Rhinos for Wrynose

PHYTEC has been busy with events, hands-on security trainings, software updates, new SOMs in development and much much more. One of our most exciting updates this quarter is our move to Yocto Wrynose for the entire Texas Instruments AM6 series of PHYTEC SOMs. Here's a clever mnemonic to help you remember how to pronounce the new release: Am I on the nose with my wry humor, or would you rather code with rhinos? Either way, you're going to have to keep saying it until 2030, since this LTS has long-term support. 

Read on for more news about what PHYTEC has been working on in Q3.  

Yocto Wrynose, Linux 6.18 : BSP updates for phyCORE-AM6 series SOMs + phyFLEX-i.MX 95

AI White Paper: AI for embedded vision systems 

SOMs in Development:  Qualcomm IQ239, Synaptics SL2610, NXP i.MX937/952

Events: NXP Tech Days: Santa Clara and Boston, Designing with NXP, embedded world NA

Engineering Insights: Cool things you can layer on top of PHYTEC's standard BSPs 

phyCORE-AM6 Series Software Updates

phyFLEX-i.MX Software Updates

Whitepaper: Designing Edge AI Success

Aligning hardware, software and AI models for production-ready embedded AI systems

With this whitepaper, PHYTEC and NXP want to provide you with valuable insights for your next AI project.
It builds on 40 years of PHYTEC experience in the design, development and production of embedded
hardware and on the close collaboration between PHYTEC and NXP to simplify access to new technologies.

• Learn how to translate an Embedded AI idea into clear system requirements by distinguishing between
  Basic AI, Multimedia AI and Vision AI.
• Understand why the best hardware choice is not defined by processor cost alone, but by the fit-for- purpose between workload, performance, software support and long-term maintainability.
• See how AI model(s), runtime target, BSP, development host and software stack must work together to reduce integration risk and move faster toward a working target system.
• Read about a proven Vision AI example that demonstrates facial recognition in the real world through the right combination of hardware, AI model and software stack.

>>Download the whitepaper

SOMs in Development 

FPSC: Swap the SOM, Keep Your Carrier Board

The following FPSC SOMs are announced and currently being developed at PHYTEC. Please subscribe to this newsletter and stay up-to-date on product updates, launches and early access hardware or get in touch with our sales team about your next project.

Qualcomm® IQ2390

phyFLEX-IQ2390

This system stands apart by bringing Qualcomm’s high-efficiency silicon into true industrial territory for the first time. By integrating local AI, ISP vision pipelines, and a real-time RISC-V MCU on a single low-power chip, it eliminates the need for extra baseboard microcontrollers or discrete accelerators. The result is a rugged, power-efficient foundation for smart cameras, HMIs, robotics, and edge controllers that require local intelligence without cloud dependency.

• Arm^® Cortex®-A78, 3x Arm Cortex®-A55
• Qualcomm Hexagon™ V66
• GPU, SiFive E61 RISC-V MCU at 600 MHz (1.1 TOPS)
• Up to 16GB LPDDR4, Up to 128GB eMMC
• USB 3.1, PCIe 2.0, 1Gb Ethernet, 1x RGMII, I3C, 9 UART
• Advanced Multimedia

Synaptics® SL2610

phyFLEX-SL2610

After decades of experience in high-performance ARM-based consumer solutions, Synaptics is now bringing its proven technology to the industrial embedded market to address the specific requirements of modern IIoT systems.

The phyFLEX-SL261x SOM brings high security for industrial applications to your device: featuring a dual-core Arm Cortex-A55 processor (up to 2 GHz) and an Arm Cortex-M52 (200 MHz), the module combines computing power with real-time capability. The Torq NPU delivers up to 1 TOPS of AI acceleration and integrates a Coral RISC-V transformer. It's ideal for smart edge applications, computer vision, and predictive maintenance and also boasts comprehensive interfaces including MIPI-DSI, MIPI-CSI, 2× USB 2.0, 4× I²C + I3C, and 8× UART. With PSA-certified protection, the SOM is perfectly equipped for demanding industrial environments.

• Dual-core Arm Cortex^®-A55 with up to 2GHz
• Arm Cortex^®-M52 (200 MHz)
• DDR3 and DDR4 Design
• Torq NPU (1 TOPS) incl. Coral RISC-V transformer
• Vision Interfaces: MIPI-DSI, MIPI-CSI (2 lane)
• General interfaces: 2x USB 2.0, 4x I2C + I3C, 8x UART
• Security Features: Secure Boot, JTAG, PSA certified

Find out more about PHYTEC's new Synaptics SOM at the Synaptics Tech Day October 20th

NXP i.MX 937/952

phyFLEX-i.MX 937/952

NXP's pin-compatible processors for scalability and optimization: the i.MX 937 provides a cost-optimized foundation for graphical user interfaces and multimedia systems, and i.MX 952 expands the range of applications with more powerful graphics, video, and image processing capabilities, such as an integrated ISP and 4K video processing.

