As integrated circuits continue to shrink in size, they bring greater performance benefits, but also create more challenges for designers tasked with making these components resistant to the extreme radiation effects of space. Meanwhile, some integrators are trading in radiation-hardness to gain performance benefits in more benign environments.
BY John McHale
Designers of spacecraft—military and commercial satellites, unmanned exploration spacecraft, and manned spaceships—demand high reliability from the electronic components on their platforms.
Commercial and military users want components that can withstand the radiation environment and last for at least 20 years, says Anthony Jordan, director of standard products at Aeroflex Colorado Springs Inc., in Colorado Springs, Colo. “They want to keep those birds up there as long as possible.”
Designing components that must last as long as two decades in outer space without failing while being battered with all sorts of radiation is not an easy task.
“All satellite systems require shielding and protection from the naturally occurring space radiation environment,” says Jim Salzman, director of technology and radiation effects at the Texas Instruments high-reliability business unit in Dallas. “This includes commercial and military alike.”
Satellites thus have limited lifetimes due to radiation damage to components and as satellite semiconductor content and complexity grow, the task of radiation-hardening is becoming more challenging, Salzman continues.
Smaller feature sizes continue to challenge designers as they press to mitigate the effects of operating outside the atmosphere, Jordan says.
“With higher performing technologies come more radiation performance challenges,” says Vic Scuderi, manager of satellite electronics for BAE Systems in Manassas, Va. “Feature size continues to shrink with an accompanying increase in susceptibility to damaging radiation effects. To mitigate these risks, our technology roadmaps have continued to blend rad-hard by design with rad-hard by process. We believe this blend is critical to designing space hardware that includes advanced application-specific ICs (ASICs) manufactured at the 90, 65, or 45 nanometer technology nodes.”
|Curtiss-Wright Controls Embedded Computing’s SVME/DMV-184 general-purpose single-board computer with Freescale 8641 PowerPC processor was tested and evaluated using radiation mitigation methods at White Sands Missile Test Range (WSMR) in New Mexico.|
“As technology evolves and feature sizes shrink, rad-hard is becoming more and more critical to overall space systems success,” says Doug Patterson, vice president of the military and aerospace business sector at Aitech in Chatsworth, Calif. “As far as standards go, total dose is only one aspect—as systems get more integrated and demand more of technology, to ensure overall system reliability, we have to address a class of radiation pathologies termed single event effects or SEEs. These include single event upsets, single event latchup, single event transients, single-event burnout, single-event gate rupture, etc. Suffice it to say, making components rad-hard is as challenging as ever.”
Users of power ICs, such as those managed by International Rectifier in Calif., “are asking for higher radiation tolerance in terms of total ionizing dose (TID) and SEE when heritage data is not available, to improve the reliability and life of their systems,” says Odile Ronat, discrete marketing manager at International Rectifier. “Where 100 kilorad TID was sufficient, customers may now be asking for 300 kilorad. There are also requests for radiation levels beyond 1 megarad.”
Radiation and small components
Two different methods may be used deal with assuring radiation tolerance in relatively small components, say experts at TRAD in Labege, France.
“With regards to very high integrated technologies—field programmable gate array (FPGA) or ASIC)—the challenges are mostly at a design level,” says Christian Chatry, director at TRAD, which makes radiation analysis software, called FASTRAD. “Already, technological trade-offs can be proposed and manufacturers will, for example, use hardening through Silicon on Insulator (SOI) CMOS processes for SEL immunity, among others.
“For a given/frozen technology, electronic designers will try to obtain immune functions with a mitigation approach by using, for example, bit scrambling—to randomize particle impact on different memory areas—or appropriate error detection codes or even error correction codes—with an error management/code size mitigation approach,” Chatry continues. “A basic complementary approach will consist in using redundant functions to prevent particle induced malfunctions.”
