As the demand for smart homes, smart cities, smart buildings, and Industry 4.0 grows, IoT solution developers are increasingly looking for reliable hardware to run these applications with zero downtime. One of the most sought-after pieces of hardware is the NXP i.MX8M Plus SoM.
Based on the NXP i.MX8M Plus processor, the NXP i.MX8M Plus SoM is a single board-level module that is purpose-built for machine learning, vision, industrial automation, and advanced multimedia applications with high reliability.
Let’s analyze the capabilities of this module in detail to see how it can help you develop reliable IoT solutions for your customers.
NXP i.MX8M Plus SoM Features
This System-on-Module borrows most of its features from the onboard chip, although it has some of its own. Combined, the board boasts of the following.
- Compact form factor (62mm x 36mm)
- 8-layer board-to-board connector (Pb-free with an independent ground signal layer and immersion gold finish)
- Multiple GPIO interfaces
- 320 output pins (includes nearly all CPU pins)
- Built-in NPU in the SoC providing 2.3 TOPS
- NPU, ISP, VPU, HiFi 4, and CAN-FD
- Dual Gigabit Ethernet ports (one with Time Sensitive Networking (TSN) for low-latency applications)
- Up to 2GB RAM and 16GB eMMC
Compatible CPUs for the NXP i.MX8M Plus SoM
SoMs are highly scalable and modular, and the NXP i.MX8M Plus SoM is compatible with eight CPUs from the i.MX 8M Plus family, which includes:
- MIMX8ML8CVNKZAB (1.6 GHz/800 MHz quad core with NPU, ISP, VPU, HiFi 4, and CAN-FD)
- MIMX8ML6CVNKZAB (1.6 GHz/800 MHz quad core with ISP, VPU, and CAN-FD)
- MIMX8ML4CVNKZAB (1.6 GHz/800 MHz quad core with CAN-FD)
- MIMX8ML3CVNKZAB (1.6 GHz/800 MHz dual core with NPU, ISP, VPU, HiFi 4, and CAN-FD)
- MIMX8ML8DVNLZAB (1.8 GHz/800 MHz quad core with NPU, ISP, VPU, HiFi 4, and CAN)
- MIMX8ML6DVNLZAB (1.8 GHz/800 MHz quad core with ISP, VPU, and CAN)
- MIMX8ML4DVNLZAB (1.8 GHz/800 MHz quad core with CAN)
- MIMX8ML3DVNLZAB (1.6 GHz/800 MHz dual core with NPU, ISP, VPU, HiFi 4, and CAN)
We used the first option in this module because it has the best features. The 1.8 GHz quad-core CPU would have been better if it had CAN-FD, but it only provides the CAN protocol.
CAN-FD (Controller Area Network with Flexible Data Rate) is an extension of the original CAN bus protocol that increases data rates and supports bigger message frame sizes.
This makes the NXP SoM better for automotive, aviation, defense, and industrial applications because it increases network bandwidth to transmit more data when communicating with sensors and controllers in real-time.
NXP i.MX8M Plus SoM Performance/Capabilities
Machine Learning
The primary advantage of this SoM is its machine-learning capabilities. It has a machine learning accelerator (2.3 TOPS NPU) that provides real-time analysis on the edge.
Smart cameras, for instance, can use the locally available trained model to identify the types of vehicles passing on specific roads and send only the necessary data to the cloud for deeper analysis. This reduces network latency and costs while keeping sensitive data near the source.
The NXP i.MX8M Plus SoC supports AWS, so it provides a complete solution for developing, deploying, monitoring, and updating large-scale deployments of edge ML gateways.
Combining the 2.3 TOPS NPU, powerful quad-core A53 processors, and a Neon coprocessor with AWS ML ops services like SageMaker Edge Manager, you can train, optimize, and package an ML model, then deploy it to several cameras in the field.
Vision
This SoM features two 4-lane MIPI camera serial interfaces that can transmit data at speeds of up to 1.5 Gbps. These can bring in data from cameras on robots, vehicles, and other autonomous systems, providing two angles of the area or object ahead for processing and analysis to determine the next move.
The SoC in the module features two image signal processors with a resolution of up to 12MP and an input rate of 375 megapixels per second. These components process the raw images from each camera and deliver them to the NPU for identification.
With a trained data set, the NPU can recognize things like road signs, good products from bad ones, human gestures from machine operators, etc., to enable autonomous movement or operations.
Real-Time Control
The NXP i.MX8M Plus processor has a quad-core Cortex A53 paired with a Cortex M7 in a big.LITTLE configuration.
As the “LITTLE” one in this arrangement, the M7 core runs efficiently at a clock speed of 800 MHz and can provide functions like real-time control to actuate machines. For instance, it can turn the eyes (cameras) in robots or CCTVs when required, or even control machine operations in Industry 4.0.