• Robust FTGA soldering module, compatible with i.MX 937 or i.MX 952
• Dual- or Quad-Core Arm^® Cortex^®-A55 with up to 1,7 GHz
• Energy-efficient real-time with Arm^® Cortex^®-M7 (800 MHz) and -M33 (333 MHz)
• LPDDR5 memory
• Integrated ISP for camera image processing (i.MX 952)

Events

NXP Tech Days and Designing with NXP, Seattle

We had a great time at three events hosted by NXP in Q3, NXP Tech Day Santa Clara and Boston, and a new event, Designing With NXP, which was hosted locally in Seattle. This was an intimate Partner and Distribution focused event where customers could come and evaluate SOM vendors and distributors. PHYTEC lead a hands-on training giving participants the chance to try out security provisioning on the phyCORE-i.MX 8M Plus. You can do this on your own or schedule a training with us using the PHYTEC SecuriPHY Security Bundle. You can also learn more about PHYTEC Security practices and integration with NXP SoCs in our Secure by Design whitepaper.

Embedded World North America

At Embedded World North America PHYTEC showcased our newest ST SOM the phyFLEX-STM32MP25x at the STMicroelectronics booth.

If you want more information about PHYTEC's ST SOMs check out the new ST Blog: PHYTEC SOMs, Production Boards, and 3 Tips to Get Started with an STM32 Microprocessor

Upcoming Events

Open Source Summit, Oct 7-9 | Register

PHYTEC will be showcasing our HandHeldOne Demo at the Texas Instruments booth. HandHeldOne is a handheld gaming device, built in 4 weeks, that highlights the speed and ease of development with PHYTEC FPSC SOMs.
>>Learn more about this demo

Synaptics Tech Day, Oct 20 | Register

PHYTEC will be presenting our newest FPSC SOM the phyFLEX-SL2610 and talking about the scalability and future proofing of our FPSC technology.

Electronica, Nov 10-12 | Register

Come visit at the PHYTEC booth: B5.229 and see our latest SOMs and Demos. We will also have demos at the NXP booth featuring phyFLEX-i.MX 95 and the TI booth featuring phyFLEX-AM62P. 

Engineering Insights:

Engineering Insights are short technical blogs to help with unique or complex development challenges. Below are a couple of sneak-peaks at a some of the insights we will be publishing on our blog in the coming weeks. 

 A/B scheme for bootloaders

What happens when your boot path fails on a remote system?

It's one of the trickiest problems in embedded systems: how do you safely update the bootloader on a device you can't physically reach?

Redundancy should be built into your system by design for when Plan A fails, whether it's an interrupted update, corrupted bootloader, or another unexpected failure. The challenge is making sure there's a reliable path back to a known good system.

In this Engineering Insight, we will look at how the phyCORE-AM64x uses redundant eMMC boot partitions and SPI NOR to provide an autonomous recovery path. If Plan A fails, the system has a defined way to fall back to Plan B without requiring physical access or external tools.

Functional C7 DSP for Edge AI

The AM62A has dedicated compute resources for edge AI, but putting them to work takes more than just running a model.

PHYTEC recently got AI models running on the C7x DSP with Edge AI on the AM62A. That means AI processing can take advantage of the SoC's dedicated compute resources while leaving the ARM cores available for other Linux tasks.

The ARM Cortex-A cores on the AM62A can run AI models too, but at a significant power and performance cost compared to dedicated silicon. The C7 DSP with MMA (Matrix Multiply Accelerator) can run inference dramatically faster and at lower power, which is the whole point of choosing a chip like the AM62A for edge AI in the first place. Now you can get the AI performance TI designed the chip to deliver, not a software workaround on the wrong core.

Our upcoming Engineering Insight will walk through what it takes to get there, from compiling the model and loading the C7 firmware to setting up the data pipeline and connecting everything back to Linux.

Partner Highlight

acontis technologies

Getting EtherCAT up and running can mean spending time on kernel configuration, hardware access, and performance tuning before you even get to your application.

This free Yocto meta layer from PHYTEC takes care of that groundwork for the phyCORE-AM68Ax and phyCORE-i.MX8M Plus Development Kits. It configures the MCU_Ethernet interface with the optimized atemsys kernel module, giving acontis EC-Master direct hardware access and a ready-to-use path to high-performance EtherCAT Master communication.

Less time working through low-level setup. More time developing your EtherCAT application.

>>Get the accessory

>>Check out the code

Following along on our phyCORE-RT1170 Developer Contest? Submissions are in, check out the latest projects and stay tuned for our announcement of the winners on Oct 15th. Which project would you vote for? Follow us on Social Media, don't miss the results.

Subscribe to this newsletter and never miss out on PHYTEC News and Events.