With regards to micro or nano technologies or MEMS (microelectromechanical systems), “the challenge is at a different level,” Chatry says. “Hardening such technologies implies the investigation of new failure mechanisms and a new testing approach as compared to classical electronic components. Such small-scale systems are generally mechanical and electrical systems—i.e., micro switches, nano- valves—which require specific physics.”
Chatry says that radiation assurance also needs to become a more collaborative process as components become more complex.
“The radiation assurance process has become more and more a multi-faceted activity requiring a collaborative approach between engineers of different fields,” Chatry says. “The Radiation Hardness Assurance (RHA) plan used in the space industry runs from the space environment calculation up to the estimate of the system response. This implies a strong partnership between different teams—space and radiation environment, satellite platform design, particle-matter calculation, radiation modeling and calculation, electronics technologies, solid-state physics, radiation testing, electrical design, etc. The results of each group must be well understood by the others in order to ensure a coherent approach and to control the result uncertainties.”
|The Xilinx V-5QVFX 130 rad-hard FPGA is a QML-certified and can be reconfigured while in orbit.|
Satellite and spacecraft designers, much like their counterparts in avionics, vetronics, and shipboard systems, want improved processor performance, greater memories, and low power in their systems.
Users of BAE Systems’ RAD750 single-board computer are demanding “more processing power, larger memories, and EEPROM replacements,” which can make radiation hardening more complex, but they need the performance, Scuderi says. “We recently held a series of RAD750 Users Workshops at which we heard [that] loud and clear from our customers.”
They also “are looking for faster interfaces to get data into and out of the processor,” Scuderi continues. “Fortunately, this is the track we’ve been on and have products in development that address each of these needs. For example, we’re complementing the general-purpose processing capabilities of the RAD750 with a high-performance digital signal processor (DSP)”—the RADSPEED DSP, a radiation-hardened variant of the commercial CSX700 processor from ClearSpeed Technology in Oxford, England.
“This single instruction multiple data (SIMD) machine features 152 processing elements each with double precision floating-point units that are organized into two independent cores and achieves an estimated performance of 70 gigabit floating-point operations per second,” Scuderi says. “This processor is well matched to applications such as image processing, radar processing, and spectral analysis.”
When it comes to the payload, “it’s all about performance and power; i.e., image generation capability, target recognition and tracking, remote sensing, communication,” Aeroflex’s Jordan says. “Payload developers turn toward state-of-the-art commercial technology for many of these systems to satisfy mission or business requirements.”
Performance vs. rad-hard
To achieve performance goals, there are some tradeoffs happening, Jordan says. Payload developers trade traditional radiation-hardened components for commercial products upscreened for their specific applications and mission.
“By understanding how a commercial integrated circuit reacts to the space environment, a payload developer is able to mitigate those effects via a number of system design techniques,” Jordan says. “The performance requirements of many high-profile missions and programs have resulted in the use of commercial parts; the community understands the additional due diligence requirements to increase the probability of success. It is an interesting dilemma for traditional radiation-hardened suppliers.”
However, not all applications have the extreme radiation-hardening requirements of space programs.
Some radiation environments, such as nuclear events that military vehicles need to survive through, do not require the extreme radiation assurance necessary for outside the atmosphere, says Michael Slonosky, product marketing manager at Curtiss-Wright Controls Embedded Computing in Leesburg, Va. They require more performance and, where possible, will trade some rad-hardness for better-performing electronics, he says.
Curtiss-Wright has a line of products for these applications called Rad-Hard Ready COTS (commercial off-the-shelf), Slonosky says. The Rad-Hard Ready line does not look at being able to set aside rad-hard requirements, but is more about being rad-tolerant for applications where extreme rad-hard requirements are not necessary, he adds.
