With over 200 million motors produced every year with fractional horse power, the opportunity to save energy by adopting an electronically controlled variable speed solution is very significant. In fact, 55 percent of total electric energy produced worldwide is used to run motors of every size, shape and efficiency. Variable speed drives can save as much as 70 percent of energy costs by varying the speed of the load. New energy efficiency standards for fans and pumps require the adoption of electronic controlled motors and present further challenges to electronic designers.
Below a 200W power range, where the motor is operated for the vast majority of the time at a fraction of the maximum power, FredFETs rather than IGBTs offer greater efficiency advantages and a more economical approach. (For more information on FredFETs, see the box “What is a FredFET?”
IR’s µIPMTM product family is the first high-voltage multichip solution based on standard QFN packaging technology. This new alternative approach pioneered by IR (patent pending), utilizes PCB copper traces to dissipate heat from the module, providing cost savings through a smaller package design and in some specific applications through eliminating the use of an external heat sink.
The newest family of µIPMTM enables designers to solve the technical challenges of designing advanced energy saving solutions, in line with the most recent energy standards for fans (heating and ventilation) or pumps (water circulation), up to 200W with a highly efficient, ultra-compact, highly integrated QFN-based solution.
A Market Perspective on New Trends
In Europe, the key legislation driving the move to increasingly efficient fan and pump applications is Directive 2009/125/EC: the European Energy Related Products (ERP) Directive. Designed to minimize energy consumption and support lower carbon emissions by improving energy efficiency, this legislation replaces the Energy Using Products (EUP) Directive and provides a framework for establishing minimum eco-design requirements for energy-using and energy-related products. The Directive is being implemented in various stages, with the key dates for fans and circulator pumps being January 1st 2013 and January 1st 2015, as defined in regulations EU 327/2011 and EC 641/2009, respectively.
The impact of this new legislation should not be under-estimated since, for example, between 30 percent and 50 percent of fans currently in the market will not be compliant with ERP requirements and fan motors consume approximately 20 percent of the total electricity produced in Europe. The situation could be even worse for the pumps employed in heating and air conditioning systems for which the directive requires an energy efficiency index (EEI) <0.27 going to <0.23 by 01/01/2015 (today’s Class A products are at <0.40).
The need to optimize the efficiency at light load (less than 50 percent of full load ratings of the motor) requires the utilization of semiconductors that are best suited to offer lower power dissipation at a fractional level of the maximum load current. Within a power level up to 200W, FredFET technology is particularly fit for this range of application by presenting an R×A FOM (figure of merit) of about 6.9ohm-mm2 for nominal blocking voltage of 500V, similar to traditional IGBT technology. Fig. 1 compares the conduction characteristics of different IGBT technologies and IR FredFet technology at 150°C.
The presence of threshold voltage in the IGBT conduction characteristics penalizes IGBT switches in overall conduction losses when compared to MOSFET. It is also important to note that in order to minimize the switching losses to a level comparable to a typical freewheeling diode normally used with IGBTs, the FET body diode needs to be designed and processed to reduce the switching recovery time and recovery charges. To achieve a fast recovery characteristic, the wafers are submitted to electron irradiation. The result is an intrinsic body-diode with fast recovery characteristics optimized for switching operation below 6~7Khz. Fig. 2 compares the reverse recovery time for an ultrafast recovery diode and a FredFET body diode.
By utilizing a FredFET. the multi-chip device’s total component count is reduced from 13 to 7 chips, hence simplifying the assembly process.
Compressor drives in refrigeration are classic applications in which efficiency at light load is considered fundamental since, for the vast majority of the time, compressors operate at a fraction of the nominal load.
By using a FredFET instead of the more traditional IGBT+Diode combination, higher electrical efficiency can be obtained at light load as shown in Table 1 that compares the results of an advanced appliance compressor inverter drive operating at 5Khz with a DC bus of 300V.
The µIPM™ PQFN (Fig. 3) is an example of a fully integrated inverter solution to utilize the PCB as a heat sink, very similar to well adopted module solutions in point-of-load and VRM applications but for the first time not limited to low-voltage products. The µIPMTM family also offers a new benchmark in size compared to any equivalent competitor solution, up to 60 percent smaller than the existing dominant 3-phase motor control Power IC.
Housed in QFN-like packages, the µIPM™ family comprises a series of fully integrated 3-phase or single phase (half-bridge) motor control circuit solutions.
As in point-of-load or VRM QFN based packages, the µIPM’s power semiconductors (500V FredFETs) and HVIC die are also bonded to the lead-frame which is exposed and soldered to the PCB. Smaller dimensions and power dissipation through the PCB, present some challenges that need to be addressed to fully achieve the optimal overall performances of the motor drive design. Fig. 4 shows the µIPM package. An advantage of up to 60 percent savings in total footprint area is clearly important for the constant demand to lower material costs, dimensions and overall weight containment.
