Multicore safety MCU targets automotive
TOKYO — In efforts to boost its lofty status in the automotive MCU market share, Renesas Electronics Corp. unveiled earlier this year what it calls “the world’s first on-chip flash memory microcontroller using a 28-nm process technology.”
While tasking the company’s own R-CAR SoCs to perform sensing and cognitive functions, Renesas is counting on its automotive MCU to further penetrate the growing market of connected, autonomous, and electric vehicles.
In the era of autonomous driving, the responsibility of safety MCUs is paramount. The MCU plays a crucial safety role in sensing, braking, and steering. Renesas’s new MCU, dubbed the RH850/E2x series, comes loaded with safety functions, including six 400-MHz CPU cores. Four of them, individually, are paired with their own checker CPU cores and two additional CPU cores without checkers.
Renesas has also been spending R&D resources to develop memory cell technology for the flash memory integrated into high-performance and high-reliability MCUs. While the company’s new on-chip flash memory MCU is fabricated in 28-nm process technology, Renesas claims that its team is advancing to the world’s first e-Flash Technology in leading-edge 16-/14-nm FinFET, as presented in technical papers at IEDM 2016 and 2017.
Renesas’s new MCU featuring a built-in flash memory of up to 16 megabytes (MB) is particularly important, said Renesas, now that over-the-air (OTA) software updates are poised to become prevalent in many connected vehicles.
Consider Tesla’s Model 3, subject of a new Consumer Reports assessment. The report pointed out that big flaws in Model 3 are “long stopping distance” in an emergency braking test and “difficult-to-use controls.” Responding to Consumer Reports, Tesla CEO Elon Musk decided to deploy software updates via OTA to the Model 3 fleet this weekend, triggering changes to the braking system.
Obviously, a large memory is prerequisite to storing the new software. But what matters more is whether the flash memory is built right onto an MCU and if the built-in memory comes with two memory banks, according to Tadaaki Yamauchi, vice president at Renesas.
Speaking of OTA, he explained, “Ideally, you want to upload the new program while you are executing on it (‘no-wait OTA’).” Furthermore, he said, you want to keep the previous program “so that when a problem happens as a result of the update, you can always go back using the previous one.”
At Renesas’s headquarters here, EE Times recently caught up with Yamauchi, who now heads the company’s automotive control solution division. He used to be an engineer developing embedded flash technology. The following is part of our conversation.
Continue reading the conversation on Embedded's sister site, EE Times: "Renesas guns for extreme safety, No-Wait OTA."