Curtiss-Wright tests its products at White Sands Missile Range, N.M., for gamma and neutron radiation, Slonosky says. The company’s rad-hard products are Curtiss-Wright COTS products that are modified so components are more rad-tolerant, he continues. For example, some NOR flash devices are more rad-tolerant than others, he adds.
|International Rectifier offers ceramic lid options on several of its surface-mount packages to mitigate X-ray threats.|
Some of these modified COTS products are flying in commercial human space flight applications where triple redundancy is built into the system to manage radiation assurance, Slonosky says. Using redundancy is less expensive than building radiation hardness directly into the card or component, he adds. This way, they can get the performance of COTS, procured in a faster manner than developing a full-space qualified part, Slonosky says.
“‘COTS’ is still an interesting word, as it is equally misunderstood in the mil and aero markets just as it is for space,” Aitech’s Patterson says. “Military and space missions don’t fall in the domain of commercial. For example, your cell phone doesn’t have to mitigate SEEs, etc. That’s not to say, however, that we can’t utilize relevant COTS technology and, in the end, that’s the real differentiator.
“At Aitech, we leverage the commercial sector’s development, and the billion-customer base testing of COTS technology to help guide us toward a viable solution for our military and space customers,” Patterson continues. “However, picking the chip is only the first step, that’s when the real work at the system-level just begins.”
Reconfiguring on orbit
Xilinx Inc. in San Jose, Calif., which offers commercial-grade FPGAs for space applications such as the ones used in Curtiss-Wright’s boards, is taking a different route with its new SRAM FPGA product—the V-5QVFX 130 FPGA—that is full QML-certified qualified and geared for extreme radiation environments.
“What has Xilinx excited though is that the FPGA can be reconfigured while in orbit,” says John Bendekovic, director of aerospace and defense sales at Xilinx in San Jose, Calif. This is the first viable, in-production rad-hard FPGA that has that capability, he adds.
Spacecraft that is in orbit for as long as 15 years do not have the ability to upgrade during that time, Bendekovic says. It enables satellite designers to no longer worry about anticipating changes and fixes to sensor interfaces, payloads, or processor structures, he says. “Now designers can make changes just by uploading code,” Bendekovic continues.
The device—based on the Xilinx Virtex V FPGA—is radiation-hardened by design, Bendekovic says. “It is unique because we went in and addressed the vulnerabilities through redesign.”
Aeroflex Colorado Springs is under contract to leverage prior government funding with the Boeing Solid State Electronics Division (SSED) in Kent, Wash., to bring a 90-nano- meter Radhard-by-Design (RHBD) library into the main stream, says Peter Milliken, director of semicustom products at Aeroflex Colorado Springs. The 90-nanometer library offers high-speed SerDes (3 1/8 gigabits per second) and high-performance embedded compiled memory (400 to 500 megahertz), and is capable of delivering ASICs with more than 30 million gates of logic.
International Rectifier in Redmond, Wash., is seeing requests for use of ceramic lids instead of metal lids in cases of X-ray threat, Ronat says. The company introduced the ceramic lid options on several of its surface-mount packages, including LCC-3 and SMD0.2, as options for applications susceptible to X-ray threat.
Crane Aerospace & Electronics is offering a new point of load converter—the Interpoint MFP Series. The MFP Series produces stable power across a wide temperature range of -70 to 150 degrees Celsius and delivers low noise performance, requiring no external capacitors.
Microelectronics officials of the U.S. Department of Defense (DOD) have certified Microsemi Corp. in Irvine, Calif., to build radiation-hardened microelectronics assemblies for space and mission-critical defense programs by granting the company MIL-PRF-38534 class H and K certifications. The MIL-PRF-38534 standard pertains to manufacturing quality based on statistical process control. Class H and K rad-hard certifications under MIL-PRF-38534 pertain to certain classes of resistance to the effects of radiation.
“Aitech currently offers the world’s first, space-rated, six-port Gigabit Ethernet managed switch, called the S750,” Patterson says. “The S750 implements and provides a built-in network interface card in a 3U CompactPCI PMC conduction-cooled form factor.”