About MOSFET Selection, Here Must be Your Dry Goods!

10-step rule: Practical MOSFET selection method.

1. Power MOSFET selection Step 1: P tube or N tube?
There are two types of power MOSFET: N channel and P channel. In the process of system design, the N tube or P tube should be selected according to the specific application. P channel MOSFET has fewer models and higher cost. If the voltage at the S-pole connection end of power MOSFET is not the reference ground of the system, the N channel needs a floating power supply drive, transformer drive, or bootup drive, and the drive circuit is complex; the P channel can be driven directly, and drive is simple.

The applications of N channel and P channel should be considered as follows:
(1) such as notebook computers, desktop, and server use the CPU and system of cooling fans, printers, motor driven feed system, vacuum cleaners, air purifier, and electric fan white household appliances such as motor control circuit, the system USES the whole bridge circuit structure, each bridge arm pipe can use P, also can use N tube.
(2) The hot plug MOSFET of the 48V input system of the communication system is placed at the high end, which can use Either a P tube or an N tube.
(3) Laptop input loop series, two back-to-back power MOSFET functions as anti-reverse connection and load switch. The N channel is used to control the charging pump with the integrated drive inside the chip, while the P channel can be used to drive directly.

2. Select the encapsulation type
After the channel type of power MOSFET is determined, the second step is to determine the encapsulation. The principles of encapsulation selection are as follows:
(1) Temperature rise and thermal design are the most basic requirements for packaging selection
Different package sizes with different thermal resistance and dissipation power, in addition to the cooling conditions of the system and the environmental temperature, such as whether there is air cooled, the shape and size of the radiator restrictions, factors such as environment is closed, the basic principle is the guarantee under the premise of the rise of temperature of the power MOSFET and the system efficiency, selection of parameters and the encapsulation of a more general power MOSFET.
Sometimes, due to the limitations of other conditions, multiple MOSFETs in parallel are needed to solve the heat dissipation problem. For example, in the applications of PFC, electric vehicle motor controllers, module power supply secondary synchronous rectification of communication systems, etc., multiple pipes in parallel are selected.
If a multi-tube parallel connection cannot be adopted, in addition to selecting MOSFET with better performance, a larger size package or a new package can be adopted. For example, TO-220 can be changed to a TO-247 package in some ACDC power supplies. DFN8*8 is used in the power supply of some communication systems.

(2) Size limitation of the system
Some electronic systems are limited by the size and internal height of the PCB. For example, the module power supply of the communication system usually adopts the packaging of DFN5*6 and DFN3*3 due to the limitation of height. In some ACDC power supplies, the TO-220 package power MOSFET pin is directly inserted into the root due to the ultra-thin design or due to the limitations of the housing, and the TO-247 package cannot be used for the height limit. Some ultra-thin designs simply bend and flatten the pins of the device, which can complicate the manufacturing process.

In the design of a large-capacity lithium battery protection plate, due to the extremely stringent size limit, most of the CSP package is now used at the chip level, to improve the heat dissipation performance as much as possible, while ensuring the minimum size.

(3) Production process of the company
TO-220 has two types of packaging: bare metal packaging and full plastic packaging. The bare metal packaging has small thermal resistance and strong heat dissipation ability, but in the production process, an insulation pendant is needed, and the production process is complex and costly; while full plastic packaging has large thermal resistance and weak heat dissipation ability, but the production process is simple.
To reduce the manual process of locking screws, some electronic systems in recent years have used clamps to clip the power MOSFET onto the heat dissipation plates, resulting in new packaging that removes the traditional TO-220 upper perforated portion and reduces the height of the device.

(4) Cost control
In the early days, many electronic systems used plug-in packaging. In recent years, due to the increase in labor costs, many companies began to switch to patch packaging. Although the welding cost of the patch is higher than that of the plug-in, patch welding has a high degree of automation, and the overall cost can still be controlled within a reasonable range. In some extremely cost-sensitive applications such as desktop motherboards and board CARDS, the DPAK encapsulated power MOSFET is usually adopted because of the low cost of the encapsulated power.

Therefore, when choosing the package of power MOSFET, we should combine the style of our company and the characteristics of our products, and comprehensively consider the above factors.

