Additional Information:
Quick View Data Sheet for the MAX5003

Additional Information:
Quick View Data Sheet for the MAX627

Additional Information:
Samples and Literature Order Form

Additional Information:
Powerhelp@design.mxim.com

The two-switch forward converter (Figure 1a) has been around for many years, and it does offer some benefits, such as reducing switch voltage stress and eliminating the reset winding for the transformer. Disadvantages include a limited duty ratio (less than 50%, normally 47% to 48% in practice) and an additional high-side driver. The two-switch forward converter found applications in high-power converters where adding the high-side switch and the driver offset the expensive snubber for the low-side switch. Also, with these high-power converters, there was increased reliability due to less voltage spike on the switches, recovery of the energy stored in the leakage inductance, and less thermal stress. However, better topologies have been invented lately, the most popular being the voltage-clamped forward converter, as shown in Figure 1b.

This topology provides a good scheme for recovering the energy stored in the leakage inductance and for clamping the voltage on the main switch. In addition, the transformer is automatically reset, eliminating the need for reset winding. There is virtually no limitation on the duty ratio of the topology, offering the freedom to optimize the converter design over a wide input voltage, as often seen in both telecom and datacom applications. All of this comes at the cost of higher-than-input-voltage stress on the main switch. However, it is still significantly less than with conventional forward converters with reset winding. As a result, the voltage-clamped forward converter is widely used in DC-DC converter and board-mounted power module designs.

*Figure 1a. The two-switch
forward converter *

*Figure 1b. The voltage-clamp
forward converter*

The single-switch forward converter with resonant reset (which also eliminates the use of reset winding) finds applications in lower-power (20W or less) isolated DC/DC converter modules, primarily because of its simple circuit structure and low cost.

What follows is a comparison of the two-switch forward converter and the voltage-clamp forward converter, using a design example, in terms of voltage stress, input RMS current, conduction loss of MOSFETs, output inductance, and the high-side MOSFET driving scheme. These are the specifications for the design example:

Input voltage: 35VDC to 75VDC

Output voltage: 3.3VDC

Maximum output current: 40 Amps

Switching frequency: 300kHz

**Voltage
Stress **

The voltage stress on the switches of the two-switch forward
converter is always equal to the input voltage. For the telecom market,
the input voltage ranges from 36VDC to 72VDC with nominal voltage around
48VDC to 60VDC. Therefore, a MOSFET with 100V-rated V_{DS} is
plenty for the job. On the other hand, the V_{DS} requirement
for the voltage-clamp forward converter is 150V (applying the 80% derating
design rule), as can be seen from the curve in Figure 2 (in this design,
the maximum duty ratio is 75%).

*Figure 2.
Maximum V _{DS} of the voltage-clamp forward converter versus
input voltage*

**Input
RMS Currents**

Input RMS currents for both converters are given in Figure 3, where it is evident that the voltage-clamp forward converter has lower input RMS current. This translates to lower conduction loss on the primary circuit, as will be seen later.

*Figure 3.
Input RMS current of (a) the two-switch forward converter and (b) the
voltage-clamp forward converter*

**Conduction
Loss of the Switches**

Figure 4 gives the conduction loss for both converters. It is clear that the voltage-clamp forward converter offers lower conduction loss, assuming that the on-resistance of the main switch is the same as the sum of the two switches in the two-switch forward converter.

*Figure 4.
Conduction loss of the switches in (a) the two-switch forward converter
and (b) the voltage-clamp forward converter*

The voltage-clamp forward converter offers lower conduction loss, because there is no duty-ratio limitation to properly reset the transformer. Therefore, for a given input voltage, the converter can be designed to operate at a larger duty ratio, resulting in a larger transformer primary to secondary turns ratio. Consequently, the peak current on the primary side is smaller and so is the RMS current. Because Q2 is used only to clamp the voltage across the main switch (Q1), the conduction loss in Q2 is negligible.

**Inductance
Values**

A good rule of thumb for designing the output inductor is to design the inductor ripple current about 20% to 30% of the rated output current, at the nominal input voltage. In this case, the nominal input voltage is 48VDC. If 20% is used, then the output inductance of the two-switch forward converter is given by:

yielding L = 0.97µH. Similarly, the output inductance for the voltage-clamp forward converter is only 0.723µH, a 25% difference! The reason is very clear: The voltage-clamp forward converter operates at a larger duty ratio, so the off-duty ratio is smaller when compared to that of the two-switch forward converter. Therefore, a smaller inductor is needed for the same output ripple current.

**The
Driving Scheme for the High-Side MOSFET**

For the two-switch forward converter, the high-side driver is required. However, for the voltage-clamp forward converter, the high-side driver is not necessary. This is because the P-channel MOSFET can be used if we can rearrange the circuit a little differently, as shown in Figure 5. This is a very popular schematic in telecom applications, particularly for an input voltage less than 100VDC.

*Figure 5.
The voltage-clamp forward converter with the P-channel clamp MOSFET*

**Conclusion**

With present MOSFET technologies, the difference of R_{ON}
from 100V to 150V is not sufficient to take advantage of using lower-V_{DS}
MOSFETS in the two-switch forward converter for 48V applications. The
duty-ratio limitation (<0.5) of the two-switch topologies is the main
reason for limiting the design optimization. Therefore, the voltage-clamp
forward converter is still the best choice for 48V applications.

Maxim's highly integrated MAX5003 high-voltage switching power-supply controller has all the features and building blocks needed for a cost-effective forward converter. It is readily applied to the voltage-clamp forward converter with an additional dual driver to drive the two MOSFETS, Q1 and Q2. The MAX5003 offers some distinctive advantages, including soft-start, undervoltage lockout, external frequency synchronization, and fast input voltage feed-forward. Voltage-mode control with input voltage feed-forward limits the maximum duty ratio as a function of input voltage, which facilitates the design of the isolation transformer for transient response. The integrated high-voltage internal start-up circuit that operates from a 11V to 110V input range greatly reduces the external component counts. A schematic of the voltage-clamp forward converter with the MAX5003 controller is shown in Figure 6.

*Figure 6.
Schematic of the voltage-clamp forward converter employing the MAX5003
PWM controller and the MAX627 dual MOSFET driver *

December 2000