Understanding the Test Criteria of Optical Fiber Transceivers Used in Space

By Jocelyn “Justin” Lauzon, Senior Technical Advisor at Reflex Photonics Inc.

Reliability is Critical for Components Operating in Space

This article was originally published in Military Embedded Systems in June of 2019.

In space and avionics applications, failure is not an option. Components must stand extreme heat, cold, radiation, shock, and vibration, yet deliver reliable performance. Devices must be tested beyond what is specified to ensure performance in harsh environments to avoid failure.
What are the criteria of a good reliability test program and how should one be designed? In this article, we examine the reliability requirements of such devices and the steps involved in designing a test program to ensure reliability and performance.

Diagram of the possible connections between the causes of failure

Diagram of the possible connections between the causes of failure. Connections are indicated with solid lines, no connection with a dashed line. Preordained failures have an internal mechanism that produces the same outcome, failing at a set rate, from the moment of creation. (Source: AT&T Tech Journal, 1985)

Electronic devices in space are bombarded by various types of radiation including visible, infrared, ultraviolet, x-rays, gamma rays abd others.

Electronic devices in space are bombarded by various types of radiation including visible, infrared, ultraviolet, x-rays, gamma rays and others.

The Advantages of Optical Fiber Extends to the Accuracy Demands of Space

In space, high performance components must be able to deliver reliably in the punishing environment. It is optical transceivers that drive transmissions, converting signals to and from a copper-resident format. Fiber optics communications provide high bandwidth and low latency signaling. Signal transmissions through fiber optic cables (FOCs) provide immunity to EM/RFI interference, crosstalk, and voltage level surges. Fiber optics’ accuracy and reliability exceeds traditional cabling. Covering 1,000 feet requires four pounds of FOC versus 39 pounds of copper wiring, and fiber optics also consume less energy than copper. To convert electrical signals from circuitries with copper output to fiber optics, optical fiber transceivers are usually required.

Reflex Photonics’ optical fiber transceivers support multiple electrical interfaces.

BGA electrical interface

MegArray™ electrical interface

LGA electrical interface

Fiber optics adds tremendous value to the entire system, therefore a suitable example within the scope of this article is a short-reach parallel multimode fiber optics transceiver supporting a large bandwidth (up to 28Gbps/channel). Intra-satellite communications require a high bit-rate. In prior years, optical transceivers represented high-performing technology, yet exhibited weakness in harsh environments. However, ruggedized, sealed optical transceivers now survive rocketing into orbit, extreme temperatures, and radiation. Testing optical devices primarily ensures quality and reliability. How does one test optical transceivers for reliable operation in harsh environments?

What Tests Are Needed to Ensure Critical Devices Can Endure the Harsh Environment of Space?

Several tests are necessary to ensure critical devices such as optical transceivers can provide reliable operation in space, including space applications testing (including radiation), mechanical, environmental, Life Tests, Live Tests, and rigorous screening tests for ensuring the reliability of subsequent lots.

How to Perform the Five Critical Tests?

In addition to a controlled set up, each test requires visual inspection for anomalies before and after the evaluation. A significant performance degradation associated with a test is considered a qualification failure. The focus of the discussion relates to mandatory testing of fiber optic transceivers (FOTs) intended for operation in space or a similarly harsh environment. All devices need mechanical and environmental testing to verify long-term integrity. Mandatory Tests also include Life and Live Tests. Such tests can help root out design or process flaws. Finally, even if units pass all tests, every subsequent lot of units must undergo screening tests.

Mandatory Mechanical and Environmental Tests

Three consecutive mechanical tests are executed on the same FOTs. Before and after completion of the mechanical integrity evaluation, units must be visually inspected and characterized over the operating temperature range to confirm no significant performance degradation occurs. Mechanical integrity evaluation tests are executed on non-operating units, in the order below:

  • Vibration tests across all three axes (20g between 20Hz and 2000Hz at sixteen minutes per axis)
  • Mechanical shock tests of five repetitions each over all six orientations, using 500g shock on a half-sine pulse duration of 0.5ms
  • Thermal shock tests consisting of twenty cycles between 0 and 100°C with ten minutes dwell time and transient time less than five seconds

Three environmental stress tests for FOTs include:

  • A temperature cycling test to evaluate mechanical fatigue over the component’s lifetime. For example, a mismatch of dissimilar materials subjected to thermal expansion can cause failure.
  • A damp heat test to ensure that a sealed FOT can continually provide resistance to a moisture-filled atmosphere.
  • A sequence of tests, applicable only to products with a BGA electrical interface compatible with solder reflow, is performed. These test for possible impact to reflow profile and cold temperature storage.

Life Tests

Life Tests validate long-term reliability via extensive, hastened lifetime tests with several units from different lots. Life tests are not intended to test for infant mortality, for which burn-in tests would suit. Life tests forecast performance degradation over a lifetime. To emulate operating for more than 20 years, an FOT operates at a bias current exceeding operating condition by 80 percent for 4000 hours at a case temperature of 100°C.

Live Tests

Live Tests, although more complex, thoroughly evaluate the performance of several electrical and optical interface configurations.
Live testing is performed with high-speed digital signals operating through every channel of the DUT (Device Under Test). As the DUT is stressed, transmission errors are measured to maintain a BER (bit error rate) that is better than 1 error in one trillion bits while operating under harsh conditions. Tested under stress, a significant signal degradation of an FOT will demonstrate a cumulative effect from both transmitter and receiver areas that would impact the error count.

Live test setup configuration.

Live test setup configuration.

Space Applications Tests

A major threat in space is radiation. Radiation testing must be split into three different categories for evaluation, covering potential complications related to geostationary or low-earth orbit environments. At least five units should be collected for each test. The tests should be repeated when FOT critical components arrive from new fabrication lots. The radiation tests are:

 

  • A non-operational test with a Total Non-Ionizing Dose (TNID) protons (5e12proton/cm2total dose).
  • A live test, Single Event Effects (SEE), at both room temperature and 85°C, including heavy ions (Ho, Cu, Ar, Ne, N, each for a total fluence of 1x107ions/cm2).
  • A biased and un-biased test with Total Ionizing Dose (TID) of gamma rays (100 krad cumulative dose).

The number of errors/events caused by live radiation tests must be compared to those occurring at acceptable levels. For non-operational tests, the performance of FOTs before and after the application of radiation dosages are compared and can only differ negligibly to pass. A ruggedized FOT that is sealed with encapsulation material must undergo an outgassing test. Live thermal vacuum tests, maintaining a vacuum of at least 5x10-5hPa for at least twenty thermal cycles extending from -40°C to 85°C, ramp temperature at ~5°C/minute, with 5-minute dwell times must also be successfully completed. Other tests for space applications include decompression tests, which can also be used to qualify parts for avionics applications. Decompression tests demonstrate that performance is not affected at pressure levels emulating a 2,438 m or 8,000 ft altitude, ramping to 15,850 m or 52,000 ft (in less than 15 s) and remaining there for one hour.

A highly ruggedized, radiation resistant optical transceiver suitable for use in satellites. The SpaceABLE28™ model has high I/O density with up to 28Gbps/lane from -40°C to 85°C.

Screening Tests

Finally, Screening Tests of six thermal cycles ramping at 8°C/minute with five-minute dwell times are completed for continuous assurance of quality COTS devices. Added to thermal cycling is a burn-in with a duration of 168 hours and a case temperature 100°C, operating at normal bias current. After a new FOT design is qualified via the aforementioned testing, units from subsequent production lots are validated with the screening process. Temperature cycling tests tend to reveal assembly defects, whereas burn-in tests exhibit infant mortality rates.

Conclusion

As discussed above, in applications such as space, avionics, and defense, failure is not acceptable. Components and systems must stand extreme heat, cold, radiation, shock, and vibration to deliver reliable performance. Finding failures, defects and marginal components through stringent testing is essential, and leads to products with a long life and high reliability in the field and in space. The five critical tests include Mandatory Mechanical and Environmental Tests, Life Tests, Live Tests, Space Applications Tests, and finally, Screening Tests. Advancing through a significant number of tests designed to demonstrate reliability in harsh environments results in fiber optic transceivers that meet or exceed operational requirements, and typically outweighs cost saving considerations for many critical applications.

VITA Technologies Round Table

This article was originally published in EE Catalog in January of 2018 and written by Anne Fisher, Managing Editor.

Sky, sea, rail—VITA technologies call them all home, and indeed, our panelists would argue, also deserve a home anywhere that cost efficiency, performance, and longevity matter.

Mark Ellins, Global Sales and Marketing, Trident Infosol

Mark Ellins, Global Sales and Marketing, Trident Infosol

Nigel Forrester, Technical Marketing Manager, Concurrent Technologies

Nigel Forrester, Technical Marketing Manager, Concurrent Technologies

Jerry Gipper, Executive Director, VITA

Jerry Gipper, Executive Director, VITA

Robert Greenfield, Business Development Manager, Acromag

Robert Greenfield, Business Development Manager, Acromag

Ken Grob, Director of Embedded Technology, Elma Electronic

Ken Grob, Director of Embedded Technology, Elma Electronic

Rodger H. Hosking, Vice-President, Co-founder, Pentek, Inc.

Rodger H. Hosking, Vice-President, Co-founder, Pentek, Inc.

Jacques Houde, President, Pixus Technologies

Jacques Houde, President, Pixus Technologies

Richard Kirk, Director, Core Computing, Abaco Systems

Richard Kirk, Director, Core Computing, Abaco Systems

Haydn Nelson, Director of Product Management for RF and DSP, Abaco Systems

Haydn Nelson, Director of Product Management for RF and DSP, Abaco Systems

Doug Patterson VP, Military & Aerospace Business Sector, Aitech Defense Systems.

Doug Patterson, VP, Military & Aerospace Business Sector, Aitech Defense Systems.

Gerald Persaud, V.P. Business Development, Reflex Photonics

Gerald Persaud, V.P. Business Development, Reflex Photonics

Rob Persons, Senior Systems Architect; Artesyn Embedded Technologies

Rob Persons, Senior Systems Architect; Artesyn Embedded Technologies

William Ripley, Trident Infosol

William Ripley, Trident Infosol

Qianqian Shao, Product Manager, Artesyn Embedded Technologies

Qianqian Shao, Product Manager, Artesyn Embedded Technologies

Valentines, vampires, VITA. Wait. There are sound reasons the three have been racing around my brain as I enjoy reading the thoughtful responses from our Round Table participants.

Valentines, because standards are on the receiving end of valentine-like praise. Take VME.
Richard Kirk, Abaco Systems, lauds its “longevity, modularity and robustness.”
As a highly mature and widely proven technology, VME remains a cornerstone for military programs in the era of constrained defense budgets,” notes Qianqian Shao, Artesyn Embedded Technologies.
And with regard to another standard, Roger Hosking, Pentek, says, OpenVPX has shown deep and sustained support by vendors, and has proven itself in numerous fielded programs.

Vampires, because, well, again, take VME. Asked “What myths about VITA technologies are you tired of addressing,” Ken Grob, Elma Electronics, zeroes in on one particular myth: “That VME will die someday.”   It’s undead. But while vampires owe their continuance to unauthorized helpings of Red Cross-type products, for VME, as Nigel Forrester, Concurrent Technologies tells us, “…there is still a sizable chunk of applications that want a low risk solution that can be satisfied with existing VME technology.
And Robert Greenfield, Acromag, points out, While Navy and Air Force tech refresh projects often require faster VPX and XMC mezzanine technologies for high-speed processing of advanced sensor data (e.g. radar, SIGINT, EW), there are many tasks interfacing mechanical equipment where parallel busses are fast enough. As the old adage goes, ‘if it ain’t broke, don’t fix it.

