SMARC MODULE - THE HIGH-PERFORMANCE LOW POWER STANDARD

Ideal solution for size, weight, power and
cost-optimized AI applications

SMARC module – credit-card-sized Computer-on-Modules

Perfect solution for size, weight, power and cost-optimized AI applications at the rugged edge

SMARC 2.1 wurde am 23. März 2020 veröffentlicht und ist ein Computer-On-Module-Formfaktor, der vom SGET-Konsortium gehostet wird.

congatec ist Editor der SMARC 2.0 und SMARC 2.1 Spezifikation. 

 

High-performance ecosystem for SWaP-C applications

SMARC has been developed to support low-power Arm and x86 processors alike. With its rugged and cost-optimized MXM 3 connector, which is backed by broad industry support due to its use with notebook graphics cards, this module standard enables exceptionally slim and rugged designs.  


Highest interface-per-watt ratio of all COM standards

SMARC takes an absolute top position among all Computer-on-Module standards thanks to its unique interface-per-watt-ratio. With a power consumption of typically 6 Watts and 314 signal pins, SMARC module executes up to as many as 4 camera inputs and 4 graphics outputs, making the standard an ideal fit for every embedded vision and situational awareness application. Additionally providing up to 4x Gbit Ethernet ports including support of hardware-based IEEE 1588 precision time protocol (PTP) and the ability to host wireless interfaces like WLAN and Bluetooth on the module, SMARC offers exceptional connectivity for any IoT application. And thanks to CAN bus support, SMARC is also well prepared for in-vehicle applications. 


Facts, features and benefits of SMARC module

FactsFeatures Benefits
Vendor independent standard                                                                                                                                            SMARC modules are available from many different vendors, all exchangeable within the same pin-out and footprint                         Increased availability and reliable multi-sourcing strategy reduce cost, improve production scalability, and boost resilience against supply chain issues 
Processor and SoC agnostic                                                                                                                                            SMARC modules can host low-power x86 as well as powerful Arm SoCs from different vendors                                                                SMARC modules can host low-power x86 as well as powerful Arm SoCs from different vendors 
Support of embedded vision and IoT centric interface technology SMARC modules support up to 4x camera inputs and 4x GbE with TSN as well as optional wireless communication   Future-oriented interface technology for an extra-long application lifecycle without bottlenecks in data transfer rates 
Reliable connector technology                                                                                                                                           SMARC modules feature the common MXM 3.0 connector designed for dedicated graphics cards in notebooks                     The 314 pin connector features a highly proven technology with a broad supporter base and cost-optimized market offerings for highest reliability and cost efficiency 

 


SMARC module positioning

SMARC 2.1 is perfectly positioned between the two well-established module standards Qseven and COM Express. Compared with the Qseven standard, SMARC offers more interfaces – especially embedded vision-oriented ports. Compared with the high-performance COM Express or even COM-HPC modules that make up the COM power class, SMARC 2.1 targets low-power applications that cannot be addressed by these performance-oriented standards with way higher power requirements. 


 SMARC 2.1 moduleQsevenCOM Express 3.1 Type 10 (Mini) COM-HPC Mini 
Footprint                 82x50 mm                                                         70x70 mm | 40x70 mm (µQseven)         84x55 mm                                                         95 x 70 mm                                                             
Construction height 1.5 mm (module bottom to carrier board top) 5 mm (module bottom to carrier board top) 5 or 10 mm (module bottom to carrier board top) 5 or 10mm (module bottom to Carrier Board top) 
Power25 W 12 W 68 W 76 W 
Signal pins 314 230220400
PCIe¹4x PCIe Gen 3 4x PCIe Gen 3 4x PCIe Gen 4 16 x PCIe Gen 5 
Graphics¹DP++/HDMI + 1x DP++ + 2x LVDS/eDP/MIPI DSI eDP/HDMI + 2x LVDS                              1x DDI +1 LVDS/eDP                                         3x DDi + 1x eDP                                              
Sound HDA + I2S 1x HDA/I2S 1x HDA/SoundWireHDA, Soundwire, & I2S
Camera in¹               2x MIPI CSI + 2x MIPI CSI (connectors on module) -2x MIPI CSI (connectors on module)                2x MIPI CSI (connectors on module)                     
Ethernet¹                  4x 1 GbE with TSN                                           1x 1 GbE with TSN                                 1x 1 GbE with TSN                                          2x 10 GbitE with TSN + 2x 10 GbitE (SERDES) with TSN 
WirelessAntenna connector for WiFI & Bluetooth Not supported Not supported Not supported 
USB¹                          2x USB 3.0 + 6x USB 2.0                                  2x USB 3.0 + 8x USB 2.0                           2x USB 3.2 + 8x USB 2.0                                  4x USB 4.0, 4x USB 3.2 x1 / USB 3.2 x1 + 8x USB 2.0 
SATA¹ 1x SATA Gen 3 2x SATA Gen 3 2x SATA Gen 3 2x SATA Gen 3 
CAN2x 1x 2x 1x 
UART 4x 1x-2x
GPIO14x 8x 8x 12x 
OthereSPI, SPI, I2C SPI, LPC, I2C, SDI LPC/eSPI eSPI, 2x SPI, SMB, 2x I2C 

¹ Not all I/Os are available in parallel. Some pins are shared.


