Last Time Buy for Silicon – Time to Change to SiC?

By Christopher Rocneanu, Business Development at UnitedSiC and Magnus Phil, Product Manager at Micropower Group

The end of the line for a critical silicon power MOSFET device could have spelled problems for charger manufacturer Micropower Group. They needed an immediate solution to keep their manufacturing up and running.  SiC cascodes from UnitedSiC not only offered a convenient direct replacement requiring only minimal changes to component values, but provided a timely introduction to the extra efficiency, ruggedness and space savings to be gained through long-term commitment to SiC.

Some power-systems designers see silicon carbide semiconductors as an option for high-end applications only – like electric vehicle drives, data center power conversion, or power conditioning for renewable-energy generators. In the past, designing with silicon carbide also presented challenges, such as special gate-drive requirements, that prevent straightforward drop-in replacement of conventional silicon devices.

The reality is somewhat different. Not only have SiC manufacturing processes matured, bringing higher levels of cost effectiveness, but a new type of SiC device has emerged: the SiC cascode. These are available – in addition to SiC MOSFETs and diodes – as a convenient drop-in replacement for existing silicon FETs that enable circuits to be upgraded with little or no net BOM-cost increase.

Sweden’s Micropower Group discovered the potential of SiC cascodes recently, after receiving a last-time-buy warning about the MOSFETs its engineers had chosen for a high-running, high-power battery charger unit. With a forward order book totaling in thousands of units per year, this impending obsolescence of a key component could have created major problems. In the event, Micropower turned these potential problems into clear advantages – for the company and its customers. Now, with valuable experience in designing and testing SiC-based products, the door has been opened to create further new products that take advantage of this exciting new technology.

 

Obsolescence as a Catalyst for Change

Micropower’s Access 100 series is a high-performance 8 kW charger, which has a large customer base in the materials-handling industry. At the front end of the original design sits a bridge converter comprising 12 silicon MOSFETs, arranged in four groups of three transistors in parallel. The chargers were fully developed, tested and qualified and had been in full production for some time – trusted by users, and in demand worldwide – when the last-time-buy notice about the MOSFETs was received.

Immediately, engineers were tasked with finding a solution, but no suitable equivalent parts were available in the market. Seeking an alternative, a superjunction MOSFET of comparable rating was the first candidate. Unfortunately, several prototypes succumbed to unexplained failures that cast doubt on the ruggedness of the transistor’s body diode. IGBTs were briefly considered but were not efficient enough for the chargers to pass the strict California Energy Commission specifications. SiC MOSFETs were also considered but would have required significant changes to the board layout and circuit design to make the most of their performance advantages.

The options appeared to have been exhausted. Until Micropower’s engineers discovered UnitedSiC’s new family of SiC cascodes.

The SiC cascode contains a SiC JFET co-packaged with a low-voltage silicon MOSFET. It can be controlled using the same gate-drive voltages and circuitry as a silicon MOSFET, although overall conduction losses and switching performance are dominated by those of the SiC JFET. Both, of course, are inherently better than any silicon MOSFET with a comparable voltage rating can offer. To be precise, the SiC cascodes’ RDS(ON) per device was less than one-third that of the obsolete parts, while breakdown voltage was 250 V higher. What’s more, they were available in an identical standard TO-247 package, and so were ready just to “drop in”.

Micropower’s engineers needed only to re-optimize the resistor values in the gate-driver circuit and adjust the driver dead-time. Then everything simply worked, even using the original gate-drive voltages and operating frequency. Performance measurements showed the efficiency at typical operating loads had actually increased by 1%. The improvement was even greater at light loads: up by 10%. At these power levels, such an apparently small difference translates into an impressive energy saving of 750 kWh over five years of service. Further refinements to the design included reducing the sizes of snubber components. In addition, two extra “Y” capacitors were needed to meet the original EMI performance. Importantly, there was no net increase in BOM costs.

The final part of the project involved a tough qualification process to ensure that production units could reliably deliver the same performance as the prototypes and survive common hazards in daily operation, such as short-circuits and over-voltages. Over the 12-month program (figure 1), the chargers were exposed to output short-circuits, load disconnections and ‘bouncing phase errors’, each performed 20,000 times. Voltage surges, and thermal overloads to simulate loss of cooling, were also applied repeatedly. No failures occurred during the entire process.Figure 1. Qualification tests applied to Micropower’s SiC-enhanced Access 100 8 kW chargers.

The final result is a significantly improved product that enjoys a newly extended production lifetime. And now – having gained experience in applying SiC technology and seeing just some of the performance advantages achievable – Micropower Group is ready to explore even more of the potential of this rapidly advancing technology. Engineers have already started a new 13 kW charger project using SiC cascodes from UnitedSiC, and the two companies have announced a long-term supply agreement. In the future, new Micropower Group products designed from the ground up using advanced SiC devices could deliver even greater ruggedness and energy efficiency, with much smaller magnetic components and other passives.

It all started with that letter about obsolescence; a catalyst for bigger and better change than anyone anticipated.

Author: Christopher Rocneanu