VITA, because the organization and the individuals offering time and talent to its working groups can take credit for the valentines bestowed here and elsewhere on the benefits of VITA standards, and for VME longevity. Doug Patterson, Aitech, observes, In time, components in electronic systems are replaced by the next generation of more technically capable components, resulting in the obsolescence of the earlier devices. This is a natural progression, but by implementing a standardized, structured approach, such as VITA technologies, that maps out the next electronics evolution, you can effectively guard against, or at least prolong, such obsolescence.

Edited excerpts of our panelists’ responses follow:

Critical Demands

Richard Kirk, Abaco: Our customers are being pushed on two fronts in relation to VITA standards. Many have systems designed around the VME platform, which faces issues of aging and obsolescence. On another front, customers are pushing the state of the art with VPX.
While VME certainly isn’t the most current form factor, it has withstood the test of time with its longevity, modularity, and robustness. There are many VME systems out there with a lot of investment dollars behind them; naturally customers want to keep these investments alive if possible. New system investments are going towards next generation platforms; it only makes sense to save money by routing R&D dollars towards the future.
Abaco is responding by offering modular options for modernization of VME platforms such as with our modular Micro-Mezzanine System (MMS) on a VME carrier and with our Product Lifecycle Management (PLM) service, where we alert customers to component obsolescence and the opportunity to make a life time buy of those components and have Abaco safely store them in a stable environment until the customer needs to take delivery of their board. We have also designed our own VME bridge so that we have control over its supply in the long term.

Rodger Hosking, Pentek: The drivers include high-speed data connections between cards and chassis, thermal management of high-density components like FPGAs, tight synchronization and timing between cards for multi-element or MIMO systems, consistent protocols for moving software radio payload data between systems, and standard implementations of backplane I/O connections for coaxial RF signals and optical signals. Many of the new VITA extensions are addressing these needs. The 2017 releases of VITA 65.0 and VITA 65.1 help define new backplane I/O configurations and add provisions for timing signal distribution. Additional extensions to VITA 48 provide new methodology for cooling, and VITA 49 extensions now provide digitized RF and IF signal definitions for receive and transmit, as well as new status and control features.

Nigel Forrester, Concurrent Technologies: We see three critical demands for our VITA products, driven primarily from defense customers. Two of these demands are long standing: Do more in the same or smaller footprint with less power and provide very long life cycles with configuration control capabilities so that a product can be certified once and then deployed for many years. A third demand, to provide an extremely high level of security, has become much more important during the last year, driving some customers to look at enabling additional security features on existing solutions. All our VITA products have been designed so that we can enable enhanced hardware/firmware/software security capability as required and we’re continually developing our security offering to include more advanced features.

Ken Grob, Elma Electronic: Higher bandwidth over the backplane and more elaborate, smaller, rugged packaging are drivers affecting our customers, as well as pushing VITA technologies forward.
Looking at some of the deeper technical issues, at the chip level, we’re responding to interface speed changes in both PCIe to Gen 3 and looking towards Gen 4 as well as Ethernet adapting to 10Gb and 40Gb rates. Work has already begun on a new backward-compatible, high-speed VPX connector. Also, our new slot profiles and architectures driven by the emerging Hardware Convergence standards include new features like timing boards and radial clocks.

What drivers pushing VITA technologies forward most affect your customers and company, and how are you responding?

And as System-of-Systems Analysis (SoSA), C4ISR Modular Open Suite of Standards (CMOSS) and hardware open systems technology (HOST) drive new SBC, timing, and switch profiles as well as reference backplanes, Elma is producing reference backplanes that address these new backplane topologies. Our participation within the standards groups and developing new backplane and switch technology to prove the new designs is contributing to the VITA standards, via the consortia.
Elma’s approach is to address customer needs at all levels to ensure we’re positioned to respond to market demands, and going further, we take what we have learned and share it as leading members of open standards committees. To support our customer-centric initiatives, we’ve invested in signal integrity modeling and testing, thermal simulation and testing, modern printed circuit design and manufacturing, and enhanced mechanical design and fabrication capabilities.

Jacques Houde, Pixus Technologies: On the Systems Platforms side of our business, more powerful processors are generating more heat in the system. Pixus has introduced our RiCool ™ line of OpenVPX Chassis Platforms. They feature powerful reverse impeller blowers that reside directly above the card cage. They pull the air from the front lower part of the enclosure and blow the exhaust 90 degrees out the rear, providing front-to-rear cooling. The hot-swappable fans provide 284 CFM in the system and only take up 1U of the chassis height. This is especially beneficial for systems with Rear Transition Modules (RTMs), which are common in OpenVPX designs.
On the Components side of our business, the high insertion forces of VPX led us to create special rails and ramp up our rugged insertion handles with a metal engagement claw. With the offset spacing of the panel/PCBs in OpenVPX, we have created card guides, filler panels, and threaded inserts for the architecture. This ensures a precision fit for both aesthetics and EMC considerations.

Gerald Persaud, Reflex Photonics: Electronic Warfare is evolving faster than ever, as computing power increases exponentially and AI algorithms and sensor technology become more sophisticated. With new Electronic Warfare threats arriving at an alarmingly fast pace, VITA systems are being architected to scale quickly and cost effectively.
Fiber optics, with its enormous bandwidth, is the best interconnect technology for scaling systems. Once the fiber optic infrastructure is installed, it supports multiple technology generations with no change. We support scaling with an array of ruggedized optical transceivers offering line speeds of up to 25Gbps and I/O densities up to 24 lanes in small chip size modules. In the near future, we will double line rate to 56Gbps, and we’ll further ruggedize our optical transceivers for more harsh environments, such as space, where radiation hardness is needed.

Mark Ellins, Trident Infosol: Trident Infosol is an active member of the VITA Standards Organization (VSO) and a leading solutions provider for embedded COTS hardware, signal processing solutions, telemetry systems, and enabling software and drivers catering to the real-time computing and embedded system marketplace. A Trident Infosol representative chairs the VITA-74 Technical Committee, which recently ratified and published the ANSI/VITA-74.0-2017 (VNX) standard.
We see a growing demand from the embedded computer customer base for rugged Small Form Factor (SFF) COTS computers. It was a logical decision to develop the rugged SFF ANSI/VITA-74.0-2017 standard with VITA rather than with other standards bodies more focused on the commercial or industrial space. VITA standards such as VME and VPX served the market well for decades and are still preferred for most MIL/Aero applications over other industrial standards or proprietary options. VNX is a SFF derivative of VPX, much the same as VPX was an evolution of VME, and benefited from the lessons learned during the VPX / VME maturation process.

Jerry Gipper, VITA: The demand for Open Systems Architectures for hardware and software is growing each year. Existing standards need additional work and new standards need to be developed. This means that working groups are very busy on documenting standards and taking them through the review and approval cycle. VITA is encouraging more involvement from both product developers and the user community to contribute and collaborate during the standards development process. VITA has been working to improve the tools and the processes to ease the efforts and to make standards development more streamlined.

Relieving cost anxieties

The Artesyn Embedded Technologies MVME5500 SBC. Originally released in September 2002, it will continue to be supported to at least 2025.

The Artesyn Embedded Technologies MVME5500 SBC will continue to be supported to at least 2025.

Our customers are asking for ways to build older military systems that they have supported for years. An amazing number of new systems are being fielded based on programs awarded well over 10 years ago. Of course, new systems designs are being awarded, and VPX based products may be an option. But many programs cannot afford a complete redesign that would allow the potential use of VPX. Having said this, these same customers fear the cost to maintain these older systems because of VMEbus board obsolescence or component obsolescence on existing VMEbus boards. We’re helping customers reduce risk by maintaining our current set of products, including, for example, our MVME5500 SBC to 2025 and beyond and doing so without any major changes to these products. Our commitment to support our products based on the IDT Tsi148 has had a significant effect on our customers and their customers in maintaining some very old designs.

— Rob Persons, Artesyn Embedded Technologies

Incorporating Innovation

Doug Patterson, Aitech: Processor and memory technology roadmaps are a critical element in developing new COTS products for our customers. Their appetite for increased performance, more RAM and ROM memory, and smaller size, weight and power, coupled with the demand for lower cost continues to pull the industry as technological advances push the market.

What trends will be most pertinent to systems integrators working with VITA technologies?

Bill Ripley, Trident Infosol: The most pertinent trend we see, even happening today, is a trend to merge VITA standards with commercial and industrial standards such as PICMG’s COM Express, PCI-SIG’s MiniPCIe and SGeT’s SMARC modules, to name a few. In the short span of a decade, with ever-increasing integration and functionality, everything is changing. With the evolution of small module-based computing techniques and the nearly universal adoption of System-on-Chip (SoC) components, the end user can now receive better solutions, at lower costs, with lower development risk, which more closely meet their needs.
Our customers still want to adhere to VITA standards, but they are also wanting to take advantage of best practices used in the commercial, medical, and heavy industrial market. VNX is such an example of on the one hand following an ANSI/VITA standard designed for rugged SWaP computers, and on the other hand using popular industrial standards for I/O, memory, and System on Chip (SoC)/ System on Module (SoM) processors.

Rodger Hosking, Pentek: VITA initiatives will continue to accommodate market demands by incorporating innovations developed by embedded product vendors and systems integrators. This infrastructure not only helps each member of the vendor community, but also embedded system customers, particularly defense organizations with 20-year life cycle concerns. Now that VITA 65 OpenVPX has shown deep and sustained support by vendors, and has proven itself in numerous fielded programs, it remains the most widely adopted architecture for high-performance embedded computing since VMEbus.

Qianqian Shao, Artesyn: VME will maintain a sizable market for the next five years or even another decade. In this market, PowerPC architecture based VME solutions account for more than 80 percent of market share. As a highly mature and widely proven technology, VME remains a cornerstone for military programs in the era of constrained defense budgets. VME is expected to continue its important role in system refreshes and upgrades as sequestration has extended the life cycle and altered the terms of maintenance and upgrade for many existing programs. On the other hand, VME still represents an optimal solution for the new programs requiring low risk and low cost with its salient competitive advantages of low power, small system size, and proven experience in deployment.
As part of the group of innovative companies that invented VME technology over 35 years ago, Artesyn has laid the groundwork and consistently worked to enhance and extend VME technology. Artesyn will continue to be committed to VME technology. To underline our commitment, we have been investing heavily in our VME offerings. We have secured a number of critical EOL components, including the Tsi148 VME to PCI-X chip and the Marvell system controller chip, to ensure that we can continue to offer an extensive portfolio of VME boards, including up to at least 2025.

Nigel Forrester, Concurrent Technologies: Our biggest opportunity is still based around the acceptance of using Commercial Off the Shelf (COTS) technologies. More than 20 years since Secretary of Defense William Perry issued his memo mandating the use of COTS by the Department of Defense, there are quite a few defense solutions that are designed from scratch rather than constructed using ‘good enough’ COTS building blocks. This is often because a ‘home made’ product exactly fits the requirements, whereas COTs building blocks provide a superset of features.
VME I/O boards were relatively simple to design, but today’s high-performance VPX based server and processing solutions are much more complex. Conversely there is high pressure to save program cost and to deliver on time, both of which are helped by using off the shelf boards when available. To help make our COTS products easier to use, we continue to develop our middleware solutions that provide security and connectivity enhancements, allowing system integrators to concentrate on their application.

Gerald Persaud, Reflex Photonics: The trend towards autonomous or SMART systems is one that will greatly affect system integrators working with VITA technologies. VITA technologies will have to support machine learning or artificial intelligence capable of accurately characterizing the environment and recommending the best course of action based on massive data input from sensors and other data sources. VITA systems will need to deliver enormous processing power in a smaller and smaller footprint. From the start, Reflex Photonics was structured to make the smallest rugged optical modules capable of supplying enormous BW and optical channels. Today our rugged technologies are field proven and well positioned to take advantage of the trend for smarter and smaller systems.

