As high-power MOSFETs present a significant load to their associated gate drive circuit, efficient operation requires the proper drive. One requirement for the gate driver in high frequency applications is to minimize the effect of parasitic circuit elements by placing the high-current driver physically close to the load. Also, newer power supply control ICs that target high operating frequencies may not incorporate onboard gate drivers at all. Their PWM outputs are only intended to drive the high impedance input to an external gate driver. In addition, the control device may be under thermal stress due to power dissipation, and an external driver can help by moving the heat from the controller to an external package.
Another important attribute for the gate driver is to provide maximum drive to a power MOSFET during the Miller Plateau Region of the switching transition. This interval occurs while the drain voltage is swinging between the voltage levels dictated by the power topology, requiring the charging/discharging of the drain-gate capacitance with current supplied or removed by the driver device.
Gate drivers provide an interface between low-power controllers and MOSFETs. One application is as a high-power buffer stage between the PWM output of the control device and gates of the primary power switching MOSFET.
Texas Instruments' UCC37321/2 family of gate drivers is intended for high-power MOSFETs. They eliminate external circuits, allowing a reduction is space, design complexity, and assembly cost. The two standard logic options are inverting (UCC37321) and noninverting (UCC37322). Typical rise and fall times are 20 ns with a 10 nF load. The supply voltage is 4V to 15V. Typical propagation delay times are 25 ns with a falling input and 35 ns with a rising input. In addition, the UCC37321/2 provides an enable (ENBL) function for better control of driver applications. ENBL (pin 3) was previously left unused in the industry standard pinout. It is internally pulled up to VDD for active high logic and can be left open for standard operation. Fig. 1 is a functional block diagram of the IC.
The inverting driver (UCC37321) generates inverted gate drive signals from controllers that have only outputs of the opposite polarity. For example, this driver can provide a gate signal for ground referenced, N-channel synchronous rectifier MOSFETs in buck derived converters. It can generate a gate drive signal for a N-channel MOSFET from a controller designed for P-channel applications.
Input to each driver should be a signal with a short rise or fall time, which is satisfied in typical power supply applications, where the input signals are provided by a PWM controller or logic gates with fast transition times (<200 ns). The IN threshold has a 3.3V logic sensitivity over the full range of VDD voltages; yet, it is equally compatible with 0V to VDD signals. Inputs can withstand 500mA reverse current without either damage to the device or logic upset.The TrueDrive™ output stage can supply ±9A peak current pulses and swings to both VDD and GND. TrueDrive is a unique hybrid output architecture used in TI BiCMOS MOSFET drivers. It uses bipolar and CMOS transistors in parallel. The bipolar transistors provide the gate drive current during MOSFET transition, and the paralleled CMOS transistors pull the output to VDD and GND rails at the end of the switching cycles. The TrueDrive architecture maximizes current output while minimizing charge for efficient current delivery at low suppy voltages.
The peak output current rating is the combined current from the bipolar and MOSFET transistors. The output resistance is the RDS(ON) of the MOSFET transistor when the voltage on the driver output is less than the saturation voltage of the bipolar transistor. Each output stage also provides a very low impedance to overshoot and undershoot due to the body diode of the internal MOSFETs.
This design inherently minimizes shoot-through current, and the outputs of these ICs can provide high gate drive current where it is most needed — at the Miller plateau region during the MOSFET switching transition. With this drive architecture, you can use the UCC37321/2 in industry-standard 6A, 9A, and many 12A driver applications. Latch up and ESD protection are also included.
In addition to SOIC-8 (D) and PDIP-8 (P) package offerings, the UCC37321/2 also comes in the thermally enhanced but tiny 8-pin MSOP PowerPAD™ (DGN) package. The PowerPAD_ package drastically lowers the thermal resistance to extend the temperature operation range and improve the long-term reliability. Rated operation is from -40°C to 105°C.
High-Side Gate Driver
Linear Technology Corp.'s LT1910 (photo, below) is a rugged high-side N-channel MOSFET driver with a wide operating supply range of 8V to 48V and protection against -15V to 60V supply transients.
