How power modules address unique appliance requirements

By Ki-Young Jang, Yo-Chan Son, Man-Kee Kim, Sung-Il Yong and Bum-Seok Suh
Fairchild Semiconductor

Thanks to the improved performance and efficiency they provide, the demand for inverterized drives and brushless DC (BLDC) motors continues to grow. These motors are widely used in consumer applications such as inverterized air conditioners, washers and fan motors. A BLDC-based inverter system is faster, quieter and more energy-efficient than conventional solutions such as brush DC motors or AC induction motors with on/off control. The key to enabling this progress is inverter technology, especially the integrated power module. Today, power module technology that integrates power switches and their gate-driving circuits can provide a compact, reliable and cost-effective inverter solution.

In addition to meeting the requirements of ever smaller, energy-efficient motors, another challenge for designers of home appliances is the ability to deliver power in a manner that the driver can handle appropriately. Power Factor Correction (PFC), EMI and thermal resistance are important considerations in applications that have to meet international standards for efficiency.

This article describes the benefits and techniques of applying "smart power" module technologies to two different motor drive applications. It gives two design examples of Fairchild Semiconductor's Smart Power Module (SPMTM) technology and provides test results of induction motor designs for inverter systems. It also enumerates the many benefits of this technology and explains how these modules support the package and design requirements within the consumer market that requires cost-effectiveness and high performance¹.

Harmonic distortion in higher power applications
Harmonic current that is generated by a non-linear load deteriorates the power quality of a utility, and is a source of electrical interference for equipment that is connected to the point of common coupling. Today, this problem is regulated by domestic or international regulations, such as IEC61000-3-2, and compliance to these regulations is mandatory. To meet these regulatory requirements, it is necessary to have a pre-regulator instead of a diode rectifier to reduce the distortion of input current. Air conditioners, whose power ratings range from 1kW to 4kW, are classified as Class A equipment according to IEC61000-3-2 (input current is less than 16 Arms), and most of these air conditioners are adopting power factor correction (PFC) circuits as the pre-regulator to mitigate harmonic distortion.

PSC circuit approach to PFC
One of the more cost-effective approaches to PFC uses a "Partial Power Factor-Correcting Switching Converter" (PSC) circuit. With this method, the circuit topology is the same as that of the commonly used high-frequency PFC circuit, yet the switching frequency is just twice the utility frequency. Compared with the high-frequency PFC, the performance of a PSC circuit is limited—and not suitable for—high-power air conditioners over 3kW since it limits harmonic regulation. However, most room air conditioners are below 3kW, so this method provides acceptable performance as well as low EMI due to its low switching frequency. Moreover, because of its simplicity, the PSC method can be controlled by the microprocessor used for inverter control without requiring a dedicated control IC. For this reason, the PSC approach is popular for 1kW to 3kW room air conditioning systems, and its variants are widely adopted achieving approximately 97 percent of the input power factor.

New PSC module concept
To address Power Factor Correction requirements in 1kW to 3 kW air conditioning systems, Fairchild Semiconductor has developed a new integrated PSC module, the PFC-SPM. This PSC module employs circuit topology similar to the high-frequency switching method but offers cost advantages for these specific applications.

Figure 1(a) is the external view of the PSC module's package and Figure 1(b) shows the cross section diagram of the PSC. The PSC module is a direct-bonded-copper (DBC) based transfer-molded package. A typical lead-frame-based package, presents a difficulty when changing the circuit topology since the lead-frame has to be changed. A DBC substrate, on the other hand, allows designers to easily create a new topology without sacrificing cost. Moreover, due to the thickness (0.68mm) of the aluminum oxide isolation layer, it is possible to for the DBC-packaged module to offer very low thermal resistance while maintaining an isolation voltage of 2.5kV for 1 minute.