Jerry Gipper, VITA: The open systems architecture demand is very important to VITA technologies. VITA is all about open standards and more of the critical embedded computer community is realizing the value of the work that goes into developing the right sets of standards. No one solution fits all, and new needs drive the demand for additional efforts. VITA has a roadmap study group focused on key technology areas to track trends that may influence the development of standards to meet future requirements.

Ken Grob, Elma Electronic: Optical networking within the chassis will become the primary path for high-bandwidth, card-to-card communication. Security and chassis management will become a major factor in the choice of solutions.
Increasing technology changes and demands are compounding the complexity of the overall system—boards, backplane, and SWaP—putting pressure on design, design verification, manufacturing, test and project management. Elma addresses this complexity all the way through the process, from requirements review to delivery.

Jacques Houde, Pixus Technologies: We see a growing trend of special requirements for RF and optical. Pixus is incorporating VITA 66 (Optical) and VITA 67 (RF) designs into our OpenVPX backplane offering.
Haydn Nelson, Abaco: We’ll see continued adoption of standards like VITA 65 as the drive for lower size, weight, and power underlines the benefits of the 3U VPX form factor. On the semiconductor side, there has been a convergence of processing technology with GPU cores included in some versions of Intel processors and multicore Arm processors being included in FPGAs, such as Zynq from Xilinx. As Moore’s Law continues to provide more resources on silicon, chip designers will find creative ways to take advantage of the space. Many heterogeneous digital architectures today blend CPU, GPU, FPGA, and RF. This convergence of technology at the chip level will lead to smaller designs and more adoption of the smaller form factors such as 3U VPX.
Greater integration at the chip level will also lead to greater data bottlenecks., the limitations of data movement will lead to many customers looking to optical interconnect for its speed, signal integrity, and security properties. Abaco is already capitalizing on these trends with VITA 66.4 offerings as well as being early to market with the Zynq Ultrascale+ device family on the VP880 product line.

If It Ain’t Broke…

Doug Patterson, Aitech: Three reasons…customer, customer and customer. If the technology remains relevant and customers still demand it, then it makes sense to continue to build it for as long as is feasible. At Aitech, we support a product’s declared lifecycle as it aligns with our COTSLifecycle+ program, which provides a minimum of 12+ years of products availability and support.

Robert Greenfield, Acromag:  Acromag remains dedicated to supporting VME and VITA 4 IndustryPack (IP) modules developed 20 years ago. Our customers continue to specify these product technologies because of their proven track record of dependable operation, compatibility between vendors, and cost-effectiveness.
In many aerospace and defense applications, IP modules hosted on VMEbus carrier cards still deliver an acceptable level of performance and are quite economical compared to newer embedded computing technologies. While Navy and Air Force tech refresh projects often require faster VPX and XMC mezzanine technologies for high-speed processing of advanced sensor data (e.g. radar, SIGINT, EW), there are many tasks interfacing mechanical equipment where parallel busses are fast enough. As the old adage goes, “if it ain’t broke, don’t fix it.” So, many subsystems employing KHz-speed A/D and D/A converters, discrete-level I/O, and RS-232/485 interfaces are carried forward with little change. The biggest challenge supporting these older products is with end-of-life components. However, Acromag has successfully redesigned many IP modules with form, fit, function compatibility using inexpensive FPGAs for extended availability to prevent any unwanted software or system changes.

Three main reasons your company continues to support the VITA technologies you’ve supported the longest?

ControlSafe® SIL4 certified systems serve the needs of customers who employ VME for traditional SIL2 applications

Rob Persons, Artesyn, notes that VME has a place in traditional SIL2 applications within the rail industry, while the company’s ControlSafe® SIL4 certified systems serve the needs of customers who employ VME for traditional SIL2 applications, but also need to add safety critical control systems.

Rob Persons, Artesyn:

  1. Military budget constraints represent a big opportunity to continue to support legacy VMEbus support. The military continues to rely on systems that were designed a long time ago that were designed using VME. As other companies chance new programs with VPX, we see a nice business supporting these older programs.
  2. If it ain’t broke, don’t try to fix it! We have a number of large commercial customers who have control systems based on VME with custom designed VMEbus boards. They are happy we are continuing to support some of our legacy products and are planning on new products based on VME. This reduces their need to reinvest in new custom products for another platform architecture. They can add capabilities by migrating to newer VMEbus technologies.
  3. For one older product family and one that is brand new, we see a lot of synergy with VMEbus products. With the huge adoption of ATCA in Navy programs, many times the subsystems that are attached to these ATCA based tactical systems are based on VMEbus. We have the ability to support both types of applications. We also see customers who have traditionally bought VME products from us to build SIL2 based systems for the rail industry have the need to address new safety requirements with SIL4 based systems. Our ControlSafe® SIL4 certified systems are a nice complement for those customers who need to add safety critical control systems while still using VME for traditional SIL2 applications.

Ken Grob, Elma Electronic: It comes down to the relationship between the supplier to us, to the customer and to the integrator. Many technologies can do the job, but supplier relationships to the market drive technology choice.
And the technology continues to evolve, so it maintains its viability, and is still highly regarded among well-established defense contractors. The argument can be made that in terms of processing power per unit of space, VPX is out in front. It may not be the cheapest, but if there are no other options, then price, at least for now, is not much of an issue.

Gerald Persaud, Reflex Photonics: Reflex Photonics VITA Technology focus is on optical interconnects such as VITA 66 and 67 standards. We see great benefit from active optical blind-mate connectors to simplify system assembly and upgrades. As well, optical blind-mate connectors reduce space and enable field servicing. Our latest LightCONEX blind mate connector integrates an optical module into the backplane connector, thus eliminating the need for separate optical modules, cables, and cable routing. LightCONEX simplifies board assembly since there is no optical cable to assemble and no chance of cracking the glass fibers during assembly.

Mark Ellins, Trident Infosol: First, our customers demand computers that are built using COTS technologies, based on the notion that using COTS often minimizes technical risk, economic costs, and the potential for schedule creep. The COTS solutions are generally based on ANSI/VITA standards such as VME, VPX, and VNX. Second, since we are also system integrators, there are many options for suppliers of ANSI/VITA standard-based modules to choose from. Last, purchasing standard-based products not only allows for better prices overall with less need for NRE, they also promise a longer life cycle, thus complying to typical obsolescence management requirements.

Richard Kirk, Abaco: We recently released a white paper titled “VME Forever.” The title is somewhat tongue in cheek, but the paper describes why it is that, so many years after its introduction, VME is still relevant to customers and to the market. The adoption of VME for new programs may be shrinking, but it has an extensive installed base, it’s rugged, it’s proven and many customers—especially those who were late adopters—remain committed to it as a trusted and safe option.
It’s clear that VME isn’t the platform of the future. VITA 65 with VPX has shown a lot of traction and is poised to see growth in adoption over the coming years. Abaco has been committed to VPX since the standard was first launched, and was the first company to launch a 3U VPX single board computer. We see VPX as having many of the same benefits as VME—modular, rugged, and incrementally adaptable. Just as VME went from 3-row connectors to 5-row, from 32-bit to VME320, we see VPX being adapted to use higher bandwidth connectors including optical, and adapting to fabrics such as PCIe and Ethernet as it moves through successive generations.

Bill Ripley, Trident Infosol: The laws of physics and thermodynamics constitute our biggest challenge. Customers continually want stronger processors, GPUs, and even GPGPUs for their applications. They envy the processing power of what’s available for the gaming market. Small Form Factor computing doesn’t always allow the luxury of the stronger processors given the challenge in cooling the system. While it can be said that companies such as Intel, NVIDIA, and AMD have drastically reduced the power budget needed to achieve a given processing benchmark, because of the ever-increasing desires of the industry, it is still a challenge to satisfy an insatiable appetite for higher performance. The VITA-74 committee understood this and created a standard that is optimized for conduction cooling and for taking heat out of small packages.

Typical thermal VNX 19mm module design

This figure shows a typical thermal VNX 19mm module design, with one module’s end plate removed. Here, the VNX 19mm compliant plug-in module is made up of two stacked boards; with the lower board having the 400-pin high-density VNX backplane connector. Hot components can be thermally connected to the outer enclosure using compressible thermal interface materials (TIM) and a metal skyline mechanical heat spreader, contoured to maximize TIM efficiency, which then interfaces to the module packaging. Multiple plug-in modules are installed in a chassis with side-plates designed to conduct heat to the chassis fins or heat exchanger.

What developments are you keeping an eye on with regard to the more efficient processing of graphical information and/or with regard to other current challenges?

Haydn Nelson, Abaco: Graphical information processing is critical to many military embedded systems deployed today. By and large, the GPU is the core technology element in graphical processing as these are designed with graphics in mind. However, we have seen some adoption of GPU technology for other applications such as cognitive algorithms, neural networks, and deep learning systems.
One area that has seen growth is where graphical processing is used in systems with extremely low latency requirements. Low latency is often required in systems such as those for degraded visual environments (DVE), automated self-protection systems, and autonomous vehicles. In each of these three areas, there are common video processing elements using FPGAs, standard CPUs, and graphics processors. The FPGA processor is often used for streaming portions of the algorithm that can be fixed in gates; however, the GPU for its parallelism is often center stage. At Abaco, we have a long-standing relationship with top GPU chip providers such as NVIDIA, and we offer many GPU capabilities in rugged form factors such as the GRA113.
With the GRA113, we significantly increase performance per watt, which translates into either more performance in the same space and/or reductions in overall heat, space, spares, maintenance, and money.
Furthermore, the ability to configure video output offers an increased flexibility in applications that often need to output the video to displays with differing resolutions and legacy standards such as RS-170. This capability has the potential to translate into lower costs by combining configurable output on one board rather than having to buy separate boards and offers truly rugged GPU capability in products such as the 3U VPX form factor GRA113.

Doug Patterson, Aitech: Leapfrog advances in GPGPU technologies from companies including NVIDIA and others…from handling increased graphics output to operating as DSP engines, GPGPUs can take on the cumbersome task of image characterization as well as identification. This continuous learning process provides a solid output for the system to ‘act on.’
Using sophisticated pattern recognition, this type of processing works like the brain, employing a multilayered, neural network approach. System intelligence and efficiency enhances, since images can be identified more easily, while context assigned to the image itself provides an even deeper understanding of the surrounding environment that image resides within.

Qianqian Shao, Artesyn: Our VME solutions are used in airborne radar systems and semiconductor lithography systems machines by leading companies worldwide. For the design of new VME boards, selection of processors with powerful graphical information processing capability is key for us to stay relevant to the target applications. In addition, in-house or third-party mezzanine card solutions remain a flexible approach for us to provide add-on processing power while using existing VME boards as infrastructure.

Ken Grob, Elma Electronic: In graphics or the use of GPGPUs, we’re seeing customers use the NVIDIA Tesla architecture like the M6 for number crunching, instead of doing this work with an FPGA. Also, lower power applications are using the NVIDIA Jetson SOM. There’s some trade off with Intel processors now including GPGPU cores within the chip, so for certain applications the local capability can be used.

What myths about VITA technologies should be put to rest?

Richard Kirk, Abaco: Three myths come to mind as they are related to VITA standards.

  1. The end of VME is here… Yes, it is declining, but it will be a slow fall off as systems with VME are phased out. It doesn’t make economic sense for the entire Navy to abandon VME when every ship uses it for many systems such as a navigation and weapons control. Abaco is committed to maintaining VME for as long as our customers need it–and right now, that looks like a very long time!
  2. Other standards such as COM Express will push VME and VPX into the background. Bearing in mind that VME and VPX are primarily used in defense and aerospace applications, there is absolutely no sign that any other standard is going to make a significant dent in this market. Small dents, maybe–but in general, system developers are focused on rugged, expandable, high performance solutions which is exactly what VME and VPX are all about. That’s no surprise, as the users played a big part in defining the specifications in the first place,
  3. FPGA Mezzanine Cards (FMC) aren’t inherently rugged. While it is rare, we occasionally face fear of using a modular mezzanine approach in systems that have stringent environmental–specifically, high G shock and vibration – requirements. There is nothing about FMC that prevents a high ruggedization level. In fact, the FMC form factor connector is designed to accommodate conduction cooling and to survive in the harshest of environments. Modularity doesn’t necessarily mean fragility. The XMC and PMC mezzanine standards have stood the tests of ruggedization and FMC is no different.

