­

Although the introduction of silicon carbide (SiC) and gallium nitride (GaN) technology has already changed the power industry quite considerably, there is still further room for improvement in the wide bandgap devices themselves, supporting components and control systems. To get the maximum benefits from SiC and GaN systems will require all three of those ingredients to be designed to work together as a whole. That is according to Leon Gross, who heads up Microchip’s Discrete Products Group. He also claims that current implementations have been held back because designers lack the knowledge and experience in how to design with wide bandgap materials. Leon took some time to talk to PSD about the optimisation of wide bandgap designs and how today’s problems will be overcome in the future.

PSD - What do you think are the main factors that have stopped wide bandgap designs achieving their full potential?

LG - When wide bandgap materials were first introduced, up until relatively recently, engineers concentrated on the metrics that are important when working with silicon MOSFETs and IGBTs, and these are are different than from the ones that they need to focus on when designing with wide-bandgap products. Wide bandgap devices require a lot more attention on other details, including control of the overshoot, the undershoot, and drain saturation as the device is turned off. These all have different requirements from designing with silicon and new skills are needed. To get around that knowledge deficit, engineers have slowed the switching down so that they can design in the way that they always have, instead of figuring out how to utilize the new materials to get the best out of them.

PSD - Is that situation currently still the case, or are have you seen any improvement?

LG - Up to around three years ago, this was a huge issue, especially for silicon carbide. Engineers just didn't understand how to make SiC products work in their systems. They've gotten over that to a large extent, but still aren't fully taking advantage of the materials. I’ve found that there is still a lack of understanding, especially with the components needed around the switches. Magnetics and other components are needed in order to handle the power ramp up and ramp down, as well as the impact of switching at frequencies of 1 MHz and over, where you end up with EMI problems and start generating RF signals. That needs to be managed better so that it doesn't become a problem around other parts of the circuit. Another thing that is holding the industry back is that we don't see a lot of companies that have both digital and analog expertize. Both are required to optimize those types of circuits.

PSD – Is the situation the same in designing with GaN?

LG – It is very similar with GaN. In addition to that, there are many smaller companies that have a lot of really good, smart people developing GaN products. They will succeed in launching improved switches, but GaN designs require much more than efficient switches. Those companies are mainly focused on the wide bandgap materials and devices themselves, and most don't have the resources or expertise to take care of the control aspect, and that is the main area where engineers need assistance – to make the analog and digital parts of the design work together seamlessly.

PSD - How can Microchip help engineers to overcome these drawbacks?

LG - Microchip may not the first name that springs to mind when thinking of power solutions, but the company already has a broad portfolio of power products, including both GaN and SiC devices. More importantly, we also have expertise in all of the other main areas required to take wide bandgap designs to the next level. Microchip is best known as a company for its huge range of digital products, which include microcontrollers and digital signal controllers (DSCs) along with the software tools, code examples and reference designs required to make them operate in any system, including power designs. DSCs in particular are a combination of microcontroller and digital signal processor (DSP) that have been specifically designed for demanding real-time control applications and widely used in digital power designs over the last decade (note - Microchip recently launched its latest DSC family, which was covered in a previous TechTalk here). As well as the processing unit that runs the control algorithms, Microchip digital controllers have integrated peripherals that can make designs simpler and smaller. The company also has a wide portfolio of discrete products that are used in wide bandgap designs.

Just as importantly as the ability to supply a wide range of products, Microchip has also developed expertise in integrating that product portfolio into a huge variety of designs. That experience allows the company to look at individual applications and design a solution specifically for each task. For example, if we look at data centres, which could be the largest problem facing power designers over the coming years, there are multiple areas that require power solutions. There is the power that comes into the main transformer that feeds the building, distributing that power to different areas in the datacentre, and then feeding individual racks and boards. Each of these areas has different requirements from the switch, the driver and the controller to to provide the greatest efficiency at that point in the data centre power system. Understanding the exact requirements of the individual application allows the design of a tailored solution that includes the perfect switch for it, as opposed to a generic off-the-shelf one. We look at the problem from that system standpoint, and try to solve it at that system level.

PSD - That sounds like you see Microchip moving toward providing complete solutions?

LG - Yes, we started out as a component provider and the company is now evolving into a system solution provider. We don't necessarily know what the individual devices will look like until we take on that application. The switches will be designed as part of a complete system and work best as a part of that solution. In that process, we look for synergies and identify common traits that will allow us to define the switch and then fine tune it for use in other solutions. However, we still have to provide solutions for a few more applications before we get to that point of understanding. When those switches are developed, they will likely go on sale as individual components and be an improvement over the products available now, but to get the best from them will require them being employed in the solution that they were designed for. Only Microchip and a few other companies have the depth of product and experience in all of the technologies required to tackle the issue in this systemic manner that will get the very best out of wide bandgap technologies.

PSD – Do you have any examples of this type of holistic solution currently available?

LG - One example of where the company has taken this more holistic approach is in its 3-Phase Vienna rectifier reference design. The Vienna solution uses Microchip’s dsPIC DSC control, its mSIC MOSFET switches and multiple Microchip analog devices to address both electric vehicle charging needs, as well as high power switch mode power supply applications. The solution can achieve up to 98.6% efficiency at its maximum 30 kW output.

Figure 1; Microchip's Vienna 3-phase power factor correction reference design was designed with a holistic approach