By utilizing an innovative packaging solution, the µIPM family delivers a new benchmark in device size, offering up to a 60 percent smaller footprint than existing 3-phase motor control IPM.
Available in an ultra-compact 12x12x0.9mm PQFN package, the µIPM family comprises a series of fully integrated 3-phase surface-mount motor drive circuit solutions. The new alternative approach leverages the PCB copper traces to dissipate heat from the module, providing cost savings through a smaller package design and, even eliminating the need for an external heat sink. Fig. 5 shows the cross-section of the µIPM
As in point-of-load or VRM QFN based packages, the µIPM’s power semiconductors (500V FredFETs) and HVIC die are also bonded to the lead-frame which is exposed and soldered to the PCB.
In general, IPM current capability depends on the DC bus voltage, the ambient temperature, the switching frequency, and for all these elements, the higher they are, the higher the losses are), the modulation scheme (i.e. 3-Phase vs. 2-Phase), dV/dt of phase voltage and, of course, the FET characteristics (RDS(ON), IRec etc.). Fig. 6 shows the current capability vs. Frequency (3-phase and 2-phase modulation technique case)
In the case of a surface-mounted solution like the one offered by the µIPM family, the current capability also depends on the PCB design and specifically to the copper thickness, the copper pad areas, the number of layers and ultimately by the maximum allowable PCB temperature. In other words, the maximum junction temperature of the power semiconductors is, in fact, less of a concern than the maximum PCB temperature. Fig. 7 is a thermal photograph of the µIPM.
By increasing the PCB copper thickness the overall thermal resistance Junction to Ambient is lowered and consequently, the PCB temperature. Lowering the temperature enables higher current capability. Fig. 6 shows the direct impact of the increased PCB copper thickness from 1oz. to 2oz, to the current capabilities.
Output current capability increases with higher ΔTCA and also increases when a 2-phase modulation versus a 3 Phase modulation scheme is used. Similarly by lowering the switching frequency, lower switching losses enable higher output current.
All topologies based on silicon-based switches have intrinsically limited improvement capabilities. Based on state-of-the-art active components and passive components, constrained integration opportunities pose a limit to the technology evolution. For future integrated modules, GaN based switches have a better potential figure of merit than other power components based on Si or SiC material. Fig. 8 shows the relative comparison of the specific RON FOM for Si, SiC and GaN based switches. The potential improvement achievable from GaN technology is significant based on the material limits.
To improve overall conversion efficiency, all of these topologies require the power switch with the lowest possible specific RON×QG Figure of Merit. GaN-based MOSFETs show great potential in FOM improvement over the coming years.
Preliminary tests have proven that GaN-based power device technology has a positive impact on the overall efficiency and size also in traditional inverter topologies. In conjunction with a radical improvement in parasitic charges Qg and Qgd when compared to the best-in-class silicon, GaN-based switches have delivered >3x power loss reduction.
Lower Switching Losses
The first generation of GaN based switch prototypes have 20x Lower Qrr compared to an IGBT Copak and more than 200x less than that of Super Junction body diode. Eoff (off switching loss) of these GaN Switches has been measured 72% less than for IGBT and 30% lower than HV Super Junction as shown in Fig. 9.
For the inverter stage, where IGBTs are today the standard, the improvement of FOM: VDS(ON) x ETS (Conduction Voltage drop multiplied by Total Switching Energy) provided by GaN based switches is almost 3x at nominal current density of 2A/mm2. Fig. 10 illustrates the progress of the FOM of IGBTs and GaN based switches.
Because GaN technology is characterized by an intrinsic lateral structure, it enables flip-chip packaging and virtual elimination of parasitic elements due to wire-bond stray inductance and parasitic resistance. Moreover, the device’s lateral structure will eventually enable the potential to monolithically integrate multiple power devices and drivers. This will drive the technology roadmap to smaller, more efficient, and more economical packages.
FredFETs are fast recovery diode FETs are designed to provide a very fast recovery (turn-off) of the body diode, making it more suitable than traditional HV MOSFET for driving inductive loads such as electric motors. Fig. 11 shows the FredFET’s construction.
When the body diode is forward-biased, the N-epi layer of the device becomes flooded with holes. During the transition from the conduction to the blocking state, these holes need to be removed. This takes a finite amount of time and is known as the reverse recovery time (trr) of the diode.
Increased injection, to reduce the forward voltage drop, implies more charge that needs to be removed from the intrinsic region before the diode will be able to block voltage. This, therefore, adversely affects the reverse recovery time.
By introducing recombination centers in the n-epi layer, the trr of the device can be significantly improved.
This is done using a life-time killing technique – in this case electron irradiation after passivation layer is deposited.