Doug Patterson, Aitech:

  1. Can we get past the bastardization of “COTS” once and for all? Add the word “-available” after Commercial to get “Commercially-available Off-The-Shelf” and it solves a figurative ton of misnomers for the military and defense markets/community at large.
  2. Not every application needs to scream data with PCIe speeds over a system backplane. Get over it—there are defense and space markets for “simple” control systems where the “speeds and feeds” (and added costs of OpenVPX) are just not warranted.
  3. New, SWaP-oriented, “standardized” small form factor (SFF) “systems” should concentrate on standardizing the mechanical form and size, not the electronics inside it, or the connectors it uses. Those are application-specific requirements, let the application dictate the “guts” and move on. Can you imagine the “cluster” if the VITA standards 30 years ago decided to define what an SBC was and the functionality within it? We’d still be stuck with a separate system controller with one interrupt controller and a system clock crystal oscillator on it and separate memory RAM/ROM boards on VME.

Rob Persons, Artesyn: The obvious one, that VMEbus is dead. Although the market is not growing in real terms, the decline is by no means steep. There is plenty of need and Artesyn Embedded Technologies understands how to support products that must have a long life. We design and manufacture our products and own our supply chain and this gives us a huge amount of capability to support our customers for the long run.

Nigel Forrester, Concurrent Technologies: For nearly 10 years there have been predictions that VME is dead. This is still far from true. Whilst new applications requiring high bandwidth interfaces are advised to utilize VPX technologies, there is still a sizable chunk of applications that want a low risk solution that can be satisfied with existing VME technology. Conversely, the hype level around new and exciting technologies always seems to exceed real demand.

Ken Grob, Elma Electronic: That VME will die someday.
Rodger Hosking, Pentek: One myth is that OpenVPX is “too open,” i.e., it offers too much flexibility in defining cards, slots and backplanes. In 2017, this myth was significantly shattered with the advent of the latest releases of VITA 65.0 and 65.1 in two major ways.
First, details of the “profiles” for cards, slots and backplanes were migrated from the base standard into the new VITA 65.1 standard, making it easier to document the details of each and easier to define new ones. Secondly, many of the options for connectors for optical and coaxial backplane I/O were more consistently defined for improved compatibility across vendors.

Gerald Persaud, Reflex Photonics: Myths we hear are the ruggedness of fiber optic interconnect and the cost of fiber-optics versus electrical interconnects.
Yes, fiber-optics may require some additional handling during installation and maintenance but this is small compared to the numerous benefits such as immunity to EMI, light weight, superior bandwidth, reach and scalability. As well, there are many technical approaches to manage the handling issues and Reflex Photonics can provide solutions.
With respect to cost, fiber optics interconnects are generally more expensive, but one has to consider the overall effectiveness of the systems and the cost of upgrades over the life of the equipment. For example, if the cost of fiber optics increases the system cost by 5% but enhances the system by 20% then one has to decide if the additional 15% effectiveness is worth it. For an expensive aircraft, this may mean the difference of losing that aircraft because it was unable to counter a threat quickly enough. As well, the cost of upgrades could be far greater if one has to replace the electrical with optical interconnects to support more advance technologies.

Mark Ellins, Trident Infosol: The key issue in the embedded computer market is being as flexible as possible to customer requirements without having to charge NRE. Standards have a stigma of being a less flexible or compromised design compared to proprietary design. And in the “old-school,” this stigma was prevalent, even with VITA standards. The major recent paradigm shift is related to the introduction of System on Chip (SoC) and System on Module (SoM) technologies that allow the use of proven industrial standards like COMe, MiniPCIe, mSATA, and SMARC. Manufacturers can now simply design hybrid carrier boards that are compliant to VME, VPX, or VNX, and utilize mezzanines that support the industrial standards and can support the requisite CPU / GPGPU / FPGA processing, I/O, video, communication, serial capabilities. Taking full advantage of appropriate standards based solutions debunks the oft-misplaced notions that the use of COTS yields a non-optimal design.

Jerry Gipper, VITA: Myth: The technology is too expensive.
Expense is relative. A critical embedded computing solution requires a completely different cost reference than a common consumer electronics product. Considerations must be made for reliability, interoperability, life cycles, unique needs, and many other areas that are not a concern for a typical commodity consumer product. While the commodity consumer market is driving the development of much of the electronics in use in all computing platforms, education on cost considerations for critical embedded computing platforms will help in understanding the additional expense.
Myth: Time to develop new standards takes too long.
The time to run a concept through the development and approval process is a direct relation of the energy put into developing the standard. Focused working groups driven to get a standard approved can often complete the efforts in well under a year. The process has certain time gates to pass to ensure that contributors and reviewers have adequate time to review the materials, but no other process time restrictions apply. Working group chairs that are well motivated can get a standard through in a very reasonable period of time.

Merging

There’s an application drive for more processing—both CPU and FPGA—and on the analog side, the thirst for wider bandwidth I/O is never quenched. In 2017, we released a state of the art 3U VPX FPGA carrier that supports a single VITA 57.4 FPGA mezzanine card (FMC+). We then followed on with our first FMC+ board, the FMC134, which delivers on the need for more analog bandwidth. The FMC134 uses two TI ADC12DJ3200 devices, which can produce upwards of 24GBps of data into the FPGA. All this data is passed to the FPGA and must go onto very wide internal data interfaces. This is kind of like merging from the Autobahn in Germany onto a 12-lane freeway in Houston Texas—you are going to need a lot of width. Fortunately, the VP880 features the latest from Xilinx with its Virtex Ultrascale device to meet the need.Once the data is inside the processing fabric, there are two options; reduce the data or transfer it off the boards. The latter leads to a significant data rate issue, as 24GBps of data is beyond what is possible with a copper interface. This is somewhat reminiscent of the old saying “you can’t connect a firehose to a garden hose.” We see many customers looking to VITA standards like VITA 66.4 to solve this high-speed data offload problem. We have helped our customers by recently offering our latest FPGA carrier, the VP880, with this option.

Haydn Nelson, Abaco Systems

More Capability on a Single Board

Ken Grob, Elma Electronic: Using high density 3U VPX cards allows system footprints to be reduced, and we are looking towards new NVIDIA and AMD GPGPU products as well as Xilinx and Altera FPGA products through our partners. High definition applications require all the above and Elma has the design capability to integrate these solutions.
To continue to support SWaP goals as performance increases, new cooling techniques will be required to allow higher power, small footprint boards to be cooled. As one compresses the design, the power density increases. Adopting new cooling schemes, like VITA 48.8, can address higher performance 3U VPX implementations.
Proving next generation connectors will allow VPX to address PCIe Gen 4 and follow-on Ethernet standards.

Jacques Houde, Pixus Technologies: Pixus continues to develop rugged ATR and rackmount solutions that reduce SWaP-C. It starts with a highly knowledgeable team that knows how to efficiently design a system platform to reduce the size and weight. We are introducing a compact 3 payload slot OpenVPX ½ ATR that is robust and designed to meet avionics needs such as MIL-STD-810, MIL-STD-461, MIL-STD-704, and DO-168. This rugged conduction-cooled system maximizes heat dissipation for the higher OpenVPX wattages. Leveraging over 25 years of modular chassis designs helps us minimize costs and lead-times. Additionally, Pixus is developing more compact and cost-effective power solutions to help reduce SWaP-C.

Gerald Persaud, Reflex Photonics: In addition to reducing size, weight, power and cost, Reflex Photonics produces the most rugged line of optical modules to meet high reliability and long operational life. As well, we plan to release modules with up to 600 Gbps in a chip that is less than 4.5 cm2. Recently, we released our LightSPACE radiation hardened optical module that meets the harshest environmental requirements for space.

Raptor VNX computer, from Trident Infosol

Mark Ellins, Bill Ripley, Trident Infosol: We are one of the first suppliers of the recently approved ANSI/VITA 74.0-2017 standard, better known as VNX. VNX is a COTS module standard for SFF, conduction-cooled, inherently rugged modules, including Single Board Computers, signal processors, storage modules and I/O devices which are to be used as part of an integrated system. Utilizing Modular Open Systems Architecture (MOSA) principles, VNX provides the first standards based, slotted module approach to building conduction-cooled SFF systems. Trident has released a VNX computer called the Raptor

A VNX computer, the Raptor, from Trident Infosol.

How does your company use VITA technologies to reduce SWaP?

Richard Kirk, Abaco: There are two fronts where we can see how reduction of size, weight, power, and cost relates to VITA standards. With the emergence of many ‘systems-on-chip,’ we expect to be able to offer more capability on a single board into the future. While the size of a single 3U VPX board isn’t getting smaller, technology convergence at the chip level has given Abaco the ability to offer highly integrated solutions that often reduce the number of cards needed in total. Fewer cards means the whole system can shrink in size, weight, power, and cost. For example, a customer may need an Intel single board computer, an ADC, discrete digital I/O, an IRIG signal, and three ARINC 429 ports. With Abaco’s Micro Mezzanine System, all this processing and I/O can fit into a single 3U VPX slot (the Abaco SBC329) and an MMS XMC Carrier.
Another factor is the advances in thermal management techniques, which mean that a high-power CPU’s maximum performance can now be maintained even at the highest operating temperatures—potentially meaning fewer CPUs in the system, and all performing optimally. The convergence of high performance SoCs with this new thermal management technology opens up many opportunities.

Doug Patterson, Aitech: Not all VITA technologies are widely adopted by the customer base. Having said that, Aitech’s product marketing managers will continue to canvas our key customers regularly to determine the technologies needed for their project successes. And if there’s sufficient market traction for that SWaP technology, with sufficient ROI to make the exercise worthwhile, we’ll produce those products for our customers, so they have what they need or want into the future.

Qianqian Shao, Artesyn: While extending the life cycle of our VME products, Artesyn is also planning to develop new VME boards that will enhance our tiered product portfolio, which includes committed research on VME bridge solutions for future portfolio additions. Artesyn’s extensive VME portfolio based on Power Architecture processors offers customers flexibility to migrate between boards when they seek optimal solutions for their applications. With a deep understanding that software compatibility is vital to make a product migration successful, Artesyn always goes the extra mile to provide technical support to help customers migrate smoothly.

Taking the Fast Bridge between Neural Networks

This article was originally published in Embedded Systems Engineering in Spring of 2018 and written by Lynnette Reese, Editor-in-Chief.

Gerald Persaud, V.P. Business Development at Reflex Photonics gets interviewed by Embedded Systems Engineering

Gerald Persaud, V.P. Business Development

Gerald Persaud is responsible for overseeing global marketing, business development and customer initiatives related to the Reflex Photonic's product lines, as well as managing product development and customer technical support. 
[expand title="Read more" swaptitle="Close"]Gerald has over 20 years experience in telecom and defense. Prior to joining Reflex Photonics he held senior management roles in engineering and business development at Nortel, General Dynamic Canada, and Celestica. Gerald has developed many leading products in optical communication, wireless and advanced computing. Gerald doubled revenues at start-up Coresim in one year and precipitated an acquisition by Celestica. He also won the largest design contract ever in Celestica for an OTN switch.
Gerald holds a B.S. in Electrical Engineering at McMaster University.[/expand]

How do neural networks scale up to 20,000 processors and beyond?

How does machine learning scale massively if connections are a potential weak link in the race for acceleration?

