Optimizing Power Supply Efficiency with MOSFET Selection

Get valuable resources straight to your inbox - sent out once per month

We value your privacy

Introduction

When designing power supplies, it is always critical to optimize the design for efficiency. Increased efficiency saves energy, reduces heat generation, and can even reduce the power supply’s size.

This article will discuss how to choose the ratio between the number of high-side MOSFETs (HS-FETs) and low-side MOSFETs (LS-FETs) to improve efficiency in power supply designs.

Figure 1 shows a simplified circuit with an HS-FET and LS-FET

Figure 1: Circuit with an HS-FET and LS-FET

When selecting MOSFETs, power engineers may struggle to distribute the internal resistance of the HS-FET and LS-FET for optimal efficiency.

MOSFET Structure and Loss Composition

When selecting a MOSFET with efficiency in mind, the designer must optimize between conduction loss and switching loss. Conduction loss occurs when the MOSFET is closed, and is due to current flowing through the on resistance. Switching loss occurs when turning the MOSFET on and off, and is due to the fact that the MOSFET does not switch on and off immediately. This is caused by the capacitive behavior of the semiconductor structure within the MOSFET.

A MOSFET is an integrated, multi-component structure that is comprised of multiple MOSFET semiconductor structures connected in parallel. There is a tradeoff between resistance and capacitance. The more MOSFET crystals are connected in parallel, the smaller the on resistance (R_DS(ON)), but the larger the parasitic capacitance. A small R_DS(ON) reduces conduction losses but increases parasitic capacitance, which then increases switching losses.

Selecting MOSFETs requires significant consideration, and onboard testing can require too much time and resources. Therefore, it is more advantageous to create an accurate mathematical model to analyze losses and help with MOSFET selection.

Calculating Conduction Loss

First, calculate conduction loss, as this is a relatively simple calculation. In a single cycle, the conduction loss can be calculated by considering the current flowing through the MOSFET and the ripple current.

For accuracy, consider the relationship between R_DS(ON) and temperature when making this calculation. The MOSFET’s internal resistance is not a fixed value, and it increases as the temperature increases.

Figure 2 shows how to calculate conduction loss, where I_O is the nominal current, ΔI_O is the current ripple amplitude, T_J is the junction temperature, and k is the temperature coefficient.

Figure 2: Conduction Loss Calculation

Switching Loss (Parasitics)

It can be difficult to calculate switching loss. When making this calculation, consider the parasitic inductance caused by the inductors in each loop, as well as the MOSFET’s nonlinear parasitic capacitance at different voltages.

Figure 3 shows the two types of parasitic components to consider when calculating the switching loss.

Figure 3: Two Types of Parasitic Components

Switching Loss (Turn-On Loss)

There are three stages for turn-on loss, described below.

Stage 1 (HS-FET stage): When the HS-FET turns on, the drain-to-source voltage (V_DS) begins to drop. In contrast, the drain-to-source current rises until the HS-FET’s V_DS(TOP) voltage drops to 0V, or the HS-FET current (I_HS) rises to the output current (I_OUT).
Stage 2 (reverse recovery stage): During reverse recovery, I_TOP reaches its peak value, and the LS-FET starts to build up its voltage.
Stage 3 (oscillation stage): Oscillation ends when I_HS stops fluctuating.

Figure 4 shows the switching losses related to turn-on loss.

Figure 4: Turn-On Loss

Switching Loss (Turn-Off Loss)

There are two stages for turn-off loss, described below.
Stage 1 (DS voltage rising): As V_DS rises, I_DS falls. This stage ends when I_DS drops to 0A.
Stage 2 (oscillation): Oscillation continues until V_DS stops oscillating.

Figure 5 shows the turn-off loss.

Figure 5: Turn-Off Loss

Mathematical Model and Analytical Validation

After understanding the circuit’s various losses, it is time to establish a mathematical model.

Set the required circuit parameter values. The MOSFET values should be analyzed using the equations above, as well as their nonlinear parameter compensation values.
Run a transient simulation over at least one switching period for the converter while in steady state operation.
Integrate the voltage and current waveforms to obtain the MOSFET’s switching loss and conduction loss. This can be accomplished using a power probe or by multiplying the current and V_DS waveforms.

Once the model is available, use the data obtained from the power losses found above to obtain the simulated efficiency, and compare that value to the efficiency curve obtained from the board (or the datasheet). If the calculation’s efficiency error is within 0.5%, the model is accurate (see Figure 6).

Figure 6: Mathematical Model Verification

MOS Selection Based on Mathematical Tools

For this article, we used 10 MOSFETs with a total internal resistance of 100mΩ, then used the model to calculate the efficiency curves under different ratios. For example, consider a ratio of 1:9. That would mean there is one HS-FET (high R_DS(ON), low capacitance) and nine LS-FETs (low R_DS(ON), low capacitance).

By comparing the curves, we can conclude that the MOSFET ratio of 3:7 is the optimal specification for 12V to 3.3V, 10A applications (see Figure 7). This efficiency curve cluster shows that different ratios will result in different efficiency curves, even with the same number of MOSFETs. Find the optimal ratio to optimize your efficiency curve.

Figure 7: Optimal MOSFET Ratio

Figure 9 shows how, for the same input and output specifications, as well as the same output current, the minimum loss point in the circuit is found at different MOSFET ratios. Therefore, designers must keep these specifications in mind when selecting the MOSFET ratio.

Figure 9: Optimal Efficiency Comparisons

Conclusion

MOSFET selection is critical for circuit efficiency, and accurate mathematical models can simplify the selection and design of MOSFET transistors. To obtain an accurate model, consider the circuit’s parasitic parameters and verify your results using an efficiency curve.

This article introduced how to select the appropriate HS-FET and LS-FET ratio through accurate mathematical modeling to optimize your power supply’s efficiency. For more information, MPS’s MOSFET drivers and power supply solutions can be found online.