Figure 1a: PSC Module. (a) package view

Figure 1b: PSC Module. (b) cross section diagram

Figure 1c: Internal block diagram

By virtue of its DBC substrate, a PSC module can share the same package with Fairchild's Motion -SPM series, which has a three-phase inverter topology. Sharing a package with the Motion-SPM device has potential advantages in the assembly of an inverter PCB. For example, Figure 2 shows a Motion-SPM and a PSC module mounted on the same PCB board. Unlike discrete packages, such as a TO-220, smart power module packages are either non-standardized or they have their own packages. With a power module, the heat sink mounting can be very troublesome when different modules have to be placed on a single PCB, which may lead to decreased assembly productivity. The adoption of a PSC in the inverter PCB that uses a Motion-SPM can increase productivity.

Figure 2: PCB mounting with SPM3

Referring to the internal block diagram in Figure 1(c), the PSC module integrates two IGBTs for boost converter operation and four rectifier diodes as shown in Figure 1(b). This module also has an optional thermistor and a gate-driving low-voltage integrated circuit (LVIC) that provides a protection function. In the PSC circuit, the boost inductor and switch are installed at the AC input of the rectifier (Figure 3). The PSC module can be controlled by tying together the input signals IN(R) and IN(S) for IGBT Q1 and Q2.

Figure 3: Operation of PSC with Vin> 0

Circuit operation
When the input signal is applied to PSC, the inductor current flows through Q1 and D4, as indicated in Figure 3

if Vin is positive. If Vin is negative, Q2 and D3 will conduct. At zero crossing of the input voltage, Q1 or Q2 (depending on the Vin sign) is turned on, and the inductor is charged. After turning off the IGBT, the inductor is discharged, and the current will flow to the load and the DC capacitor, as depicted in Figure 3. Since the IGBT is turned on alternately at every period, the power dissipation of each IGBT will be half that of a typical full switching PFC circuit. For this reason, the IGBT current rating can be half. When combined with the DBC-substrate package characteristics, it is possible to use a much smaller IGBT in the PSC circuit.

Moreover, at the moment of IGBT turn-on, the collector current will start from zero, so there is virtually no reverse recovery current and no turn-on switching loss. This means that normal rectifier diodes can be used for the upper diodes: D1 and D2, which have small forward voltage drop and large recovery current. In the PSC example, all four diodes are normal rectifier diodes that provide smaller conduction loss than the fast recovery diode D5 in typical full switching PFC circuit.

By understanding and utilizing the merits of the substrate package and the control method, it is possible to increase performance while decreasing the cost of the module. In addition, the PSC uses an AC reactor and provides protection function in LVIC. It also has an optional thermistor for ease of temperature monitoring, which is another advantage that the PSC brings. Figure 4 illustrates the typical application circuit of the PSC.

Figure 4: SC application circuit (Click to view full image)

PSC module—experimental results
To verify the capability of the PSC module, a unit was used to control a 1.8kW air conditioner with a maximum power of 2.4kW. Figure 5 shows the experiment results. Here Vsh is the voltage across the shunt resistor (Rsh = 25 m?) in Figure 4 used for current sensing. Under the operation conditions shown in Table 1, the case temperature was maintained at 48°C where the ambient temperature is about 30°C.

Figure 5: Experimental results. Time: 2 ms/div

Table 1: Operating conditions for the experiment

Using this simple scheme, it is possible to meet the regulation limit in most of the low-frequency regions (f

This example shows that the PSC power module provides a compact and reliable solution of front-end rectifiers in the 1kW to 3kW inverter used in air conditioning systems. For much lower power appliance motors, a different approach is required.

Integrated MOSFET inverter module for low-power appliances
Motor drive applications below 100 watts, such as small air conditioners, air purifiers, dryers and dishwaters require a different design approach. For these lower power applications, a MOSFET is typically chosen over an IGBT. Moreover, the body-diode of a MOSFET can be used as a fast recovery diode (FRD) essential in IGBT inverters, reducing the number of components and the drive system cost. Using MOSFET technology for the output devices, Fairchild developed the Smart Power Module, the Motion-SPM inverter module, to control up to 0.1kW BLDC motor applications. This module consists of six MOSFETs and three dedicated gate-driving high-voltage integrated circuits (HVIC) in a new package as a compact, reliable and environmentally compatible solution for small motor drive systems.

