congatec - The Rhythm of Embedded Computing

concept & advantages

Computer-on-Modules are small circuit boards that are plugged into a standard connector on an application-specific carrier board and contain the processor with all of its vital peripherals.
Most interface circuitry is situated on the module, including PCI Express, Ethernet,  SATA, display interfaces, general purpose I/O, and similar interfaces. The interfaces are routed to the appropriate ports on the carrier board.  The application-specific circuit parts are accommodated on the carrier board.  When designing your carrier board, you decide which of the electrically available computer  properties you actually need and which extensions you want to develop.

 

Concept

• CPU module with standard PC core functions
• Carrier board with customer specific function&size
• Logical alternative to a chip-down design effort

 

Benefits

• Faster time to market
• Reduced development costs
• Scalable product range
• Allows customer focus on system features
• Faster reaction to market trends
• Second source philosophy
• Minimize inventory cost

Lower costs 

COMs save money. The cost of the development and end product are dramatically reduced when compared with a full custom design.
This holds true for the product‘s entire life-cycle.
COMs provide cost advantages from the start:

• Lower engineering cost
• Lower product cost
• Lower cost of life cycle management

Improved Flexibility 

COMs are flexible and can meet all performance requirements. The modules support a wide range of performance levels starting from NXP i.MX6 up to the Intel® Xeon® processor, as well as future architectures. The COM standards are well established and are already prepared for the future.

• Scalability
• Easy performance and technology upgrades

Reduced risk 

COMs minimize risk. Basic changes during the design phase, or in the middle of a product‘s life cycle, are easily managed. Simply plug in the next-generation COM module and continue.
COMs allow for easy upgrades

• Lower design risk
• Lower transition risk

Time-To-Market Advantage 

COMs put you in a leading position. The use of customized carrier boards reduces necessary engineering effort by separating your design work from the embedded PC technology.
Focus on your own core competency.

• Faster time to market
• Faster engineering
• Faster reaction time to market changes

SMARC modules stand out with their rich choice of graphics, camera, sound, network and optional wireless interfaces. They offer embedded system developers a complete, off-the-shelf, credit-card sized embedded computing core that is ideal for IoT connected multimedia platforms and many other graphics-intensive low-power applications. 
The updated specification SMARC 2.1 extends this feature set even further. Two PCI Express lanes can be used alternatively as additional Ethernet ports. Extra two MIPI CSI interfaces and additional GPIOs also extends the possibilities of SMARC modules. 

The 314 pins of the SMARC 2.1 connector, which is also used for the MXM 3.0 graphics card standard, provides space for up to four video outputs, underlining SMARC 2.1’s particular suitability for multimedia applications.

 

Quick overview 

SMARC 2.1 is perfectly positioned between the two well-established module standards Qseven and COM Express. Compared with the Qseven standard, which allows low-cost entry into the world of computer modules and integrates various x86 and ARM low-power processors for the process and field levels, SMARC offers more interfaces – especially more multimedia ports. Compared with the high-performance COM Express or even COM-HPC modules that make up the upper class of COMs, SMARC 2.1 is targeted at the low-power processor segment and also supports other interfaces than COM Express.

 

Connector

SMARC 2.1 utilizes a highly reliable, high speed certified but affordable 314 pin 0.5mm MXM 3 connector.

 

Extensive video interface options

SMARC 2.1 offers a rich choice of internal and external video interfaces. Two dual-mode DisplayPorts (DP++) are provided for flexible external screen connections via DisplayPort, HDMI or VGA. For internal displays 2x24 Bit LVDS is implemented. Alternative use is defined to support two independent embedded DisplayPort (eDP) or MIPI Display Serial Interface (DSI)

 

Up to 4 Ethernet interfaces yield greater precision 

SMARC 2.1 implements two Gigabit Ethernet ports including SDPs (Software Defined Pins) to allow for hardware-based IEEE 1588 Precision Time Protocol (PTP). Two of the 4 PCI Express lanes (PCIe C & PCIe D) can be used alternatively as SERDES interfaces to provide additional Ethernet functionality. In total up to 4x Gigabit Ethernet ports are possible now.

 

Wireless

SMARC 2.1 provides a special area on the module that is dedicated to the placement of the miniature RF connectors to allow for wireless interfaces like WLAN and Bluetooth.