Neural networks have been around since the 1950s. The advent of fast, massively parallel processors like the Graphics Processing Unit (GPU) have made neural network applications like object recognition feasible. Neural networks are one means used to create Artificial Intelligence (AI). The latest iPhones now have an AI chip, primarily to offload face recognition tasks.1 Voice translation tasks would also benefit from an AI chip. Google provides voice translation as long as there is access to a cloud. The ability to translate directly from a phone without requiring Internet access to Google engines would be advantageous, and it’s possible that the iPhone is headed in that direction.

AI often needs real-time operation. China’s answer to policing during the New Year’s travel crush involves providing police officers with smart glasses for rapid facial recognition. If using traditional video cameras, the suspect has left the scene by the time policemen arrive. China’s police force is equipped with smart facial-recognition glasses connected to a pocket-sized data module for identifying up to 10,000 individuals. The module’s database dramatically reduces latency versus a cloud access system. Recognizing a face in as little as 100 ms, police can immediately act upon the information.2 At the other end of the spectrum, enormous banks of GPUs in server farms can require low latency. Scaling up to 20,000 processors and beyond needs relatively high bandwidth between GPUs to communicate quickly. Connectivity can be the slow link in a system with the fastest GPUs. A company located in Quebec, Canada, called Reflex Photonics, has a connectivity solution that fits the bill, enabling a machine learning or AI infrastructure to scale up with more processors. Reflex Photonics provides tiny optical transceiver chips that can move a tremendous amount of data, which reduces the latency between GPUs so that they can appear as though working seamlessly, in parallel.

The Mercedes Freightliner Inspiration truck is the first road-approved truck for autonomous operation

The Mercedes Freightliner Inspiration truck is the first road-approved truck for autonomous operation. (Source: Daimler)

One area that Reflex Photonics serves includes VPX systems. According to the VMEbus International Trade Association (VITA.com), “VPX is a broadly defined technology utilizing the latest in a variety of switch fabric technologies in 3U and 6U format blades.” VITA technologies are well known in military and aerospace. Interconnects for serial switch fabric in venerated technologies like XMC, VPX, and VXS, as well as new standards that include Gigabit Ethernet, PCI Express, and Serial RapidIO are in the VITA ecosystem. Form factors in this technology include credit card-sized processing platforms up to the 6U Eurocard.
Technology for optical interconnects is vital in the industry, since processor speeds are outpacing copper wires for bandwidth and latency among devices, including VPX. During our recent interview, Gerald Persaud, VP Business Development at Reflex Photonics, told me this is a solvable problem, even in the harsh environments of military and outer space. Edited excerpts follow:

Lynnette Reese (LR): What is Reflex Photonics currently developing?

Gerald Persaud (GP): Today, we are focused on aerospace and defense, and industrial markets. Our expertise is delivering chip sized rugged high bandwidth optical transceivers that work in the harshest environments, such as space. For example, we were recently selected for a major satellite program because our parts could meet the required 20 years lifetime in space. Many optical transceiver suppliers claim high bandwidth operation at 25Gbps per channel but only for an operating temperature of 0 to 70oC. All of Reflex Photonics’ rugged transceivers operate error-free over a temperature range of -50 to 100oC while also meeting severe shock, vibration, damp heat, and thermal cycling requirements.

Machine vision will benefit from the high-bandwidth of the LightVISION industrial optical transceiver

Optical interconnect for high-speed, high- bandwidth 10GigE and 40GigE cameras used in machine vision.

LR: Reflex Photonics’ expertise is in ruggedized optical communications. How did your process for dealing with the challenges of harsh environments evolve?

GP: In 2002 when we started the company our goal was to create a chip-size optical module that could be solder reflowed to support low-cost board assembly. This was much harder than we had imagined due to differences in material properties such as thermal expansion, thermal conductivity, and curing processes. Over the years we were able to incrementally improve our manufacturing processes from a commercial offering to a full space-qualified part. An excellent understanding of materials and processing is critical to the successful production of high-bandwidth rugged optical modules.

LR: What is on your roadmap?

GP: We plan to release higher channel speeds up to 56Gbps, more I/O density such as 24 transmitters or receivers in a chip size optical module. As well, we will continue to harden our parts to meet even wider temperature extremes of -65 to 125oC. Another product we recently released is active blind-mate optical connectors called LightCONEX®. We have gained a great deal of interest in this solution from the VPX community, as it frees up a lot of board space and simplifies field upgrades.

LR: Can you give an example where Reflex Photonics has a play in VPX for machine learning?

GP: One example of this is in unmanned vehicles where machine learning is critical for autonomous operation. Many sensors are interconnected to machine learning VPX compute farms via an optical switch. Optical interconnect, with its long reach, high bandwidth and light weight, is the only viable solution for advanced Autonomous Vehicles (AVs). From the start, Reflex set out to make the smallest rugged optical modules capable of supplying enormous bandwidth (BW) and optical channels. Today, Reflex Photonics’ rugged technologies are field proven and well positioned to take advantage of the trend for smarter, smaller, and robust systems.

LR: How are you dealing with power challenges in a Small Form Factor (SFF)?

GP: Power is indeed a challenge for mobile vehicles, which have a limited amount of power to supply onboard electronics. Today a 150Gbps chip consumes about 1.3W. However, as bandwidth demand grows from 150Gbps to 2400Gbps over the next five to 10 years we cannot scale power linearly or the same chip will consume 21W. And there are multiple chips per board!  We will need to introduce techniques to improve optical coupling efficiency and lower laser bias currents. As well, laser drivers and amplifier will need to operate at lower voltages. Closer integration of the drive electronics with optical transceivers could save a lot of power as the need for Clock and Data Recovery (CDR), equalizer, or pre-emphasis could be eliminated.

LR: What are your competitors doing? How is Reflex Photonics any different?

GP: Everyone including Reflex is racing to increase BW and interconnect density. However, in the aerospace and defense sector, suppliers must also meet the challenges of operating in a very harsh environment while keeping space, weight, and power [SWaP] to a minimum. Reflex is different in that we were the first to deliver a 150Gbps chip-size optical module that could operate from -50 to 100oC while consuming 1.2W. Most recently Reflex launched the first radiation-hardened parallel optical chip for space applications. These chips passed extreme environmental test conditions that our competitors were unable to meet. This is excellent news for the space industry, where size and weight are critical and smallsats are expected to do far more than their predecessors.

LR: I have always considered price to be a specification. How is your pricing affected by ultra-hardening for space?

GP: The price differential is not as significant as most would expect. In the old days when you said “space,” it meant 10 times the price. Those days are gone. There might be 30% increase in price for space grade over a military grade device. One grade down from military is the industrial device, which has similar operating temperatures but is not expected to have as long a life as Space and MIL grade parts.

The LightCONEX blind mate optical interconnect for VPX embedded computing systems, consists of a module connector and a backplane connector compatible with the VITA 66.4 standard.

LightCONEX 50G and 150G is a rugged blind mate optical interconnect for VPX embedded computing systems. Used in military systems, optical interconnects are faster and more resistant to noise than copper-based connection systems. The potential for use in automotive reflects the need for high bandwidth communications at real-time speeds for modern automotive AI systems.

LR: Can you detail some of the challenges for optics at extreme operating temperatures?

GP: Optical transceivers require exact alignment (less than five micrometers) of the laser or photodetector to the optical coupler. One challenge is maintaining this alignment over a wide temperature range. Reflex developed a patented approach using materials with low coefficient of thermal expansion and a simple coupling structure with no intermediate lens to maintain alignment over a wide temperature range of -57 to 125oC. Another challenge is having a cost-effective sealing method (for moisture resistance in the optical path) that will withstand many thermal cycles without compromising the mechanical integrity of the module. Of course, there are other challenges like radiation hardening, solder reflow temperature survival, low power, optical sensitivity, and signal integrity.

LR: What are the different grades of products that you have for harsh environments?

GP: Most of our sales are for MIL, Space and Industrial grade parts. We offer some commercial grades such as QSFP and CFP for Telecom/Datacom markets. Our industrial components are used in many applications such as commercial aircraft, semiconductor wafer inspection, and instrumentation and tests. Most recently, we have had a number of automotive applications for our industrial parts.

LR: Where would the automotive or transportation industry need rugged optical transceivers?

GP: The automotive industry is quite large and includes cars, city buses, transport trucks, and other vehicles. We expect as self-driving or assisted driving goes mainstream fiber-optics will interconnect all systems in the vehicle. Compact AI engines will connect many sensors to automate driving. The vehicles of tomorrow will provide great energy efficiencies, less pollution, and a comfortable and productive driving experience. NVidia is now offering small form factor AI engines that are already deployed in Unmanned Aerial Vehicles.

LR: Any optical transceiver is still going to need fiber to transport the signal in a system. Isn’t vibration a real problem for this kind of signaling in a vehicle?

GP: No. Our parts have been tested to MIL-STD-883, Method 2007.3 for vibration and Method 2002.4 for shock. Vibration is 20 to 2000Hz, 20g, 16 minutes per axis and shock is 500g, 0.5ms pulse, 5 repetitions, 6 directions. These tests were done while transmitting and receiving 150Gbps with no errors.

LR: That’s impressive. What distance and latency are we talking about?

GP: Distance in AVs are typically less than 100 m, and latency is less than 1 microsecond.

LR: Do you see Reflex Photonics involved in Autonomous Vehicles (AVs) someday?

GP: Yes, AVs will require fiber-optics for security, bandwidth, latency, and SWaP. As the leading provider of rugged high bandwidth optical transceivers, Reflex is well positioned to deliver the most reliable optical interconnect for AVs. For large AV industries like commercial automotive one big challenge will be reducing the price of optical transceivers while keeping all the ruggedization testing in place. This will happen over a number of years, and so we will invest accordingly to track market prices.

LR:  When do you think AVs will start to get traction?

GP: When the technology is considered safe adoption will happen. This will require years of education and trials. One area of concern is cybersecurity—nobody wants a hacker taking over their vehicle at 60 miles per hour. An effective strategy will be needed to isolate critical control functions from infotainment. This separation is done in commercial aircraft and similar standards will be imposed on AVs. Fiber-optics provide the first level of defense since they are immune to electromagnetic interference and therefore harder to disrupt. As well, learning machines will be smart enough to initiate automatic protection from dangerous threats. Protection techniques commonly used by military aircraft could be deployed.

LR: How do you think the Autonomous Vehicle is going to play out, in reality?

GP: The benefits of autonomous vehicles have long been known, but safety has always been a barrier. The recent advances in AI and low-cost sensors has generated great hope for convenient, safe and cost-effective people transport. Like everyone, I see a gradual shift to AV starting with assisted driving available now to special lanes for AV followed by AV completely dominating the roads. I see China embracing this technology to solve local pollution issues while seizing the opportunity to lead the automotive industry.

1: Novet, Jordan. “Apple Packed an AI Chip into the IPhone X.” CNBC, CNBC, 12 Sept. 2017.
2: Chin, Josh. “Chinese Police Add Facial-Recognition Glasses to Surveillance Arsenal.” The Wall Street Journal, Dow Jones & Company, 7 Feb. 2018.

Sign up for Reflex Photonics' Newsletter

Sign up for Reflex Photonics' Newsletter

Reflex Photonics Newsletter Stay up to date and informed on new products and announcements with our newsletter.
The newsletter is published 6 times a year. (English only)

Your name (required)

Your Email (required)

Your Company

Sign me up for Reflex Photonics' Newsletter

By using this form you agree with the storage and handling of your data by this website. You may withdraw this consent at any time. Visit the Privacy Policy page to learn more.


VPX Technology Development Trends

John Koon, Contributing Editor

John Koon, Contributing Editor

John Koon was the Editor-in-Chief of the RTC Magazine and COTS Journal. As a researcher, he has presentd findings of technology trends such as Aviation, AI, autonomous driving, robotics and automation, low-power WAN, medical innovations, wireless technology including 5G and low-power WAN, Fog Computing (beyond cloud) and edge, IoT, NB-IoT, LoRaWAN, cybersecurity, blockchain, automotive including autonomous driving, automation, robotics, m2m, software, aerospace, manufacturing and COTS advancements.
His writings can be found in eBooks, blogs and technical trade journals. He held a BS in engineering (California State Polytechnic University, Pomona) and an MBA (San Diego State University).