The SMARC ecosystem

congatec offers a comprehensive SMARC ecosystem allowing agile and reliable realization of future-oriented vision-based AI applications. OEMs get direct access to modules with suitable cooling solutions, carrier boards, and ready-to-go embedded vision starter sets.  


Unsere SMARC Module

Mehr über conga-SMX95

conga-SMX95

  • NXP i.MX 95 processor series
  • Up to 6-core ARM Cortex-A55
  • NXP eIQ® Neutron NPU
  • Advanced Security
  • Industrial temperature support
Mehr über conga-SA8

conga-SA8

Virtualization Ready
  • Intel Atom x7000RE Series
  • Up to 8 Cores
  • Up to 16 GByte soldered LPDDR5
  • Virtualization ready with RTS Hypervisor
  • Industrial temperature support 
Mehr über conga-STDA4

conga-STDA4

  • TI processor TDA4VM or DRA829J
  • Dual Arm® Cortex®-A72
  • Deep learning hardware accelerator
  • Arm® Cortex®-R5F MCUs for real-time communication
  • Industrial temperature support
Mehr über conga-SMX8-Plus

conga-SMX8-Plus

  • NXP i.MX 8M Plus application processor
  • Integrated Neural Processing Unit (NPU)
  • Quadcore Arm® Cortex®-A53 / M7
  • Up to 3 independent display outputs
  • Industrial temperature support
Mehr über conga-SMX8-Mini

conga-SMX8-Mini

  • NXP i.MX 8M Mini application processor
  • Up to quadcore Arm® Cortex®-A35
  • 1x Arm® Cortex®-M4F for real-time processing
  • GC NanoUltra 3D graphics
  • Industrial temperature support
Mehr über conga-SMX8X

conga-SMX8X

  • NXP MX8-X application processor
  • Ultra-low power design (2-5 W)
  • Up to quadcore Arm® Cortex®-A35
  • 1x Arm® Cortex®-M4F for real-time processing
  • Industrial temperature support

conga-SMC1 3.5-inch carrier board

The size-optimized SMARC 2.1 carrier board in 3.5-inch form factor is application-ready and can be deployed off-the-shelf in small to mid-sized series in combination with any congatec SMARC Computer-on-Module. Tailored to make 3.5-inch SBC designs modular, it is available in optimized x86 and Arm variants. Its slot for SMARC 2.1 processor modules provides processor socket independent scalability, making OEM solutions highly flexible and long term available. The carrier board is also prepared for rapid customizations for best time to market. 

conga-SMC1 x86     conga-SMC1 ARM


Embedded vision starter set

At the heart of the starter set is the SMARC 2.1 Computer-on-Module conga-SMX8-Plus. The 3.5-inch carrier board conga-SMC1/SMARC-ARM directly connects the 13 MP Basler dart daA4200-30mci BCON for MIPI camera without any additional converter modules. Next to MIPI CSI-2.0, USB and GigE vision cameras are also supported. On the software side the set comes with a bootable SD card with preconfigured boot loader, Yocto OS image, matching BSPs, and processor-optimized Basler embedded vision software enabling immediate AI inference training on the basis of captured images and video sequences. 

Request your starterkit


Design-in training Arm

Our Arm design training program provides a sound entry into the complex Arm world by introducing the Arm architecture, special bootloader requirements, and different Operating Systems including Yocto Project and Android. Hands-on training for setting up the hardware, booting a sample image, and BSP setup as well as a presentation of our Technical Service Center (TSC) and its services round off this 3-hour online training program.  

Register

Carrier board design-in training

Our carrier board design training program shares best practice knowledge on how to design-in SMARC modules. The courses guide engineers through all the mandatory and recommended design essentials and best practice carrier board schematics. OEMs participating in the program will be able to efficiently simplify their carrier board design projects. 
 

Register

Erfahre mehr über SMARC Computer-on-Modules

Downloade unsere Whitepaper!

NPU trainierte Augen

Die Embedded-Vision-Integration entscheidet über den Erfolg von KI-basierten Systemen – egal, ob bei autonom fahrenden Vehikeln, der Videoüberwachung oder bei kollaborativer Robotik. Damit Entwickler schneller zu ihren individuellen Lösungen kommen, bieten sich vorkonfigurierte Embedded-Vision-Baukästen an.

Whitepaper herunterladen

SMARC modules with NXP i.MX 8M Plus Processor

congatec is expanding its SMARC platform with a new module with an NXP i.MX 8M Plus processor especially for embedded AI applications. Thanks to the extensive ecosystem with application-ready 3.5-inch carrier board, Basler cameras, and AI software stack, fast proof of concept is possible.

Download Whitepaper

 

Kontaktiere uns