3. Select withstand voltage B_VDSS
In most cases, it seems that the selection of power MOSFET withstand voltage is the easiest thing for many engineers to do because the input voltage of the designed electronic system is relatively fixed, the company selects some material number of specific suppliers, and the rated voltage of the product is also fixed. For example, in the laptop adapter and mobile phone charger, the input is 90-265V AC, and the primary power is usually 600V or 650V MOSFET. Laptop motherboard input voltage 19V, usually choose 30V power MOSFET, do not need any consideration.

The breakdown voltage B_VDSS of power MOSFET in the data table has certain test conditions and different values under different conditions, and B_VDSS has a positive temperature coefficient, which should be considered comprehensively in practical application.

It is often mentioned in many materials and literatures that if the maximum peak voltage of V_DS of power MOSFET in the system is greater than that of B_VDSS, even if the duration of the peak pulse voltage is only a few or dozens of nS, the power MOSFET will enter the avalanche and cause damage.

Unlike triode and IGBT, MOSFET power has the ability of resistance to avalanche, and many of the large semiconductor power MOSFET avalanche energy on the production line is full inspection, 100% inspection, which is in the data it is a guarantee of measured values, avalanche voltage BVDSS usually occurs in 1.2 to 1.3 times, and the duration of the usually uS and even mS level, then the last only a few or dozens of nS, far below the peak of avalanche voltage pulse voltage is not damage on the power MOSFET.

Why is it required in the actual design that, in the most extreme case, the maximum V_DS voltage of the power MOSFET must be lower than the B_VDSS, with a certain amount of reduction, such as 5%, 10%, or even 20%?
The reason: is to ensure that electronic systems are productive and reliable in mass production.

In any electronic system design, the actual parameter can have a certain range, it is sometimes hard to ensure that more extremes meet together, thus causing problems to the system, especially under the condition of high temperatures, the other elements of power devices and the system temperature coefficient of the drift can produce some unimaginable problems, derating design margin can reduce damage in these extreme conditions as possible.

4. VTH is selected by driving voltage
The drive voltage selected by the POWER MOSFET of different electronic systems is not the same. The ACDCD power supply usually USES 12V drive voltage, while the notebook main board DC DC converter USES 5V drive voltage, so the power MOSFET with different threshold voltage V_TH should be selected according to the drive voltage of the system.

In the data table, the threshold voltage V_TH of power MOSFET also has certain test conditions and different values under different conditions. V_TH has a negative temperature coefficient. Different driving voltages of V_GS correspond to different on-resistance. In practical applications, temperature changes should be considered to ensure that the power MOSFET is fully switched on and that the peak pulse coupled to the G pole will not be triggered by mistake during the turn-off process, resulting in a direct or short circuit.

5. Select the on-resistance RDSON. Note: it is not the current
Most of the time, engineers are concerned about RDSON because RDSON is directly related to conduction loss. The smaller RDSON is, the smaller the conduction loss, the higher the efficiency, and the lower the temperature rise of power MOSFET is. Similarly, the engineer tries to use existing components from previous projects or material stores without much consideration for RDSON's actual selection method. The temperature rise of the selected power MOSFET is too low. For cost consideration, RDSON larger components will be used. When the temperature rise of power MOSFET is too high and the efficiency of the system is too low, RDSON smaller components will be used or adjusted by optimizing the external drive circuit and improving the way of heat dissipation.

Think about it: if it is a new project and there is no previous project to follow, then how to select the RDSON of power MOSFET?

When designing a power supply system, the input voltage range, output voltage/output current, efficiency, operating frequency, drive voltage, and of course other technical indicators, mainly related to the power MOSFET are these parameters.

The steps are as follows:
(1) Calculate the maximum loss of the system according to the input voltage range, output voltage/output current, and efficiency.

(2) The stray loss of the power loop, the static loss of non-power loop components, the static loss of IC, and the driving loss are roughly estimated, and the empirical value can account for 10%-15% of the total loss. If the power loop has a current sampling resistor, calculate the power consumption of the current sampling resistor. The total loss minus the above loss is the power loss of the power device, transformer, or inductor.
The remaining power loss is allocated to the power device and transformer or inductor in a certain proportion. If it is not certain, the power loss is distributed equally according to the number of components, thus obtaining the power loss of each MOSFET.

(3) The power loss of MOSFET is allocated to switch loss and conduction loss in a certain proportion. If it is uncertain, switch loss and conduction loss are equally distributed.