Where various systems need to work cohesively, and system integrators are under constant pressure, VPX technologies respond to the needs of complex embedded systems.
Today’s military and other users of embedded systems have an endless appetite for greater integration, more autonomous operation, faster connectivity, and increased computing muscle. Especially on the military side, we are seeing new money and new technologies driving new strategies, platforms, and capabilities. That bodes well for our industry, because the electronic content in military platforms is rising dramatically.
Reasons that VPX is becoming the technology of choice for high-end systems designers include high performance per watt; VPX open-systems standards that lend themselves to interoperability and a broad technology ecosystem; and the inherent ruggedness of VPX, which enables it to stand up to high shock and vibration environments that demand conduction cooling and high signal integrity. User-definable and flexible, the VPX backplane offers a multitude of options, so it’s no wonder we’re seeing engineering companies using the platform in novel and unexpected ways.

Matching high reliability to rugged environment applications, VPX technologies are predicted to grow at a CAGR of more than 10% over the next five years. VPX is designed as a platform with a built-in feature that fully supports multiprocessing and provides multiple sources at the chip, board, and system level.
VME has long been used in mission-critical embedded systems. However, its VPX (VITA 46) progeny surpasses VME in bandwidth and power. But VPX designers soon found that VPX yielded its share of system-level interoperability issues: Multiple vendors’ boards and components were often incompatible within the larger system design, presenting a major obstacle to VPX system development. The answer: OpenVPX (VITA 65), a leading system architecture standard in field-deployed, mission-critical embedded systems. OpenVPX realizes a framework for system-level interoperability for COTS-based VPX Line Replaceable Units (LRUs). LRUs’ main plus is that they are Two-Level Maintenance (2LM) compatible.

Quick Return to Operation

The ultimate goal of two-level maintenance in all complex defense systems, including OpenVPX, is to enable a field technician without special maintenance tools to quickly identify a faulty board and swap it in a harsh and hostile environment, rapidly restoring the system to operation. OpenVPX’s proven field-deployed readiness and advanced systems design flexibility hasten the delivery of next-generation customizable COTS solutions to the front lines.
The need for speed has never been more critical. And VPX doesn’t disappoint at all. Since embedded defense systems must often be developed and deployed rapidly, it is very important that potential conflicts between the various profiles are eliminated as quickly as possible. VPX breaks out from the traditional connector scheme of VMEbus to merge the latest in connector and packaging technology with the latest in bus and serial fabric technology. It also combines best-in-class technologies to assure a very long technology cycle similar to that of the original VMEbus solutions. “VPX is a complicated technology that was purposely defined to be as open and expandable as possible to easily accommodate unknown future needs,” stated Jerry Gipper, VITA Executive Director. “The user community is taking advantage of this by defining OpenVPX profiles with supplier inputs to meet real-world application needs.”
Technologies called for in VPX include:

  • 3U and 6U formats
  • 7-row high speed connector rated up to 6.25 Gbps
  • Choice of high-speed serial fabrics
  • PMC and XMC mezzanines
  • RF and optical connectors, all in one big ecosystem

Here are examples of the latest VPX products showing the future trends.

OpenVPX Carrier Cards

One vendor offering 3U and 6U OpenVPX carrier cards is Acromag. The company uses surface mount technology and high precision pick-and-place equipment. Its carrier cards are for XMC , PMC and AcroPack I/O mezzanine modules. These carrier cards feature a high-speed PCI Express interface which meets the demands of high-performance industrial, defense, and scientific research systems. Choose from air-cooled or conduction-cooled models.
Rugged 6U carriers feature two PMC or XMC mezzanine slots with support for front or rear panel I/O and deliver 25W of power to each site. Connection to the OpenVPX board is via the Expansion plane or the Data plane. VPX REDI 3U models offer a single XMC slot and can be used as a VPX switch card, allowing a host CPU to communicate with up to three downstream cards in addition to the XMC card. Also offered are VPX bus carrier cards for AcroPack modules. These carriers support three AcroPack or mini-PCIe modules in any combination of I/O functions.

3U and 6U OpenVPX carrier cards for XMC , PMC and AcroPack I/O mezzanine modules (courtesy Acromag)

For easy integration with real-time software application programs, Acromag offers C libraries for VxWorks and other operating systems. The libraries provide generic routines (source code included) to handle reads, writes, interrupts, and other functions. Demonstration programs enable the developer to quickly exercise the I/O modules before attaching the routines to the application program. This diagnostic tool can save hours of troubleshooting and debugging (Figure 1).

3U and 6U OpenVPX carrier cards for XMC , PMC and AcroPack I/O mezzanine modules (courtesy Acromag)

6U VPX

Concurrent Technologies, a leading supplier of processor technology for demanding environments, has launched VR E7x/msd, a new 6U VPX computing board based upon the latest generation Intel® Xeon® processor E-2176M (formerly known as Coffee Lake). Supporting 50% more processor cores within the same power envelope, the Intel Xeon processor E-2176M has six cores compared to previous generation quad-core processors from the same product family.

In addition to improved processing capability, VR E7x/msd includes the option of front panel mounted USB 3.1 ports, 10 Gigabit Ethernet connectivity, enhanced storage options, and improved digital graphics outputs. Direct attached storage options include a SATA flash disk and two M.2 modules. These utilize PCI Express® connectivity and NVMe support for a high capacity solution that is suitable for use in challenging environments. VR E7x/msd is designed to fulfil a system management role for high performance 6U VPX processing solutions and so includes the option for dual XMC modules to support I/O expansion.

Figure 2: VR E7x/msd, 6U VPX Computing Board (courtesy Concurrent Technologies)

For security conscious customers, Concurrent Technologies offers several options: a TPM 2.0 device and Secure Boot are standard features with Sanitization Utilities and Guardian Security Package also available. Guardian enables customers to deploy sensitive applications through a range of hardware, firmware and software features that deter tampering and lock access to intellectual property.

VR E7x/msd, 6U VPX Computing Board (courtesy Concurrent Technologies)

Low-Profile Fiber Optic Transceivers

Low-profile LightABLE LM transceiver module

The demand for high-speed data networks drives the growth of fiber optics solutions. The main components to convert the digital data to light signals are fiber optic transceivers. Reflex Photonics offers a family of transceivers with different form factors and support speeds up to 300 Gbps. The LightABLE transceiver module measures 23 mm x 14 mm x mm (L x W x H) and has a low-profile form factor. It can operate in rugged environments with temperature ratings from -40 º C to 100 ºC and delivers reliable throughput with bit error rate (BER) of 10-15. Compliant  with the VITA standard and MIL- TD 883, the module consumes 1.4 W per module (12-lane) and can reach 300 meters over OM3 fiber cable.

Low-profile LightABLE LM transceiver module.

High Performing Backplanes

Elma Electronic provides VITA 46/65 VPX and OpenVPX backplane design and manufacturing. Newly released designs support VITA 66 and 67 optical and RF connectivity including VITA 67.3 RF/Optical modules. Instrumental in the development of the OpenVPX set of standards, Elma engineers developed the industry’s first VPX. Additionally, Elma offers the selection of 3U and 6U OpenVPX backplane profiles in slot counts from two to 12.

OpenVPX development a supporting C4ISR/EW modular open suite of standards (CMOSS) and the new sensor open systems  architecture (SOSA) profiles. (courtesy Elma Electronic)

As signal speeds and system complexities increase, Elma’s backplane designs provide support for VITA 66.4 optical, 67.1 RF and 67.3 RF/Optical modules with a host of designs supporting 1000BASE-BX, 1000BASE-KX, 10GBASE-KX4, 10GBASE-KR, or 40GBASE-KR4 connectivity. Features include precision timing protocol and network synchronization along with a radial clock slot.

Now available are OpenVPX development platforms supporting C4ISR/EW Modular Open Suite of Standards (CMOSS) and the new Sensor Open Systems  Architecture (SOSA) profiles. (courtesy Elma Electronic)

High-performance Layer 3 Switch

A high-performance Layer 3 switch (The ComEth4590) implementing two Ethernet switch matrices that control and separate physically the Control and Data Planes for highly secured VPX 3U systems has been introduced by Interface Concept. These two Ethernet packet processors, managed by the Dual core Arm processors and the Interface Concept Switchware, offer switching capabilities and multiple Giga, 10, and 40G Ethernet configurations.

Switch which can be used for 3U architecture (courtesy Interface Concept)

The ComEth4590a provides unrivalled performance and reliability in the most demanding static, mobile, shipboard, and airborne environments. Facilitating control and data plane switching are Ultra Thin Pipe ports routed on rear I/O (P2), operating in 10GBase-KR or 1000Base-KX modes – seven of them in all and a front optical port supporting 10GBase-SR or 1000Base-SX. The ComEth4590a also implements a microcontroller dedicated to the management plan (VITA 46.11). The ComEth4590a is available in air-cooled and conduction cooled versions.

A switch which can be used for 3U architecture (courtesy Interface Concept)

XMC for Signal Processing

Pentek launched its newest member of the Jade family of high-performance data converter XMC modules based on the Xilinx Kintex UltraScale FPGA. The Model 71800 is a co-processor module with an XMC PCI Express Gen 3 interface and general purpose I/O using parallel LVDS and gigabit serial ports.

The Jade Architecture embodies a streamlined approach to FPGA-based boards, simplifying the design to reduce power and cost, while still providing some of the highest performance FPGA resources available today. Designed to work with Pentek’s Navigator Design Suite of tools, the combination of Jade and Navigator offers users an efficient path to developing and deploying FPGA IP for data and signal processing.

High-performance data converter XMC module based on the Xilinx Kintex UltraScale FPGA. (courtesy Pentek).

The Model 71800 is pre-loaded with IP modules for DMA engines, a DDR4 memory controller, test signal and metadata generators, data packing and flow control to speed up the development process. The 71800 is available with the Kintex UltraScale KU035, KU060 and KU115, supporting a range of processing power. The majority of the Kintex UltraScale FPGA resources are available for customer installed IP for processing and management of I/O. The Model 71800 XMC module is designed to operate with a wide range of carrier boards in PCIe, 3U and 6U VPX, AMC, and 3U and 6U CompactPCI form factors, with versions for both commercial and rugged environments. They are designed for air-cooled, conduction-cooled and rugged operating environments, the Model 71800 XMC module with 5 GB of DDR4 SDRAM onwards. Additional FPGA options are available.

A high-performance data converter XMC module based on the Xilinx Kintex UltraScale FPGA. (courtesy Pentek).

Upcoming Challenges

Presently, everything is going gung-ho related to the VPX technologies supporting switched fabric, spreading network traffic across multiple physical links with higher throughput. Let’s take a moment to ponder over future challenges in the embedded systems world. As the number of embedded systems grows so do security concerns, with approaches to this issue varying somewhat depending on whose perspective we consider. Security should cover all aspects of embedded system design from architecture to implementation. As interoperability is key to embedded systems for a seamless experience, secured embedded design is of prime importance. Finally, applications are always going to need more bandwidth and improving system performance is always a challenge facing the VPX industry. Recognizing this, the VITA organization is working on improving performance with optical technology. How VPX technologies will grow in this context remains to be seen.

Improving interoperability in various electronic modules a common problem that faces anyone developing a complex embedded computing platform. Getting all of the modules to work together is a formidable integration challenge for system integrators. Components and modules need to connect easily and quickly into the system: plug-and-play, not plug-and-fight. Are design standards enough? Should something else be done to ensure compatibility and interoperability? Should the standards be strengthened or new standards added?