Advantages of an integrated MOSFETFor small motor drive systems, under 100W, a MOSFET is better suited to other power transistors in terms of power dissipation, cost and performance. The forward characteristic of a MOSFET is ohmic and its conduction loss is proportional to the square of the drain current. Therefore, the conduction loss below 1A can be less than that of an IGBT with the same rating since an IGBT has the threshold voltage at its on state. The significance of the IGBT's voltage drop increases as the inverter output power decreases. Most fan motors used in air-conditioner applications are below 50W and a MOSFET inverter can provide better efficiency than an IGBT inverter in such cases.

Another advantage of a MOSFET is its ruggedness. The MOSFET is more rugged than an IGBT and provides a wider area of safe operation (SOA) compared to other devices with the same rating. Each MOSFET in the Motion-SPM module can sustain short-circuit current for 80 microseconds under the typical operating condition (Vcc=15V, Vdc=300 V, Tc=25°C) as shown in Figure 6. Moreover, under a surge condition, the MOSFET inverter is stronger than IGBT inverter with the same voltage rating, which is proved by the avalanche rating of the switching device. For this reason, for 200V utility line voltages, a 500V MOSFET can be used; whereas, under the same condition, a 600V IGBT is required. However, a conventional MOSFET exhibits an extremely fast switching speed.

Figure 6: Short-circuit capability of MOSFET in module (a) Test circuit

Figure 6: Short-circuit capability of MOSFET in module (b) short-circuit withstanding time

Dead time and delay time
In order to get the best performance for dead time and delay time without incurring instability (dV/dt induced shoot-through), it is necessary to have a MOSFET customized for the motor drive. Besides adjusting gate resistors, optimal selection of Qg and Vth of MOSFET is necessary. In the module, the Qg ratio(=Qg/ Qgs) of the MOSFET has been set at approximately 2.0 to prevent the shoot-through problem under the worst-case operating condition. Based on this gate charge information, the range of applicable gate resistance was decided. Delay times of the power MOSFET are logarithmic functions of Vth. For this reason, the variation range of Vth plays a big role determining the worst-case delay time and the dead time. While meeting these requirements, the output dV/dt should be small to minimize EMI. The switching characteristics are the result of meeting 2kV/?s dV/dt, 1.0 microseconds dead time and 2.5 microseconds on-delay from signal input to current stabilization, and under the worst-case operating conditions, considering the dispersion of gate resistance and other device parameters. This has been achieved by selecting the appropriate turn-on gate resistance and threshold voltage.

In addition to these predetermined characteristics, the module allows users to control switching speed. As is the case with other SPM series products, the module provides open-source terminals for high-side MOSFETs, allowing users to implement their own impedance cell to customize the switching speed of the high-side MOSFET for optimal trade-off between switching loss and EMI.

Packaging and high power density
The module shown in Figure 7(a) integrates six MOSFETs, as well as three half-bridge HVICs dedicated to MOSFET characteristics—all within a transfer-molded full-pack package as shown in Figure 7(b). This module is fabricated using the same transfer-mold process typically applied to normal integrated circuit packaging. Power MOSFET and HVIC chips are attached on the copper lead frame. Electrical interconnections between the chips and the lead frame are made using gold wires and an ultrasonic bonding process.

Gold (Au) wires are used to connect the MOSFET source terminal as well as the signal connection. In conventional power modules, aluminum wires are usually used for the high-current path for a high fusing current level. Gold wires are used for signal-level interconnection since the wire diameter is usually small, making its current capacity relatively small as well. However, in small power applications, the current is low enough to use gold wires in bonding even for its fusing current. Consequently, the bonding process is simplified, enabling streamlined production. After wire bonding, the module undergoes epoxy molding and then the leadframe is trimmed to form individual pins, and prepared for device insertion on the PCB.