This real-time control function is supported by dual CAN-FD, which increases network bandwidth, while the Ethernet Gigabit port with TSN reduces network latency.
General-Purpose Computing
The “big” A53 core is suitable for handling general-purpose computing in the heterogeneous processor and can run functions like database management to store captured data locally.
Multimedia Capabilities
For multimedia applications, the NXP i.MX8M Plus SoM provides video encoding and decoding. It uses 1080p60 AVC/H.264 and 1080p60 HEVC/H.265 encoders for the former and 1080p60 HEVC/H.265 Main, Main 10 (up to level 5.1), 1080p60 VP9 Profile 0, 2, 1080p60 VP8, and 1080p60 AVC/H.264 Baseline, Main, High decoders for the latter, forming its vision processing unit.
The SoC’s graphics department is also well-equipped, featuring a GC7000UL GPU for 3D graphics (supports Vulkan and OpenCL) and a GC520L GPU for 2D acceleration. There is a super tile status buffer between these two GPUs to enable render target compatibility.
Audio is handled using 18x I2S TDM, DSD512, S/PDIF Tx + Rx, eARC (HDMI), and ASRC, with input coming in using an 8-channel PDM mic input.
An 800 MHz Cadence Tensilica HiFi 4 DSP integrated into the chip as a secondary core processes the audio output, providing noise cancellation and suppression. This low-power voice accelerator also has a neural-network-based voice assistant that can provide voice recognition in smart speakers, soundbars, TVs, etc.
All these functions, including the other capabilities, are supported by a Neon coprocessor, which accelerates signal processing algorithms to speed up voice recognition, deep learning, video processing, and computer vision.
High Industrial Reliability
Smart devices installed in factories and industrial settings are exposed to harsh conditions, so they must have features that ensure reliability and longevity.
The NXP i.MX8M Plus SoM incorporates these features, using industrial-qualified DDR inline ECC that can run nonstop for 10 years in extreme temp ranges (-40°C to 105°C). So you can use this module in low or high-temperature applications, such as inside welding robots in car manufacturing plants.
This longevity is critical for industries that want to develop new, complicated manufacturing processes or have to undergo safety procedures before approval. These can take time, and the NXP SoM will provide reliable performance all through.
On top of that, the physical NXP chip is made using the 14nm FinFET process that has a low SER (Soft Error Rate). This ensures maximum performance, reliability, and energy efficiency.
Applications of the NXP i.MX8M Plus SoM
Automotive
Since this NXP SoM provides real-time control, it can run the motors that push heated or cooled air into the cabin in a vehicle’s HVAC system. It can also control the actuators that open and close vents to regulate the climate inside the cabin.
The SoM’s multimedia capabilities can also come into play here, enabling high-quality audio output from the infotainment system, as well as voice control to keep the driver’s hands on the wheel and eyes on the road when manually adjusting the HVAC and other vehicle functions.
Industry/Factory Automation and Control
Combining machine learning, vision, control, and industrial reliability provides intelligent and dependable management of various industry processes, such as AC, building safety, motor drivers, heat metering, factory automation, energy generation, etc.
The SoM can be embedded in various devices, such as robots, industrial computers, Human Machine Interfaces, and IoT gateways in these industrial applications.
Smart Cities
The machine learning and vision capabilities of this SoM make it suitable for automatic vehicle identification, fleet management, service drones, traffic monitoring and flow optimization, fleet analytics, driver monitoring, and surveillance in smart cities.
Smart Homes
Smart homes incorporate various components and the NXP i.MX8M Plus SoM is handy for tasks like home automation, smart robots, remote control using voice assistants, surveillance (using smart cameras), etc., using its multimedia, machine learning, vision, and real-time control capabilities.
Wearables
Smart headphones (hearables) can use this SOM to provide superior audio quality with minimal battery power consumption, while smartwatches can use its machine learning capabilities to make edge health monitoring smarter and multimedia capabilities to enable voice control.
Conclusion
The versatility of the NXP i.MX8M Plus SoM is its biggest strength. Machine learning and vision capabilities make it relevant for various applications requiring camera inputs, such as robotics, surveillance cameras, drones, and autonomous vehicles.
With the additional real-time control, multimedia, and high industrial capabilities, the use cases expand drastically, especially in Industry 4.0.
The SoM can act as a gateway to monitor and control various processes from a single point. It can even be embedded in robots to enable quick, low-latency analytics and decision-making at the source, such as when using them in warehouse management.
Check out the NXP i.MX8M Plus SoM to get more technical details about it, and contact us if you find it suitable for your projects. We can fine-tune it to meet your project requirements at a reasonable cost, including fitting it with the most suitable SoC from the compatible options.