Security is a critical challenge for a wide range of embedded systems applications. System integrators must optimize operational capability, maximize competition for systems development, ensure interoperability, and maintain commonality to reduce life-cycle cost.

Overcome Challenges in Embedded Optical Interconnects Design

Gerald Persaud, V.P. Business Development
David Rolston, Ph.D., Chief Technology Officer

This article was also published by Electronic Design magazine. 

Introduction : Optical Technology in Embedded Design

The emergence of IoT in cloud computing and the demand for 4G and 5G networks worldwide are driving the increased use of optical transceivers in a wide variety of applications: business, government, industrial, academic, and cloud servers in public and private networks. Both local area networks (LAN) and wide area networks (WAN) are demanding more bandwidth packed into smaller spaces, and traditional copper interconnects cannot satisfy the insatiable appetite of all the network servers and gateway devices. Furthermore, the next generation of networking devices will need to be even more compact and faster. According to market research firm Radiant Insight, the optical transceiver market will reach $9.9 billion by 2020, three times its 2013 level.

Optical transceivers are the key components which transform electrical signals to light over optical cables. At the receiving end, another optical transceiver will convert the light back to electrical signals as shown in Figure 1. Most transceivers operate with speeds of 10, 40 and 100 Gbps. When higher speed is needed, multiple lanes are used in parallel to deliver the required bandwidth.

Demonstration of conversion of digital data and fiber optic light signals.

Figure 1: Demonstration of conversion of digital data and fiber optic light signals.

What are the Advantages Optical Networks?

Fiber optical networks or embedded communication systems have three key components: the optical transmitter, the fiber optical cable and the optical receiver. As described above, the transmitter converts electrical signals to light using either light-emitting diodes (LED) or laser diodes. At the receiving end, a photodetector is used to convert the light back to electrical signals. A transceiver combines the transmitter and receiver in one module.

The advantages of using fiber optics instead of copper include higher bandwidth, longer distance links, reduced weight, immunity to electromagnetic interference,and increased security.

Applications for Embedded Fiber Optics

Applications for embedded fiber optics are very broad and they often include projects that require very large bandwidth in confined areas, typically co-located with high-speed, high port count FPGAs or ASICs. Many of the civilian and military command and control monitoring systems (C4ISR), radar, FPGA interfaces, multiprocessor interconnects, CCD/CMOS imaging sensor arrays, high fidelity radar imagery, and systems requiring secure communications use fiber optics.

The block diagram of LightABLE LM module represents the typical functions of a transceiver

Figure 2: The block diagram of LightABLE™ LM SR4 module represents the typical functions of a transceiver

Challenges in Designing Embedded Systems with Optical Interconnects

There are three major challenges in designing embedded optical interconnects.

  • Creating an embedded system to support maximum bandwidth with the smallest possible footprint.
  • Designing or selecting the best optical interconnect able to perform in harsh environments with maximum MTBF.
  • Future-proofing the embedded system to maximize the return on system investment.

Balancing performance, space, power, and cost is a constant tradeoff and challenge. The design will depend largely on the bandwidth requirements of the embedded system, which can be a single boardcomputer (SBC) or multiple boards that fit in a chassis. Some of these network systems need to support many Gbps, thus the system design will also depend on the size and throughput of the interconnect. Multiple components may be needed which can increase the size of the SBC or chassis, and power and cooling requirements can also impact the overall system footprint. Many commercial grade optical interconnects operate in a limited temperate range, and cannot handle the shock and vibration of harsh environments, and some systems may need to operate in humid conditions. All these environmental factors will directly or indirectly affect the operational performance and life span of the products. In other words, a high-performance system will typically consume more power and require better cooling resulting in larger size and costing more. Therefore, tradeoff in system design is always an important consideration.

In Search of the Best Fiber Optic Interconnect Solutions

There are many considerations in selecting the best fiber optics interconnect solution, and they often require tradeoffs. The main considerations include investment protection, product performance, form factors, reliability, and integration considerations. The following provides guidelines in selecting or designing optimal embedded fiber optic interconnect solutions.

Standard-based solutions protect purchase investments

System performance is impacted by many factors, and there are a few rules of thumb to bear in mind:

  • The smallest form factor is desirable
  • Always choose rugged and reliable solutions even if they cost more
  • Guidelines for module integration and configuration should be observed

While there are many proprietary designs available, the best approach is to select VITA standard-based solutions because they are supported by a consortium along with a large ecosystem, so the design will be supported with upgrades over time. Additionally, the standard defines the technology and the connector specifications, which enables developers and integrators to select from multiple VITA-based vendors.

System performance is impacted by many factors

Fiber optics technology is capable of providing high-speed, low-latency, long distance communication with no electrical noise interference. However, many factors will impact the true, sustained performance and potential distance of the communication links. These factors include communication error rate, link budget, and receiver sensitivity. Additionally, when doing system design, it is important that full-duplex is part of the equation. Some new configurations can achieve up to 600 Gbps bandwidth, but there is overhead involved which may impact the actual throughput. For example, an unstable transceiver with high error rate will cause the system to perform retransmission which will decrease the overall system performance. The measure of this phenomenon is referred to as bit error rate (BER).

For fiber interconnect device or systems, the minimum BER should be 10-12. Higher performance can be achieved if BER is 10-15 or better. BER of 10-12 means that one error occurs every trillion transmissions. Additionally, a link budget greater than 13 dB with receiver sensitivity of -12dBm are recommended

600G LightCONEX LC plug-in module composed of two 24-lane transceiver side by side.

Figure 3: 600G LightCONEX plug-in module composed of two 24-lane transceiver side by side.

Smallest form factor is desirable

More and more embedded systems including single board computers (SBC) are using a smaller form factor. Therefore, it is important to select modules with the smallest footprint possible. Fiber optics transceiver modules can be as small as 1.3 cm × 1.3 cm (see Figure 4). Additionally, consider low profiles modules with a height of less than 5 mm. This will allow room for the SBC or systems to add additional functions on board.

150G and 300G LightCONEX Optical Transceivers

Figure 4: 150G and 300G LightCONEX Optical Transceivers

Always choose rugged and reliable solutions even when they cost more

Many fiber optics systems are used in harsh environments. As a result, commercial grade products will often fail or, at the minimum, have a shortened service life. It is important to choose solutions that will survive the environments of the target applications. Typical operating temperature should be from -40 °C to 100 °C or better with storage temperature from -57 °C to 125 °C. In addition, if the module consumes less power (100 mW per lane or better), it will create less heat and help achieve better MTBF. Other considerations should include shock and vibration resistance, passing MIL-STD-883 or better. The module should also be sealed to prevent corrosion due to exposure to moisture. As a rule of thumb, it is recommended that products pass the following tests to ensure the highest quality.

MIL-STD-883:

  • Vibration tests, Method 2007.3
  • Mechanical shock tests, Method 2002.4
  • Thermal shock tests, Method 1011.9
  • Thermal cycling tests, Method 1010.8

MIL-STD-202:

  • Damp heat tests, Method 103B MIL-STD-810:
  • Cold storage tests, Method 502.5

Rules for module integration and configuration should be observed

Illustration of a VPX blind mate connector comprising a 24 fiber MT ferrule and 10 RF coaxial connectors. The size of this connector meets VITA 66.4 standard.

Figure 5: Illustration of a VPX blind mate connector comprising a 24 fiber MT ferrule and 10 RF coaxial connectors. The size of this connector meets VITA 66.4 standard.

A well-designed transceiver should take into account the connector choice and location, as well as ease of system integration and configuration. As shown in Figure 5, an active blind mate optical design will make connector mating much easier and reduce the chances of making connection mistakes.

Additionally, there are two other rules to follow when considering connector selection or design. First, follow the VITA 66.5 “Optical Interconnect on VPX, Spring Loaded Contact on Backplane” standard which defines the connector dimensions. Then place the board-edge, plug-in module connector near the edge of the board, integrating an active parallel optic transceiver, and a spring-loaded backplane connector developed for VPX systems (part of the VITA standard) as shown in Figure 6. This approach will limit any additional cabling needed to bring the signals to the edge of the SBC board.

 LightCONEX Active Blind Mate VPX Optical Interconnect

Figure 6: LightCONEX Active Blind Mate VPX Optical Interconnect

The Embedded Interconnects Design Check List

The following paragraph can serve as a check list for ease-of-use.

  • Select solutions based on the VITA standard including VITA Section 6. This applies to the overall systems as well as the connectors.
  • For optimal system performance, select modules with low bit error rate (BER) such as 10-15.
  • Select small form factor modules in the range of 1.5 cm x 1.5 cm, with height profiles less than 5 mm. This will provide extra space for the circuit board.
  • Select rugged and reliable solutions that pass a minimum set of tests including the following:

MIL-STD-883:

  • Vibration tests, Method 2007.3
  • Mechanical shock tests, Method 2002.4
  • Thermal shock tests, Method 1011.9
  • Thermal cycling tests, Method 1010.8

MIL-STD-202:

  • Damp heat tests, Method 103B

MIL-STD-810:

  • Cold storage tests, Method 502.5

Select design/modules that are easy to integrate. These include blind-mate and broad-edge connectors.

Conclusions

The above article has outlined the advantages and challenges of using fiber optical interconnects. To use fiber optic cables, the electrical signals need to be converted to light signals using fiber optic transceivers. While there are many challenges to embedded fiber optics design, the benefits are substantial. The guidelines and design check list provided will help developers select the best solutions for their needs.

Optical Interconnect Design Challenges in Space

Guillaume Blanchette, Space Industry Manager, and
David Rolston, Ph.D., Chief Technology Officer, Reflex Photonics

Reprinted with permission from Aerospace & Defense Technology, September 2018.

Designers of fiber interconnect solutions have to consider space radiation attacks.

More and more aerospace applications are incorporating fiber optics technology into their designs due to its many advantages over copper. The thinner fiber solutions provide higher speed over a longer distance, are more reliable, offer higher noise immunity and, in many cases, lower the cost of ownership. Additionally, for the same diameter, fiber can pack more data than copper. Fiber is faster than the category 5 and 6 copper cables, approaching the speed of light (31% lower). For copper, pushing the speed beyond 1G is a challenge, but for fiber 10G is quite common. Copper is limited by distance. Usually, signal degradation with copper will occur after about 90 meters (2.7 km maximum for custom systems), while fiber can achieve more than 1.5 km without a problem and can deliver over 80 km depending on transmission signal quality.

Perhaps the most significant advantage of fiber is that it is not affected by electrical noise because the transmission uses light instead of electrical signals. The typical electromagnetic interference (EMI) that affects copper cables will not be encountered with fiber optics. Over time, the copper will also degrade and have worse signal-to-noise ratio
Compared with copper, a fiber system can be very efficient. In the case of a fiber-based Ethernet connection, more than 99.5% of the signal can be delivered to the Ethernet hub. Different types of convertors can be used to convert signals from the popular unshielded twisted pair (UTP) Ethernet connections over fiber cable, so many lower speed UTPs can be combined to achieve, for example, 100/120 Gigabits.

Challenges of Fiber Interconnect Design in Space

According to NASA, space radiation is made up of three kinds of radiation: particles trapped in the Earth’s magnetic field; particles shot into space during solar flares (solar particle events); and galactic cosmic rays, which are high-energy protons and heavy ions from outside our solar system (Figure 1). This adds up to ionizing radiation, proton and gamma ray attacks. These attacks have a major impact on electronic circuits, described as the total Ionizing Dose (TID) effects, which is measured in rad (radiation absorbed dose). Note that 1 rad = an absorbed energy of 0.01 J/kg of material, and 1 gray = 100 rads. The impact of exposure to space radiation ranges from degradation of performance to total malfunction. In space, one would imagine that the results can be quite serious.