(4) The maximum allowable MOSFET conduction resistance, which is the RDSON of MOSFET at the highest operating junction temperature, is calculated based on MOSFET conduction loss and the effective value current flowing.

The RDSON label of power MOSFET in the data table has certain test conditions and different values under different defined conditions. The test temperature is: T_J=25℃, and RDSON has a positive temperature coefficient. Therefore, according to the highest operating junction temperature of MOSFET, the RDSON corresponding to 25℃ is calculated from the above RDSON temperature coefficient.

(5) The appropriate type of power MOSFET is selected by RDSON at 25℃, and the MOSFET can be trimmed down or up according to the actual RDSON parameters of MOSFET.

Through the above steps, the model and RDSON parameters of power MOSFET are preliminarily selected.
In many sources and literatures, it is not right to calculate the maximum current of the system, and then perform the forehead reduction, and select the device by the current value of the power MOSFET data meter.
The current of power MOSFET is a calculated value, and it is based on T_J=25℃, and the switching loss is not considered. Therefore, this method is far from the actual application and has no reference value. In some applications where high current surges require short circuit protection, the maximum drain pulse current value and its duration in the datasheet are checked, which is not directly related to the selection of RDSON.

6. Select switching characteristics: C_RSS、C_OSS、C_ISS、Q_G、Q_GD、Q_OSS
Power MOSFET produces switching loss during the switching process, which is mainly related to these switching characteristic parameters. QG affects the drive loss, which is not consumed in the power MOSFET but in the drive IC. The larger the QG is, the greater the drive loss will be.
After the power MOSFET model is selected based on RDSON, these switching characteristic parameters can be looked up in the data table, and then the switching loss can be calculated according to these parameters.

7. Thermal design and check
According to the selected power MOSFET data table and the working state of the system, the conduction loss and switching loss are calculated, and the maximum junction temperature of MOSFET is calculated based on the total power loss and operating environment temperature, to check whether it is within the design range. All conditions are based on the worst conditions and then adjusted according to the calculated results.
Slants big, if the total loss is greater than the allocation of power loss, then to choose other types of power MOSFET, can view than the selection of power MOSFET RDSON of larger or smaller, check again the total power loss, this process is usually to cooperate in step 5 and 6, after several times of check repeatedly, finally determine to match the design models, until meeting the design requirements.
Sometimes, due to the limitation of the product model, products with appropriate parameters cannot be found.

The following methods can be adopted:
(1) Multiple pipes are connected in parallel to solve the problem of heat dissipation and temperature rise.

(2) The power loss is reallocated, and the transformer or inductor and other power components allocate more power consumption. When changing the power allocation, it is also necessary to ensure that the temperature rise of other components meets the system design requirements.

(3) If the system allows, change the way of heat dissipation or increase the size of the radiator.

(4) Other factors, adjusting the working frequency, changing the circuit structure, etc., such as PFC adopting a staggered structure, LLC, or another soft switching circuit.

8. Check diode characteristics
In the bridge circuit, such as the full bridge, half bridge, LLC, and the down tube of the BUCK circuit, there is the problem of the reverse recovery of the internal parasitic diode. The simplest method is to adopt the power MOSFET of the internal fast recovery diode. If the internal fast recovery diode is not carried, the reverse recovery characteristics of the internal parasitic diode should be considered: I_RRM, Q_RR, T_RR, T_RR1/_TRR2, for example, T_RR is less than 250nS. These parameters affect the voltage peak, efficiency, and reliability of the power off. For example, during the startup and short circuit of LLC, the system enters the capacitivity mode. If the diode reverse recovery performance is poor, the direct connection of the upper and lower tubes may be damaged. If the control device has a capacitive mode protection function, this factor is not considered.

9. Avalanche energy, UIS, and DV/DT
Avalanche energy and conditions for testing are described in great detail in the following article. Except for flyback and some motor driven applications, most structures do not suffer from this single voltage-clamp avalanche. In many applications, the combined effects of DV /dt, overtemperature, and high current during diode reverse recovery produce dynamic avalanche breakdown damage. See the following article for more information.

10. Other parameters
The size of the internal RG, the problem of load switches and hot-swappers working in linear zones, SOA characteristics, EMI-related parameters, and so on.