The best flyback solution depends on the application

The best flyback solution depends on the application

Know what regulations, certifications, and devices you should consider for flyback converter designs

BY ALEXANDER CRAIG
Fairchild Semiconductor
San Jose, CA
http://www.fairchildsemi.com

The most ubiquitous power supply topology that manufacturers use today is the flyback topology. This article examines the choices made when designing a flyback power supply and explores the regulations, certifications and topology pros and cons that steer the decision-making process. It also provides guidance for selecting devices used in flyback converters.

When starting a power supply design an engineer needs to look at the requirements of the end applications as well as the regulations and certifications that are needed to satisfy the design. The basic points that a designer needs to know are

• Input voltages (assume universal input 85 to 265 Vac) and output voltages and currents

• Isolation requirements

• Certifications and standards such as Energy Star, EN61000-3-2, 80 PLUS

• Size, cost, and performance

Other special needs that are determined by the application

For low-power isolated designs of less than ~150 W, a flyback is typically employed. The flyback is ideal for multiple output converters because multiple outputs are obtained with minimal additional components.

When choosing a flyback you need to look at the different flyback options and the different methods for operating a flyback to see what it offers to the overall design. An engineer can chose different operational methods from fixed-frequency hard-switched, valley-switched variable frequency, and quasiresonant, primary-side regulation, constant power, constant current, constant voltage or some combination of these. An engineer can also choose the implementation from an integrated solution in which the PWM IC and the high-voltage MOSFET are combined is a single package. These implementations tend to be easier to design and offer a higher level of protection than a nonintegrated solution. But these integrated devices do impose certain restrictions on the overall application.

The basic off-line flyback converter consists of a bridge rectifier, capacitor, PWM controller, high-voltage MOSFET, flyback transformer and output rectifier diodes. Using the basic off-line flyback converter in Fig. 1 , let’s determine how to select the semiconductor devices outlined in red-dashed lines.

Fig. 1. A typical flyback.

Starting with the bridge rectifier, this is a simple power calculation. If we assume 75-W output power supply designed for a universal input 85 to 265 Vac(rms) then we need to size the bridge rectifier. Assuming an 85%-efficient power supply this means the power in is 75 W/0.85 = 88.25 W, so: IAC(rms) max = 88.25 W/85 Vac(rms) max = 1.04 A(rms).

With a 265-Vac(rms) max input, a 400-V bridge rectifier will work but most engineers will choose the 600-V bridge rectifier for design robustness. Given the needs of a 600-V and greater than 1.04-A(rms) the DF06S is a good choice.

Choosing the flyback control IC and MOSFET is a bit more complicated and depends on the needs of the specific application and design priorities such as cost, space and ease of design. Additionally, certifications and standards that the power supply needs to meet (such as, Energy Star, EN61000-3-2, 80 PLUS) in which standby power and efficiency are specified for power levels and applications, must be considered.

When selecting the control IC, first choose from the different operational methods. Fixed-frequency, hard-switched, secondary-side regulated is the most basic approach. This approach can offer an advantage in that a fixed frequency can provide well-defined quiet spots in the EMI spectrum.

Most fixed-frequency control ICs today offer a frequency dithering of a few percent to spread the EMI profile. The secondary-side regulated solutions tend to have the tightest output voltage regulation.

Valley-switched variable frequency, and quasiresonant converters offer lower power losses in the switching MOSFET and a lower EMI profile. The valley-switched variable frequency or limited quasiresonant converters can also offer well-defined quiet spots in the EMI spectrum.

In order to switch at the right time, the control IC needs some form of information about the drain voltage. This is typically provided from the VCC winding on the transformer. True quasiresonant converters tend to offer the most-efficient solutions for higher power flybacks (100 to 250 W). Primary-side regulation control ICs offer an inexpensive option for many lower-power supplies with no need to have an optocoupler and voltage reference. The regulation may not be as tight as with the secondary-side regulation but it is typically good enough for many applications.

Many control ICs today have a mode of operation that cause the converter to operate in a lower-power mode for applications that need an Energy Star-type rating in which the system is required to draw less than a certain amount of power in standby mode. If the target application needs to meet an Energy Star-type specification, the engineer needs to be mindful of the control-IC solution to this operational mode.

