News

MCU Roundup: 3 New MCUs Push Performance at the Edge

April 02, 2024 by Aaron Carman

From small appliances to edge AI, the latest MCUs from Renesas, Ambiq, and Nuvoton provide higher performance and efficiency.

As AI and edge devices become more prolific, companies must release new microcontrollers (MCUs) to support higher computing power and novel use cases. In this roundup, we will examine three newly released MCUs from Renesas, Ambiq, and Nuvoton, highlighting their features, performance, and applications.  

 

MCU roundup

 

Renesas Releases Custom RISC-V MCU

Up first, Renesas has recently announced its newest 32-bit RISC-V-based MCU: the R9A02G021 MCU family (datasheet linked). The new family builds upon Renesas’ existing RISC-V portfolio, including its ASSP and RZFive devices, but includes an in-house-developed RISC-V core. Renesas says it designed the R9A02G021 family from the ground up to target applications including IoT sensors, consumer electronics, and small appliances.

 

Block diagram of the R9A02G021

The R9A02G021 family includes many different peripherals and a custom-made RISC-V core. Image used courtesy of Renesas
 

Alongside its custom 48 MHz RISC-V core, the MCU includes 128 KB code Flash, 4 KB data Flash, and 16 kB SRAM. The device works across a wide operating temperature range from -40°C to 125°C, with supply voltages from 1.6 V to 5.5 V and offerings in QFN and WLCSP packages. 

In addition to the MCU, Renesas has also released an example design using the R9A0G021 family. 

 

Ambiq Supports Power-Efficient AI at the Edge

Ambiq has released its Apollo510 MCU, purpose-built to provide designers with power-efficient AI hardware for the edge.

The Apollo510 leverages Ambiq’s Subthreshold Power Optimized Technology (SPOT) to improve the MCU's efficiency. Early reports from Ambiq show up to 30x improved power efficiency and 10x improved speeds compared to previous generations, all the while maintaining a small form factor for edge devices.

 

Block diagram for the Apollo510 MCU Apollo510

Block diagram for the Apollo510 MCU Apollo510. Image used courtesy of Ambiq
 

Under the hood, the Apollo510 leverages an Arm Cortex-M55 processor, along with several other security features to ensure an easier development experience. The device also includes 4 MB NVM and 3 MB SRAM to ensure sufficient memory for AI applications. 

The Apollo510 is currently sampling, with general availability expected in Q4 of this year.

 

Nuvoton Introduces Endpoint AI Platform

Lastly, Nuvoton recently announced its endpoint AI platform, which is built to expand the AI ecosystem to the microcontroller domain. The new platform uses some of the latest Nuvoton products and promises to aid designers in developing and deploying advanced AI/ML models to the edge.

The NuMicro M55M1 leads the edge AI platform using a 200 MHz Arm Cortex-M55 CPU and an Ethos-U55 NPU. The combination of these two processing units allows the parent microcontroller to effectively achieve inferencing without needing a high-power system. The M55M1 includes 1.5 MB SRAM and 2 MB flash, alongside memory expansion and peripheral support.

 

Nuvoton M55M1

The Nuvoton M55M1 offers advanced edge AI performance thanks to the Ethos NPU included in silicon. Image used courtesy of Nuvoton
 

In addition to the M55M1, Nuvoton has also revealed the MA35D1 microprocessor and M467 series, each targeting unique applications. The MA35D1, with its dual-core architecture, 800-MHz clock, and USB camera support, is a candidate for object detection. The M467, on the other hand, includes a 200 MHz Arm Cortex-M4F core with a built-in FPU and DSP instruction set, allowing it to target various other AI endpoint applications.

In addition to hardware, Nuvoton’s endpoint AI platform includes a dedicated ML software development stack, further simplifying the development process for edge AI applications.

 

Augmenting Power at the Edge

While each of the MCUs shown here addresses unique applications, they all share a common theme: bringing more computing power to embedded devices. The newest MCUs offer major performance and efficiency increases, both of which are critical to enabling next-generation AI/ML edge devices and the IoT.

1 Comment