The environment in space is harsh and demanding. Commercial-off-the-shelf (COTS) devices have to be able to endure the extreme temperature swings and the constant vibration. Failure is not an option in a space mission. Adding to this is the challenge to deliver maximum performance with minimum space, weight and power (SWaP), high mean-time-between-failure (MTBF), and reliability.

Designing for aeronautics is very different than designing for the earth environment. Aeronautical applications, such as spacecraft, satellites, and military aircraft are much more challenging. Designers of fiber interconnect solutions have to consider specific requirements to deal with those challenges. The three major challenges are:

  • Space radiation attacks
  • Operation in harsh environment
  • Achieving space, weight and power requirements (SWaP) and reliability
Spacecrafts experience constant attacks of space radiation from magnetic fields, solar flares and galactic cosmic rays.

Figure 1. Spacecrafts experience constant attacks of space radiation from magnetic fields, solar flares and galactic cosmic rays.

Best Practices for Optical Interconnect Design

paceABLE is a radiation-resistant optical transceiver created by Reflex Photonics. The modules measure less than 3 cm2 and weigh less than 15 g.

Figure 2. SpaceABLE SM is a radiation-resistant optical transceiver created by Reflex Photonics. The modules measure less than 3 cm2 and weigh less than 15 g.

Defend Against Radiation with Radiation-Resistant Design

What are the design considerations to meet the requirements as described above? It is important to defend against the radiation from ionizing, gamma, and other attacks. There are several methods to protect the device from radiation, including shielding, error correction, and using radiation-resistant components, which some refer to as radiation hardening. Shielding works for low-level radiation. Error correction works if the amount of radiation only temporarily impacts the device. However, heavy error correction will slow down the performance of the device.

Increasingly, more designs are incorporating radiation-resistant components to protect the device. Radiation-resistant silicon uses a different approach from the typical semiconductor wafers. The common approach is silicon on insulator (SOI) and silicon on sapphire (SOS), which enable radiation-resistant components to withstand an attack of ionizing radiation. While commercial-grade silicon can withstand between 50 and 100 gray (5 and 10 krad), radiation-resistant solutions can withstand 5 to 1000 times more depending on the types of components involved (Figure 2).

Design to Work in Harsh Environments and Follow Standardization

For the interconnect devices to survive in harsh environments, in addition to radiation resistance, they must include other parameters that may not be required for commercial-grade components. This includes meeting requirements for shock and vibration as specified in MIL-STD 883. It is strongly recommended that the devices be sealed from moisture and thermal shock within a wide range of operating temperature (typical -40°C to +100°C). Keep in mind that some devices may slow down when the temperature goes to the extreme, so it is important to measure sustained performance at those temperatures.

A different view of the SpaceABLE fiber-optic transceiver shows the connector for fiber-optic cable connection. At the bottom is the view of the ball grid array (BGA) for surface mount soldering.

Figure 3. A different view of the SpaceABLE SM fiber-optic transceiver shows the connector for fiber-optic cable connection. At the bottom is the view of the ball grid array (BGA) for surface mount soldering.

Designing or selecting open standard-based (VITA 66) interconnect devices ensures that the solutions will follow the lifespan of the standards and will not be easily obsoleted, as is often the case in proprietary or custom designs. To ensure that the devices meet minimum standards, they should meet – but are not limited to – the following industry standards:

  • MIL-STD-883, Method 2007.3 (vibration tests)
  • MIL-STD-883, Method 2002.4 (mechanical shock tests
  • MIL-STD-883, Method 1011.9 (thermal shock tests)
  • MIL-STD-202, Method 103B (damp heat tests)
  • MIL-STD-810, Method 502.5 (cold storage tests)
  • MIL-STD-883, Method 1010.8 (thermal cycling tests)
  • MIL-STD 883 (shock and vibration)
  • MIL-STD-883G, Method 1019.7 (total Ionizing Dose and Cobalt 60 gamma rays tests)
  • Total Non-Ionizing Dose (TNID) tests
  • Open VITA 66 standards
  • ECSS-Q-ST-60-15 Space Assurance

Achieving SWaP and Reliability

Weight becomes increasingly significant in space transportation and applications. The cost of sending 1 kg is estimated to be $50,000. Designing products to achieve optimal SWaP and high reliability with high MTBF is always the ultimate goal.

In space and military missions, failure cannot be tolerated. Satellites will be in orbit for many years, and repairing failed parts is not only difficult but also very costly. Therefore, designing for compact-size, ruggedness and high reliability will help developers stay competitive in the race to space. For example, the SpaceABLE interconnect solution with multiple lanes can yield as much as 150 Gbps. For reliability, a combination of sealing, ruggedness and radiation-resistant design plays into the longevity of the device. Its lifespan can range from a few years to over 20 years. The total cost of ownership including maintenance can be kept to a minimum with high-reliability devices.

Conclusions

Aeronautical applications face many design challenges that are unique to their intended environment. The best practices for optical interconnect design for space applications include the use of radiation-resistant technology to defend against space radiation, the use of components and devices that are designed to operate in harsh environments, and meeting SWaP and long-term reliability requirements. Finally, it is recommended to follow open standards like VPX and to look for solutions that comply with MIL and quality standards.

Optics all around: interview of Gerald Persaud

Amelia Dalton from EE Journal interviews Gerald Persaud, Reflex Photonics' V.P. Business Development.

This week we investigate embedded optical modules with Gerry Persaud from Reflex Photonics - The Light on Board Company. Gerry and I discuss the benefits of  their chip-sized embedded optical modules and why Reflex Photonics stands out in the optical module ecosystem. 

Gerald Persaud

Gerald Persaud is responsible for overseeing global marketing, business development and customer initiatives related to the Reflex Photonic's product lines, as well as managing product development and customer technical support.
Gerald has over 20 years experience in telecom and defense. Prior to joining Reflex Photonics he held senior management roles in engineering and business development at Nortel, General Dynamic Canada, and Celestica. Gerald has developed many leading products in optical communication, wireless and advanced computing. Gerald doubled revenues at start-up Coresim in one year and precipitated an acquisition by Celestica. He also won the largest design contract ever in Celestica for an OTN switch.
Gerald holds a B.S. in Electrical Engineering at McMaster University.

The Internet of space and radiation hardened transceivers

We are on the verge of a new era of human connectivity and communications – the Internet of Space (IoS) is upon us. The explosion of worldwide communications over the past 25 years has led to the pervasive use of mobile and land communications equipment with an abundance of platforms, applications and devices all driving the growth of many of the largest businesses in the world. There is no doubt that this trend will continue through the Internet of Things (IoT), along with improvements to the underlying network infrastructure. However, the next, ‘Small Step’ for man in terms of ubiquitous communications will be the ‘Giant Leap’ into the Internet of Space.

Internet of Space

The Internet of Space (IoS) is a long-term vision that leaders in some of the most technologically advanced companies in the world have begun to seriously consider. Both the European Space Agency and NASA have prepared plans that involve the deployment of networks of satellite around the Earth, Mars and the Sun. These networks are composed of complex communications networks for MIMO microwave antenna arrays and free-space line-of-sight laser links.
These technologies will be responsible for the communications of manned missions to Mars and will have to have the best in terms of redundancy, speed, and network management as most of what we send up, will never be fixed. Further to this however, will be the machine-learning A.I. systems on-board exploratory robots and landers for the moon and Mars including asteroid mining that will be tasked with resource extraction. For example, before the arrival of astronauts on Mars, dozens of intelligent, self-exploring robots and rovers will have to have found water on the planet for them. Basically, the self-driving cars of today will become the self-exploring robots of space.

New players in space exploration

OneWeb Satellites is planning the launch 900 satellites into low Earth orbit beginning in 2018, to deliver Internet access globally.

Coverage areas of each of OneWeb's planned satellites. (OneWeb)

The success of companies like SpaceX have shown that many of the traditionally held ideas about space exploration are breaking down and commercial opportunities are staring to be explored.
With more reliable, lower-consumption, smaller and more powerful computer systems, it is now possible to truly envision complex space networks.
OneWeb Satellites is planning the launch 900 satellites into low Earth orbit beginning in 2018, to deliver Internet access globally.

TeleSat, a Canadian company, has a plan to deploy almost 300 LEO satellites by 2021 to serve as an interconnect for continual 3G data networks for ship and aircraft connections over the oceans. However, as more commercial-off-the-shelf (COTS) parts are targeted for space applications as a mean to take advantage of powerful technologies at lower costs, a more meaningful business-case for vendors can now be made to support the space-vendor-ecosystem.

Space environment

However, there is a catch… The space environment itself is extremely severe; outside the protective cushion of the earth’s magnetosphere the exposure to radiation and extreme temperatures can destroy terrestrial electronics. Therefore, to really open these markets, the vendors will have to meet the space community half-way, and do what they can to “space-ify” their COTS products.

SpaceABLE radiation hardened optical transceivers

SpaceABLE is available in different configurations: 50G (4 TX plus 4 RX lane per device) and 150G (12 TX or 12 RX lane per device).

Reflex Photonics has a plan to do just that – with the cost of sending even just 1 kg into space at over $50 k, the advantages of using the Reflex Photonics’ line of small, lightweight, high-density SpaceABLE parallel optical transceiver modules inside the satellites will impact this enormously.
The Reflex family of SpaceABLE modules offer extremely high aggregate data rates (over 150 Gbps), the modules are less than 3 cm2 and weigh less than 5 g. They can be placed anywhere on a motherboard or linecard linking powerful CPU’s, GPU’s and FPGA’s across multiple boards and racks.

Rigorous testing

In terms of reliability, the SpaceABLE product follows the rigorous environmental testing of MIL-STD-883 with a variety of thermal shock, vibration, humidity and cycling tests included. Furthermore, Reflex Photonics has qualified these parts under very stringent radiation exposure tests: Active Heavy-Ion testing for latch-up, SEE and SET failures, long-term irradiation from PIF and NIF cyclotrons, and long-term exposure (over several weeks) of gamma-rays using Cobalt-60 on active parts. These tests were all done with reference to the ECSS-Q-ST-60-15 Space product assurance standard - Radiation hardness assurance - EEE components.
The space community is slowly evolving from an era of mega-projects and unlimited budgets to a dynamic industry that envisions a commercial market with volumes that can support multiple business. It can no longer afford the “nine-nines” of reliability – especially when private commercial enterprises like Virgin Galactic’s SpaceShipOne and XCOR’s Lynx space vehicles are rapidly closing in on tradition institutional domains. Reflex Photonics is part of this belief and this ultimate goal of Bringing space a little closer™ by offering optical transceivers and optical infrastructure that will enable the next generation of space exploration.

Reflex Photonics makes a splash at OFC

Reflex Photonics newly released optical transceivers for Embedded computing and Space captured much attention at OFC2018.

Visitors to Reflex’s booth were surprised by the small size of the LightCONEX active optical blind mate connector for rugged VPX computing and commented that this is the first solution to solve the many real-estate challenges for small rugged embedded computers.
As well, they thought the 300GBps bandwidth over 24 fiber optic channels offered great bandwidth and I/O density for next generation manned and unmanned vehicles.
The newly released SpaceABLE radiation hardened space transceivers from Reflex was also a big hit as engineers now have small chip sized optical transceivers that meets the needs for all future space equipment and high altitude aircraft.
Many visitors commented that Reflex is a leader in rugged optical transceivers and are happy we continue to focus on this market.

Demonstration of the LightCONEX blind mate optical interconnect at OFC
Demonstration of the LightCONEX blind mate optical interconnect at OFC
Sign up for Reflex Photonics' Newsletter

Sign up for Reflex Photonics' Newsletter

Reflex Photonics Newsletter Stay up to date and informed on new products and announcements with our newsletter.
The newsletter is published 6 times a year. (English only)

Your name (required)

Your Email (required)

Your Company

Sign me up for Reflex Photonics' Newsletter

By using this form you agree with the storage and handling of your data by this website. You may withdraw this consent at any time. Visit the Privacy Policy page to learn more.