Are you still selecting Power Devices based on their parameters or figure of merits?

It is time to move on...


Abstract

We demonstrate that it is not only possible but also effective to perform a comprehensive analysis that considers thousands of Mosfets available in the market and then identify the best matching devices that meet given criteria or system specifications based on power loss modeling. The power loss modeling approach considers the interaction of the Mosfet and its parasitics with the circuit and is shown to be more effective than selecting Power Mosfets based on their parameters such as on-resistance, threshold voltage, gate charge, gate capacitance or other figure of merit.

Introduction

When selecting Power Mosfets for use in any given circuit application, engineers typically start with identifying Mosfet package, breakdown voltage and then select desired on-resistance, threshold voltage, current rating, gate charge and capacitance as they are considered to be indicative of power losses in circuit. Only after some initial Mosfet options have been identified, engineers model their power loss in the circuit and then select the devices that best meet their criteria.

Although very common, the above approach has two major limitations. One limitation is that Mosfet performance depends on the interaction of Mosfet with the circuit and optimizing parameters such as on-resistance, threshold voltage, gate charge, gate capacitance and figure of merits considered in isolation is not sufficient. Second limitation is that engineers may not be aware of all the Mosfets that are available in the market and their selection criteria considers only a very limited number of devices which often rules out some excellent alternatives. To address these limitations, we present an approach where we model power losses in the circuit for a large number of Mosfets promoted by the key Mosfet manufacturers on their website and then zero-in on the devices that meet the given criteria.

Power Loss Modeling Approach

To demonstrate our approach of selecting the Mosfets based on their power loss, the first step was to create a comprehensive list of Mosfets that were promoted by over 20 manufacturers on their websites. The second step was to tabulate the essential device ratings and parameters of all the Mosfets in our list by going over their datasheets and package drawings where we recorded over hundred parameters for tens of thousands of Mosfets so as to consider all viable devices. We then kept our list up to date to include recent releases from the manufactures. The third step was to model the power losses for each Mosfet under given set of operating conditions. We modeled the power loss by estimating conduction loss (Ids2*Rds(on)*Duty_cycle) and switching loss (0.5*Fsw*Ids*Vds*Switching_Time) where switching times were calculated based on current and voltage rise and fall times as explained in [1]. We then analyzed the Mosfet power loss with respect to common selection parameters such as on-resistance, gate charge, voltage rating, current rating, package size and narrowed our search criteria to shortlist the best matching Mosfets.

Results and Discussion

Please provide your email to download the full paper including all the charts below.
I agree to receive email from DiscoverEE on power device intelligence and product updates.


Figure 1: Calculated Total Power Loss vs. Drain to Source On-Resistance
Figure 2: Calculated Conduction Loss vs. Drain to Source On-Resistance
Figure 3: Calculated Switching Power Loss vs. Drain to Source On-Resistance
Figure 4: Calculated Switching Power Loss vs. Total Gate Charge (Qg)
Figure 5: Calculated Switching Power Loss vs. Gate Drain (Qgd) Charge
Figure 6: Calculated Switching Power Loss vs. Input Capacitance (Ciss)
Figure 7: Calculated Switching Power Loss vs. Output Capacitance (Coss)
Figure 8: Calculated Switching Power Loss vs. Reverse Transfer Capacitance (Crss)
Figure 9: Calculated Total Power Loss vs. Rds(on)*Qg Figure of Merit.
Figure 10: Calculated Total Power Loss vs. Package Area

Conclusion

The above results show that selecting Mosfets based only on their key parameters and figure of merits is not an effective approach for switching circuits as it can lead to large variations in power loss. Similar conclusions were also drawn in previous works as well. The selection approach based on power loss modeling considers the interaction of the device with the circuit and is shown to be more effective in identifying the devices that meet the power loss criteria. Further, we demonstrate that it is not only possible but also effective to perform a comprehensive analysis that considers thousands of Mosfets available in the market and then identify the best matching devices that meet given criteria or system specifications.

Need more information?