The IC can be used as a solid-state circuit breaker or load switch to protect resistive, capacitive, and inductive loads from short circuit and overcurrent conditions. It offers faster response time than electromechanical circuit breakers. It is specifically designed for harsh operating environments such as industrial, avionics, and automotive applications where poor supply regulation and/or transients may be present. Applications include stepper motor, dc motor control, robotics, heavy machinery, and electronic circuit breakers. It has a self-contained charge pump to enhance an external N-channel MOSFET without requiring external components.
It features short-circuit and overcurrent protection and a diagnostic output signal to alert the system via an open-collector fault flag. When detecting an overcurrent condition, the LT1910 turns off the MOSFET and turns it back on after a set time, eliminating the need for external control circuitry. The overcurrent value and timer are user-adjustable. The LT1910 repeats this cycle until the fault is removed or the driver is turned off.
The device uses an integrated precision current sense comparator to detect overcurrent conditions by measuring the voltage drop across an external current sense resistor. A timing capacitor such as a 0.33µF sets the length of time the power MOSFET is held off following a current limit trip. The IC easily interfaces with optoisolators for operation in harsh environments requiring isolation. The LT1910 is offered in an 8-pin SO package.
The LT1910 is rated for the industrial temperature range of -40°C to 85°C. Pricing starts at $2.05 each in 1,000-piece quantities.
Thermal Diode Sensor
Micrel Semiconductor has a low-cost, miniature thermal sensor IC for thermal management using embedded thermal diodes such as those found in CPUs from Intel and Advanced Micro Devices, Xilinx VIRTEX™ FPGAs, and other high-performance devices. The MIC281 Low Cost IttyBitty™ Thermal Sensor senses and reports the temperature of an off-chip embedded thermal diode or commodity transistor, interfacing to its host via an I2C/SMBus™ compatible 2-wire serial bus. Its low cost and tiny SOT23-6 footprint makes it a good choice for cost-sensitive applications like set-top boxes, video game consoles, PCs, appliances, and other consumer goods. Fig. 2 shows a typical application.
The device includes Micrel's proprietary A-D converter, signal processing front-end, and 1-wire diode interface. This gives the MIC281 superior noise performance in both laboratory and real-world testing. This translates into higher system accuracy. Its specified accuracy is ±3°C from 60°C to 100°C.
The MIC281 responds in a failsafe manner to diode faults and features a user-activated low-power shutdown mode. Operating power supply current is 3mA, and the quiescent current in shutdown mode is only 5µA. The supply voltage range is 3V to 3.6V. The clock and data pins are 5V tolerant, regardless of the value of VDD. They will not clamp the bus lines low even if the device is powered down.
To further enhance system reliability, the serial interface includes a timeout function. The timeout prevents the MIC281 from locking up the bus if data transmission errors occur. To preserve existing software investments, the MIC281 is 100% hardware and software backward compatible with existing and planned future thermal management products from Micrel.
The MIC281 uses standard SMBus Write_Byte and Read_Byte operations for communication with its host. The SMBus Write_Byte operation involves sending the device's slave address (with the R/W bit low to signal a write operation), followed by a command byte and the data byte. The SMBus Read_Byte operation is a composite write and read operation: the host first sends the device's slave address followed by the command byte, as in a write operation. A new start bit must then be sent to the MIC281, followed by a repeat of the slave address with the R/W bit (LSB) set to the high (read) state. The data to be read from the part may then be clocked out. The Command byte is eight bits (one byte) wide and carries the address of the associated MIC281 register.
Eight factory-programmed slave address options are available. In 1,000-piece quantities, the MIC281 in 6-pin SOT23 packaging is priced at $0.88 each. Samples are from stock, and an evaluation board is available.
Texas Instruments, Dallas
CIRCLE 347 on Reader Service Card
Linear Technology, Milpitas, Calif.
CIRCLE 348 on Reader Service Card
Micrel, San Jose, Calif.
CIRCLE 349 on Reader Service Card
