Half-Bridge Flyback Converters Outperform Conventional Types

Electronicdesign Com Sites Electronicdesign com Files Uploads 2014 08 Ifd2601 F1

Fig. 1: A conventional flyback topology experiences significant switching losses.

Electronicdesign Com Sites Electronicdesign com Files Uploads 2014 08 Ifd2601 Ivankovic

Mladen Ivankovic is a Sr. Staff Application Engineer at Infineon

A classic flyback converter stores energy as a magnetic field in a transformer’s air gap—as long as the converter switching element is conducting (Fig. 1). When the switch turns off, the stored field collapses, and the energy transfers to the circuit’s output as current. Generally, the topology includes a transient voltage suppressor (TVS) to limit voltage on the low-side switch.

In a half-bridge flyback converter with serially connected MOSFETs, the upper switch pair shares the total voltage in the same way that the low-side switch interrupts input voltage (i.e., the bus voltage) (Fig 2). Meanwhile, the high-side switch interrupts transformer-reflected voltage, along with the voltage spike caused by the transformer’s leakage inductance.

Electronicdesign Com Sites Electronicdesign com Files Uploads 2014 08 Ifd2601 Sawyer

Fred Sawyer is a Sr. Field Application Engineer at Infineon

Integrating serially connected switches into the half-bridge flyback helps enhance efficiency because they minimize turn-off losses. This comes in handy for applications dealing with wide input-voltage ranges.

Delving further into this arena, Infineon developed a prototype to study half-bridge flybacks with these serially connected switches. The prototype features an 800-V input/24-V output capable of driving up to 0.33 A.

In looking at applications, the company pinpointed a smart-meter power supply, which requires a converter with low power demands (~8 W) and a wide input-voltage range. A typical flyback topology used in such an application faces limitations arising from MOSFET technology. Foremost among those is the MOSFET’s breakdown voltage, which is on the order of 900 V. As a result, it can’t provide sufficient voltage margin to accommodate transformer-reflected voltage (~200 V) and the peak voltage (~100 V) caused by the transformer’s leakage inductance.

Electronicdesign Com Sites Electronicdesign com Files Uploads 2014 08 Ifd2601 F2

Fig. 2: To minimize losses, connect the high-side diode to the bus with a rectifier diode.

The conventional approach toward designing a reliable switch (i.e., one rated for a V_DS of at least 1300 V) demands two MOSFETs connected in series. Such an approach also must incorporate transient voltage suppression (TVS) using a high voltage Zener diode to limit the voltage on low-side switch. That is, when the low-side MOSFET isn’t conducting, transformer current would be directed through the source-gate capacitance of the high-side MOSFET and the TVS diode.

In that situation, two parallel processes are in operation: discharge of the C_GS capacitor until the high side MOSFET turns off; and voltage-limiting on the low-side MOSFET, accomplished by turning on the TVS diode. In that situation, the losses in the TVS equal the product of transformer current and TVS voltage.

A half-bridge flyback topology, using a rectifier diode that connects the gate of the high-side MOSFET to the bus (Fig. 2, again), offers a better alternative to the conventional approach. Although the turn-off current discharges the C_GS capacitance of the high-side MOSFET, the circuit doesn’t experience any additional losses.

Advantages of the half-bridge flyback include:

• Better efficiency as turn-off switching losses are minimized

• Simple low-voltage drive for high-side MOSFET

• Rectifier diodes can be used instead of a high-voltage TVS Zener diode (D1)

In the conventional solution, prior to turning ON, the MOSFET’s parasitic capacitance would be charged to its breakdown voltage. Taking events in sequence, the switching cycle starts when MOSFET Q1 is turned on by driver V1 (a pulse source).

The next step is to charge the C_GS capacitor of the high-side MOSFET using charge from the parasitic capacitance of D1. Once C_GS is charged, the high-side MOSFET will turn ON. At that point, the input voltage is applied to the primary side of the transformer, and magnetizing current increases until the low-side MOSFET (Q2) turns off.

In the turn-off sequence, while Q1 is off, the V_DS voltage increases until it reaches the breakdown voltage of D1. D1 then starts to conduct, and the low-side MOSFET current (I_DS) reroutes to D1 through the C_GS capacitance of the high-side MOSFET Q2. The I_DS current then discharges C_GS and the high-side MOSFET turns OFF. When V_DS on Q2 reaches the value of the reflected voltage, diode D3 turns ON, triggering transformer secondary current.

A coarse estimate for driving switching losses (P_SW_conv) in this scenario is established with:

P_SW_conv = V_BR_tvs × (Q_g) × f_s

where V_BR_tvs = breakdown voltage of the TVS Zener diode; Q_G = gate charge of high-side MOSFET; and f_S = switching frequency.

The half-bridge topology’s basic characteristic is that neither one of the two MOSFETs “sees” any voltage larger than bus voltage. That’s why the authors termed this topology the “half-bridge flyback.”

The switching cycle starts when driver V1 (pulse source) turns on Q1. The next step is to charge the C_GS capacitor of the high-side MOSFET using the same pulse source V1 and diode D4. When C_GS is charged, the HS MOSFET will be turned ON. Input voltage is applied to the primary side of the transformer, and magnetizing current increases until the low-side MOSFET (Q2) is turned off.

In the turn-off sequence, when Q1 is off, V_DS voltage increases until it reaches bus voltage V2. The rectifier diode (D5) starts to conduct and the low-side MOSFET current (I_DS) then gets rerouted to D5 through the C_GS capacitance of Q2. I_DS current will discharge the C_GS capacitance and the high-side MOSFET will turn OFF. When Q2’s V_DS reaches the value of reflected voltage, diode D3 will turn ON and transformer current will start transfer from the primary to secondary side. Losses occur via conduction at D5, in addition to the MOSFET losses.

A coarse estimate of this topology’s switching losses is calculated by:

P_SW_hbf = V_F × (Q_G) × f_S

Comparison of the two approaches reveals much smaller losses for the half-bridge flyback topology. (Table 1 and Fig. 3).

Electronicdesign Com Sites Electronicdesign com Files Uploads 2014 08 Ifd2601 Table

The half-bridge flyback converter employs a rectifier diode to limit the low-side MOSFET voltage to bus voltage. This improves efficiency by reducing turn-off loss when the high side MOSFET is turning off.

Electronicdesign Com Sites Electronicdesign com Files Uploads 2014 08 Ifd2601 F3 Big

Fig. 3: In a test of the half-bridge flyback, the MOSFET switching at bus voltage (red trace) is 830 V. Note the knee at the start of high-side MOSFET turn-off (yellow trace).

To simplify the driver stage of converter designs, the half-bridge flyback approach can be extended to forward topologies. A half-bridge forward topology, which will handle magnetizing current much like the half-bridge flyback, will require demagnetizing turns on the forward transformer.

REFERENCES

Ilchmann, Bernd, “Ultra Wide Input Range, HV-BIAS Supply for SMPS with ICE2B265 and SPA02N80,” Published by Infineon Technologies AG, March 2008.

Ivankovic, Mladen, “Half bridge flyback and forward” patent application (Serial No. 13/436,109), March 2012. US 8,653,881 B2, Feb. 18, 2014.

Bjoerk, Fanny; Hancock, Jon; Deboy, Gerald, Application Note: AN-CoolMOS-CP- 01 CoolMOSTM CP- “How to make most beneficial use of the latest generation of super junction technology devices,” published by Infineon Technologies AG, Feb. 2007.