously inaccurate analog multipliers and requiring many components around the control IC. With the introduction of the NCP1653/4 from ON Semiconductor, a simple, yet robust 8-pin CCM PFC controller, this complexity is removed. As illustrated in Figure 2, the NCP1654 requires very few components, yet delivers comparable performance to any other CCM controller.
Selection Criteria
With so many emerging choices for the PFC implementation, the resulting natural question is “How does one decide with which approach to go?” Here is a somewhat simplistic guide to help the designer select the right approach. A more detailed guide is available in PFC Handbook published by ON Semiconductor [1].
Power level
If the power level is below 150 W, you are better off with the CRM or DCM approach. As mentioned earlier, the NCP1601/5 offers an intriguing option of getting the best of both worlds – dynamically.
At power levels above 300 W, CCM is the preferred alternative. One will have to deal with the diode reverse recovery issues, but the peak and RMS currents are contained.
Between 150 W and 300 W, the choice depends on the magnetics design savvy of the designer (the inductor design is more challenging for CRM/DCM), but CCM is a safer, albeit more expensive, choice. With the advent of the NCP1653/4, the cost aspect has been effectively addressed.
Other system requirements: The topology choice also depends on the other system requirements. For example, if there is a need to synchronize frequencies in the system, CRM is out of the question. Additionally, if the second power stage can handle a wide input voltage range, the follower boost option makes a lot of sense. Finally, if the output voltage of the power supply is not tightly specified, there may be an incentive to go with an isolated PFC solution available in a single stage as offered by NCP1651.
Figure 3 shows the efficiency comparison among the reference designs using different topologies, CRM, DCM/CRM, and CCM, which full power is from 270 W to 300 W respectively. One thing to highlight in advance is that this comparison acts as a reference of efficiency only, not for showing which one is the best. It is because the efficiency is tested at 3 different boards. In this example, the DCM/CRM topology provides the best efficiency, up to 96 % at 270 W output, 115 Vac input. The CRM and CCM topologies just run after, not too far away, 95.4 % and 95 % at 275 W and 300W output respectively. They all have good efficiency. It really depends on the designer’s favor to make this decision which control method for PFC.
Conclusion
Many choices available for PFC implementation allow the designers to experiment and choose the best approach for their application. Availability of easy-to-use design tools [2] makes this task quick and pain free. Given the increasing involvement of regulatory agencies in energy regulation and rapid pace of globalization, the applicability of PFC circuits will continue to permeate more and more systems. In this context, it is imperative for the designers to familiarize themselves with the available options and choose the most appropriate ones for their applications.
[1] Olivier Meilhon, Kristie Valdez, Dhaval Dalal, Power factor correction handbook, HBD853/D Rev. 2, Aug 2004, http://www.onsemi.com/pub/Collateral/ HBD853-D.PDF
[2] O. Meilhon, “Computer Design Aid greatly simplifies the design of Power Factor Converters”, published at PCIM-China, March 2005.
[3] Joel Turchi, “A Novel Scheme for Current Shaping Circuits Yields Unity Power Factor in Fixed Frequency and Discontinuous Conduction Mode”, PCIM Europe, May 2004.
Click here for the illustration:
Figure 1, Figure 2, Figure 3