Is power factor correction (PFC) needed? If so, combination ICs that offer both PFC and PWM functions with synchronized switching frequencies are available. Lighting applications greater then 25 W likely will need a PFC and other applications may also require PFCs.

MOSFET selection may be part of this integrated solution or as a discrete device. In either case, the choice is all about power, conduction and switching losses. To avoid thermal issues, you need to choose devices based on PD (power dissipation) and Rθja (thermal resistance junction to ambient) calculations. The basic calculation is junction temperature (Tj ), Tj = Pd *Rθja +Ta, Rθja = Rθjc + Rθca . Rθjc (thermal resistance junction to case). To avoid abnormal input voltage hazards, one needs to select devices with a breakdown voltage 20% to 25% higher than is expected to be found in the application. Additionally, remember the reflected voltage from the secondary side and be mindful of any uncontrolled switching inductance (V = L*di/dt). This unclamped inductance is the cause of most overvoltage failures.

Is fixed frequency needed? If so, this can make passing the EMI specification more difficult. But this can also provide a quiet zone if the application needs to transmit in a certain frequency range. This tends to limit the IC options.

The power level will influence the decision whether an integrated or a nonintegrated solution is better for a design. Integrated solutions are easier to implement but impose some restrictions on the design. Nonintegrated solutions offer options such as two-level OCP for pulsed power mode and the ability to choose the switching frequency. Quasiresonant controls and products like the Fairchild’s Green FPS power switches (FSCQxxxx) offer performance for higher power levels but can have higher power losses at low-power levels than other designs. Limited quasiresonant controls and devices such as Fairchild’s FSQ0x65 devices work around this issue by limiting the variability of the switching frequency. Regulation methods are also options with secondary-side regulation being the most accurate and popular. While primary-side regulation offers an inexpensive approach, it is less accurate, but it may be an ideal solution for constant current applications.

Choosing a feedback device will depend on the above choice for primary-side regulation. The reflected voltage from the output winding through the transformer on the VCC winding is the feedback device and it places demands on the coupling of these two windings. For secondary-side regulation it is typical to use an optocoupler biased in a stable CTR region in conjunction with a voltage regulator like the KA431. Devices that combine an optocoupler and voltage regulator are also available if space is an issue.

For diode selection, major issues are breakdown and power losses. The breakdown is at least 20% to 25% greater than the device it is expected to see.

As for the power losses, the concern is with reverse-recovery losses that are incurred by the MOSFET and not the diode. This loss, in conjunction with the conduction loss, often leads to Schottky diodes and output rectifiers being used. As with the other power devices, choose devices based on PD and Rθja .

The basic calculation is Tj = Pd *Rθja +Ta , Rθja = Rθjc + Rθca . Rθjc . For diodes, the PD is typically (forward current x Vf x duty cycle) the switching loss seen by the diode, which is small compared to the conduction loss.

For higher-current, low-output voltage supplies it may be better to replace the output diode with a MOSFET configured as a synchronous rectifier. This will significantly reduce the conduction losses on the output rectifier.

To implement a synchronous rectifier, a MOSFET and a control method is needed. The MOSFET needs to have a breakdown voltage at least 20% to 25% greater than what you expect to see. Please note this is much more than the VOUT . For a typical 12-V output flyback the output rectifier could easily see 60 V so it important to calculate the needed voltage.

Vd = VOUT + (Vdc(max) * (VOUT + VF ))/VRO

Where the output voltage reflected to the primary = VRO and Vdc(max) = 1.414 * Vline(max)

So a MOSFET that is used as a synchronous rectifier needs to have a BVDss > 1.3 * Vd

There are many design options for a flyback converter. A fixed-frequency PWM vs. quasiresonant; integrated vs. separate MOSFET; and a primary- vs. secondary-side regulation are just some of choices a power supply designer must make. Each design option carries its own advantages and disadvantages. The best choice will depend on the specifications and cost target of the end application. ■

For more on flyback power supplies, visit http://www2.electronicproducts.com/Power.aspx.