Figure 7: MOSFET module (a) Block diagram

Figure 7: MOSFET module (b) package outline and cross-section of the module

The primary goal of the package design is to provide maximum thermal performance within a restricted installation space. A small motor application usually integrates its control and driving circuitry inside its chassis. As motor technology improves, the diameter and the length of a motor rotor shrinks, reducing available space for control circuit. The power module inside such a built-in motor should be as small as possible while delivering the required power to the motor. That is, its package should deliver the power dissipation of silicon chips while maintaining the junction temperature of the semiconductor below the absolute maximum rating (125°C) within the space limitation.

In 100W motor drive systems, if the inverter efficiency is 90 to 95 percent, the power devices will generate 5W to 10W power dissipation. In this condition, the package is expected to deliver the required power without exceeding the maximum junction temperature of the semiconductor chips.

Considerations from an applications perspective—experimental results

A verification experiment was conducted to check the power rating of the module using a 130 W BLDC motor (sinusoidal back-EMF) with the circuit setup shown in Figure 8(a). The effective surface area of the heatsink used in this experiment was about 100cm?. With this heatsink, the module could deliver 150 W electrical output power to the motor with 12W of power dissipation when 20kHz space vector PWM was used. At this point, the case temperature of the module was 86°C, the MOSFET junction temperature 1040C and ambient temperature 27°C. When discontinuous PWM was used in the same condition as shown in Figure 7(b), it was possible to obtain 8W of power dissipation and 95 percent inverter efficiency due to the reduction in effective switching frequency. The case temperature was 62°C and the junction temperature was 74°C. The overall system loss was 82 percent, taking the iron and core loss of the motor into account, and the inverter loss was only 27 percent of the total system power dissipation.

Figure 8: Application example of power module (a) application circuit example (Click to view full image)

Figure 8: Application example of power module (c) variation of bootstrap voltage of HVIC

Bootstrap considerations

Another interesting merit of a MOSFET inverter is in the variation of the bootstrap voltage as shown in Figure 8(c). In this case, the motor was running at 10Hz. The bootstrap voltage is depicted with the output current of the inverter. When the current was positive, the bootstrap voltage VBS was maintained close to VCC=15V, but it dropped to almost 10V when the current was negative. This is caused by different charging mechanism according to the current direction as shown in Figure 9. When the output current is positive, it flows through either a high-side MOSFET or a low-side body diode. In this case, the bootstrap capacitor CBS is charged when the low-side body diode is turned on as shown in Figure 9 (a).

Figure 9: Bootstrap voltage charging mechanism according to current direction (a) Positive output current

Figure 9: Bootstrap voltage charging mechanism according to current direction (b) negative output current

If the current is negative, the charging voltage Vchg will drop much according to the output current due to the forward drop of MOSFET when the low-side MOSFET serves as the active switching device. The bootstrap voltage is the gate-driving power supply of the high-side MOSFET, and has the meaning only when the current is positive. When the current is positive, the bootstrap voltage does not vary much due to the low Vf of MOSFET, eliminating the need to use large bootstrap capacitor.

One can use a small bootstrap capacitor to hold the bootstrap voltage against the standby current of HVIC only when the output current is positive. In the case of over-modulation high-speed motor operation, the high-side MOSFET can be fully turned on for a half period of output frequency.

Identifying application requirements
For efficient consumer appliances, addressing unique design requirements in a cost-effective manner is pivotal. This article has explored two approaches to simplify design and optimize efficiency by using Power Factor Correction (PFC) techniques and designing with integrated smart power modules. These application-specific modules for high and low power extremes provide enhanced performance and complement the existing portfolio of smart power modules developed by Fairchild, which covers the full range of 50W to 3kW appliances.

References
¹ Jun-Bae Lee, Ki-Young Jang, Dae-Woong Chung and Bum-Seok Suh, "New smart power modules for very low-power drive applications," in Conf. Rec. IEEE-PESC'03, 2003, pp. 436-441.