How to choose emerging power devices

The emerging market for SiC and GaN power devices is expected to grow 18-fold over the next 10 years. The main demand markets are power supplies, photovoltaic inverters, and industrial motor drives. SiC Schottky diodes have been around for more than 10 years, but SiC MOSFETs, SiC JFETs and SiC BJTs have only recently emerged, and GaN power devices have just appeared on the market. Who will become the protagonist of the emerging power device market in the future? Should we choose them now?

Among these new power devices, we have selected the most representative products one by one and touched their development pulse in comparison to see who will win in the emerging power device market in the future? How should we choose?

High efficiency and high reliability: SiC BJT products can achieve higher efficiency, current density and reliability, and can work smoothly at high temperatures. In addition, SiC BJT has excellent temperature stability, and there is no difference in the characteristics of high-temperature operation at room temperature. SiC BJT actually has the advantages of all IGBTs and solves all bottlenecks in the design of IGBTs. Since IGBTs are voltage-driven and SiC BJTs are current-driven, design engineers use SiC BJTs instead of IGBTs, which may be unaccustomed at first, but device suppliers such as Fairchild generally provide reference designs to help engineers design drivers. line. With the introduction of dedicated driver chips in this area, the use of SiC BJTs will be even more simplified.

Low losses reduce costs: SiC BJT's Vce is reduced by 47%, Eon is reduced by 60%, and Eoff is reduced by 67%. SiC BJT offers the lowest conduction loss in the market, with less than 2.2 milliohms per square centimeter Ron at room temperature. The SiC BJT provides minimal total losses, including driver losses. SiC BJT is the most efficient 1200V power transfer switch ever achieved, SiC BJT achieves higher switching frequency, conduction and switching losses are lower (30-50%) than IGBTs, enabling up to 40% in the same size system The output power increases.

The 2KW step-up circuit from 400V to 800V achieves only 25KHz switching frequency when implemented with silicon IGBTs, and requires 5 film capacitors. When using SiC BJT, not only the switching frequency can be 72KHz but also only needs to be used. With two thin film capacitors, the size of the heat sink and the size of the inductor are all reduced by one-third, which means that a smaller inductor can be used, thereby significantly reducing the total BOM cost of the system.

Increase the switching frequency of the power supply to achieve high frequency: The biggest drawback of traditional IGBTs is the slow switching speed, low operating frequency, and a high current at the time of turning off the tail will cause high turn-off losses. SiC BJT switching speed is fast and no IGBT turns off is the current tail, so the switching loss is very low. Under the same rated voltage condition, the conduction internal resistance of SiC BJT is also lower than that of IGBT, which can reduce the conduction loss.

The best application of SiC BJT is the power supply design with more than 3000W power. Many of these power supplies use IGBTs as switching devices to achieve cost and efficiency optimization. If a design engineer uses SiC BJTs instead of IGBTs, it is easy to significantly increase the power switching frequency, thereby reducing the size of the product and improving conversion efficiency. Due to the increase in frequency, the number of inductors and capacitors required in the peripheral circuit can also be reduced in design, which contributes to cost savings. On the other hand, SiC BJTs have fast switching speeds and can perform switching operations in <20 nS, which is even faster than MOSFETs, so it can also be used to replace MOSFETs.

Compared with bipolar IGBT devices, SiC BJT has lower conduction internal resistance and can further reduce conduction loss. The high temperature stability and low leakage of SiC BJT surpass IGBT and MOSFET. In addition, its internal resistance changes with a positive temperature coefficient and is easily used in parallel for high-power power supply designs.

According to Lan Jiantong, vice president of marketing at Fairchild Semiconductor in Asia Pacific, "Mainly hampered by the impact of manufacturing costs and product yields, the main reason for the current large-scale entry of SiC products into the market is the high price, which is generally about 10 times that of comparable Si products. I personally think that the SiC market will be formally launched in 2013, and SiC BJT devices may become the first product to be accepted by the market in the next 2-3 years, and the product yield of SiC devices will increase substantially in 2015 and the price will also increase. Decline, then SiC products may achieve scale applications."

Fairchild already has a complete product roadmap for SiC BJT products. Now the driver part of Fairchild's SiC BJT solution is still discrete. Next we will first develop the SiC BJT driver IC. The SiC BJT driver is very different from other similar devices in the past. Because the high current requires a special driver IC, Fairchild must develop a proprietary IC to prevent EMC interference. "Blue built copper.

So what are the advantages of SiC MOSFETs compared to SiC BJTs?

The SiC MOSFET was introduced to the market in mid-2010. During this period, many engineers began to contact SiC MOSFETs and their characteristics were also known. SiC MOSFETs are very similar to conventional IGBTs in terms of their use, especially in driving, so replacing IGBTs has some advantages. However, the production cost of SiC BJT is higher than that of SiC MOSFETs. In the long term, what kind of SiC solution will be accepted by the market will depend on the cost. In addition, many design engineers are also concerned about the reliability of the oxidation layer of SiC MOSFET gate oxide in long-term operation, which may affect the working life of the device. However, SiC BJT does not have this gate oxide layer in structure and is reliable. Sex is no such worry.

By 2022, SiC MOSFET revenue is expected to reach 400 million US dollars, more than SiC Schottky diodes become the most popular SiC discrete devices. At the same time, it is expected that SiC JFETs and SiC BJTs will have less than half the revenue of SiC MOSFETs by 2022, although they may have achieved good reliability, price, and performance.

Current end-user preference for SiC MOSFETs is due to cost issues. However, in order to improve product performance, SiC BJT will be the first choice. So one of the major challenges SiC BJT suppliers currently face is how to educate their potential customers to accept these new technologies.

GaN is just starting but lagging behind GaN is a wide bandgap material that offers SiC-like performance characteristics, but with greater cost reduction potential. This cost-performance advantage is possible because GaN power devices can be grown on silicon substrates at a lower cost than SiC substrates.

GaN is dominant in applications below 600V/3KW, and it is possible to replace MOSFETs or IGBTs in these applications. These applications include micro-inverters, servos, motor drives, and UPS.

As the global economic downturn and the price decline of SiC are not as large as expected, the demand market for SiC and GaN power devices has not seen strong growth in recent years. In contrast, the industry’s confidence in GaN technology began to increase as more semiconductor suppliers announced plans for GaN development. For example, Transphorm has become the first.

The key factor in determining the future market growth rate of GaN power devices is how quickly the cost and performance of GaN power devices are comparable to that of silicon MOSFETs. CNT expects this to be possible in 2019, once the industry can achieve this in 2019 We expect that the market for GaN power devices in 2022 will exceed $1 billion.

The road to GaN development has only just begun. Basic device performance represented by the quality factor RQ will be fundamentally improved. With people's further understanding of materials and processes, performance is expected to increase twofold in the next three years, and is expected to increase tenfold in the next ten years. Silicon-based GaN does not need to be packaged, so it can remove all costs associated with packaging, circuit board area, thermal resistance, resistance, and reliability problems often encountered in packaged power devices.

Looking at the contrast of these new power devices, how would you choose?