 

Camera interfaces

SMARC 2.1 provides all signals required to support digital cameras. For this purpose, two serial MIPI CSI (Camera Serial Interface) have been implemented on the module connector. Further two MIPI CSI interfaces can be implemented as flat foil connectors on the SMARC 2.1 module.

 

Low Power

SMARC 2.1 is defined for low power consumption applications only. It can be operated by 3.3V or 5V DC power and provides all additional signals for battery management.

 

Backwards Compatible

Since all new SMARC 2.1 features have been implemented as alternative functions to existing signals and/or on unused pins, SMARC 2.0 is 100% backward compatible. A redesign of existing SMARC 2.0 carrier boards is only necessary if new SMARC 2.1 functionalities are used.

 

Small Size

The module dimensions are a mere 82 x 50 mm². This means it can be easily integrated into size constricted systems. 

Harness the power of today’s multicore processors

The innovative Real-Time Systems Hypervisor permits multiple operating systems - both real-time (RTOS) and general purpose operating systems (GPOS) like Microsoft™ Windows® or Linux - to run concurrently on multicore x86 processors. By utilizing this powerful and cost-effective software solution, designers achieve increased flexibility in system design and remarkable enhancements to functionality and performance - at the same time reducing overall system cost.

 

Hard Real-Time Performance: Multiple Operating Systems in Perfect Harmony

• Combine real-time operating systems like VxWorks®, QNX Neutrino or Real-Time Linux , with e.g. Microsoft™ Windows®
• Operating systems reside simultaneously on an x86 computer while maintaining the hard real-time characteristics of an RTOS
• Reboot any operating system anytime without disturbing the execution of other operating systems
• Communication via high performance virtual TCP/IP network and flexible shared memory
• User-definable boot sequence

Advantages 

• Reduced system costs and physical size
• Hardware consolidation
• Hard real-time performance
• Maximum flexibility in system functionality
• Increased reliability (MTBF) as no additional hardware is required for additional operating system
• Works seamlessly with COTS and proprietary operating systems
• Proven in thousands of systems worldwide

About the Hypervisor 

• All operating systems operate completely independent
• User defined startup sequence of operating systems
• Any operating system can reboot without affecting other operating systems
• All operating systems safely separated and protected
• Standard development tools can be used (supplied by the operating system vendors)
• Standard drivers can be used - no special development required
• NUMA (Non-Uniform Memory Access) fully supported 
• OS independent drive sharing

The specifications for COM-HPC, COM Express, Qseven and SMARC modules include heatspreader definitions, the mechanical thermal interface. All the heat generated by power-consuming components such as chipsets and processors is transferred to the system’s cooling via the heatspreader. This can be achieved by either a thermal connection to the casing, a heat pipe or a heat sink.

High Performance Cooling

The congatec heatspreaders and cooling solutions for the high performance modules are feature heatpipes in order to boost performance and reliability. A copper block is mounted on the chip to absorb heat and to mitigate the effects of thermal peaks. Between the chip and the copper block, a phase-change material is placed to improve the heat transmission. To account for different component heights and manufacturing tolerances, the copper block is spring loaded to apply an optimized pressure to the silicon dye. The copper block and the cooling fins or heat plate are connected by flexible flat heatpipes.

The heat pipe is attached directly to the cooling blocks on the chip and the heatspreader plate. As a result, more heat is transported from the processor environment to the heatspreader, hot spots are cooled more quickly and therefore the processor is optimally cooled.
The heatpipe adapter uses the same principals as described above but transmits the heat from the module directly to standard heat pipes with 8mm diameter. This approach allows for cost optimized, ultra-flat system solutions i.e. 1 U rack units.
 

Heat spreader and passive cooling solution for Pico-ITX boards

The CPU as heat generating component is placed on the bottom side of the Pico-ITX board. This allows for a heat spreader concept for conduction cooled systems. The heat spreader with its installed phase change material and copper block for heat transient buffering is preinstalled with 2 screws to the Pico-ITX board. This combination can be mounted to a metal housing or to any other system cooling device.
 
Extreme slim passive cooling for conduction cooling. Installed phase change material for best heat transmission. Solid copper block to handle transient heat and allows for best burst performance. Through holes for easy mounting.
 

Cooling solutions for SMARC modules