Power Electronics

Revamped Planar MOSFETs Are Ready for PFC

In DCM PFC circuits, the faster switching of new planar MOSFETs enables them to achieve higher efficiency than superjunction devices at power levels below 500 W.

Switch-mode power supplies are increasingly being designed with an active power-factor correction (PFC) at the input stage to meet international regulations for harmonics. The boost topology in discontinuous current mode (DCM) is considered by many to be the most suitable PFC method for converters rated at less than 300 W of output power.

In this topology, the losses associated with turn-on of the boost switch are negligible, while turn-off and conduction losses account for the majority of the converter's power losses. Traditionally, planar MOSFETs have been widely used as the boost switch because of their low cost. In cases where higher efficiency was desired, some designers have used superjunction MOSFETs as the boost switch because of the lower RDSON associated with these devices. But the tradeoff for that higher performance has been higher cost.

However, simulation and bench testing conducted at the device and system levels demonstrates that the latest planar power MOSFETs can provide excellent performance in DCM PFC applications where up to several hundred watts are required. In these applications, the new planar MOSFETs offer a more-efficient alternative to the superjunction devices currently being used.

MOSFET Technologies

Planar power MOSFETs have high ruggedness under unclamped inductive switching (UIS) conditions. This can occur when power supplies power up during an ac line transient. The latest generation of planar MOSFETs such as Fairchild Semiconductor's UniFET family can offer even greater UIS ruggedness than previous planar devices.

For example, a 265-mΩ, 500-V UniFET shows more than 80 A of avalanche current under a low-coil UIS test. Moreover, it does not fail at all in the test. On the contrary, a conventional planar MOSFET with the same on-resistance failed at around 40 A. The improved ruggedness will ensure enhanced reliability in a system.

In terms of switching performance, gate charge is one of the benchmarks to compare different devices. The UniFET has a smaller gate charge, faster switching characteristics and lower switching power losses than conventional planar MOSFETs such as Fairchild's QFET, a previous-generation MOSFET family. Some typical electric-characteristics benchmarks are shown in the table below.

The Benefit of DCM

Generally, PFC circuits have used a boost topology because it is simple and inexpensive. There are two modes of PFC boost circuit operation. One is continuous current mode (CCM) in which there is continuous inductor current. This mode has many benefits such as lower core loss, lower ripple current and a smaller input filter. But, it requires a very fast reverse-recovery diode for the boost diode, because the boost switch is being switched on while the inductor current is not zero.

In contrast, in DCM the controller switches on the boost switch when the inductor current is zero. This allows the use of slower, less-expensive diodes. The turn-on loss of the boost switch is also negligible. However, the DCM is usually used for small power supplies — 300 W or less where inductor current is relatively low and where low cost is an important design constraint.

Simulation and Experimental Results

Conduction loss is easy to evaluate because the RDSON value is clearly stated in data sheets. However, the switching loss varies greatly according to circuit conditions. To compare the switching performance in a system, one UniFET and one superjunction device were selected and evaluated. An inductive switching test board was used to measure switching loss at turn-off transient. In this way, it is possible to control important test variables like drain current (ID) and external series gate resistance (RG).

Fig. 1 shows the energy-loss curves for planar versus superjunction MOSFETs under different conditions of RG and ID. The solid traces indicate the losses of the UniFET and the dotted traces are the losses of the superjunction device. There are four different lines per device according to the preset drain-current levels. The drain-current levels are 20 A, 10 A, 6.5 A and 1.8 A from top to bottom.

As shown in Fig. 1, the UniFET has fewer energy losses than the superjunction device at high current levels. The UniFET also outperforms the superjunction device as gate resistance increases. The test point where the superjunction device does better than the UniFET is at the lowest current and smallest gate resistance. The power loss during the turn-on transition has not been evaluated because it is negligible in the DCM PFC.

Based on the switching-performance evaluation results, a simulation was done to analyze systemwide performance. A 200-W-rated DCM PFC was assumed for the simulation, and simulation time was set to a single cycle of ac input.

The simulated conduction losses are shown in Fig. 2. The lower RDSON contributes to lower conduction losses. Fig. 3 shows combined loss curves at external series gate resistance of 15 Ω. In this figure, the estimated performance of the UniFET is better than the superjunction device due to its fast switching characteristics. The distortion at zero-crossing current regions is due to the convergence error of the simulation. With more switching energy loss data, the convergence error can be reduced.

But, what if RG were reduced? With a 4.4-Ω gate resistor, the superjunction device would be slightly better than the UniFET, as illustrated in Fig. 4. However, as Fig. 1 shows, there wouldn't be a difference in turn-off power losses when both gate resistance and drain current are low.

To verify simulation results, both devices are evaluated using a state-of-the-art game-console power supply. The devices are applied to a DCM PFC block of the power supply, and the test conditions are set as VIN = 110 Vac/60 Hz, POUT = 225 W, RGON = 22 Ω and RGOFF = 3.3 Ω.

In Fig. 5, an infrared camera has been used to measure the device temperature. The three measurement points are a PFC diode and two paralleled PFC MOSFETs. With a small gate resistor, the UniFET temperature is lower than the superjunction device by around 10°C. The lower temperature is the result of smaller switching losses, as shown in Fig. 6.

The UniFET turn-off energy loss at the peak of ac input voltage is less than half of the superjunction device switching losses. There is a little plateau in the drain current of the superjunction that makes turn-off losses bigger. There was no such waveform in the bench test. Perhaps this was due to different gate-drive circuitry and pc-board layout.

Interleaved DCM PFC

Recently, dedicated controllers for the interleaved DCM PFC have been introduced in the market. The interleaved DCM PFC technique is a good option for implementing a high-density, cost-effective converter with extended input-power range. It has quickly become a mainstream topology in switching power supplies for flat-panel displays, because the interleaving technique can reduce the total system cost compared to the CCM topology.

Although it requires a pair of boost inductors, boost switches and rectifiers, interleaved DCM PFC can use small-sized filters, smaller high-voltage aluminum electrolytic capacitors, less-expensive 500-V boost switches and slower rectifiers. With the demand for thin flat-panel TVs, smaller components are a crucial requirement for a low-profile switching power supply.

Since interleaving PFC operates in DCM, UniFET power MOSFETs show competitive performance in the system when compared to superjunction devices. To compare system performance, the UniFET and the superjunction device are tested with an interleaved DCM PFC evaluation board.

In this experiment, the evaluation board has a UCC28060 controller, and the phase-management function is activated. Two RURP860 ultrafast rectifiers are applied as boost diodes. The test conditions are set as VIN = 115 Vac/60 Hz, RGON = 10 Ω and RGOFF = 3.9 Ω. Tests are conducted at room temperature without a fan, and external bias is applied for the controller supply voltage.

The efficiency results are shown in Fig. 7. There is not a large difference in efficiency during a heavy load. Basically, the superjunction device has lower RDSON than the UniFET at the same drain current rating; therefore, it will have more conduction loss advantages as the load becomes heavier. The UniFET's smaller switching losses compensate for its higher RDSON in the heavy load area, and the UniFET shows slightly better performance. In the light load area, switching losses dominate the power losses and the UniFET surpasses the superjunction device.

Suitable Applications

When advanced planar MOSFETs are evaluated at the device and system levels, they demonstrate performance on par with superjunction devices. Therefore, given their lower cost compared to superjunction devices, planar MOSFETs can offer an optimal solution in DCM PFC applications as long as the required breakdown voltage of the boost switch is less than 500 V.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.