Budget cuts pressuring rad-hard designers to maintain quality while cutting costsStory
June 03, 2014
A lack of new programs and delayed funding for others has created a demand in military space circles to extend the life of platforms and look for ways to reduce funding in less critical systems by pursuing more commercial parts and manufacturing processes. Meanwhile, high-speed processors are creating thermal management challenges for radiation-hardened Integrated Circuit (IC) designers.
Budget constraints put in place by military leaders on Earth over the last few years are finally affecting the designs of satellites and other spacecraft at the component level. The long life cycles of space products typically lag behind terrestrial systems by about two to five years just due to the rigorous qualification requirements for electronics in the high radiation environments of space. The uncertainty about what will be funded in the long run is also making it difficult for suppliers to plan future development strategies.
While some programs such as NASA deep space missions and manned spacecraft platforms will still require the utmost in radiation-hardening of their electronic components, military leaders are crunching numbers to see where they may able to get by with less protection in less critical applications. This puts pressure on radiation-hardened (rad-hard) electronics suppliers to find ways to meet these cost reduction demands while maintaining the reliability of their parts for space missions.
“The overall market is down from the last couple of years primarily due to the poor financial situation in the U.S. and Europe,” says Larry Longden, Vice President and General Manager for Microelectronics at Maxwell Technologies in San Diego, CA. “This situation has led the U.S. to cut most major space programs and the associated new technology developments. This will have a long-term negative impact on the U.S. space market.”
“We are seeing related rad-hard budget pressures for military space satellites,” says Siobhan Dolan, Vice President of Business Development, Commercial Air/Space at Microsemi Corp. in Aliso Viejo, CA. (See Figure 1.) “They want longer life out of their existing designs as new funding for new programs is reduced or delayed. In the past, designs were typically required to last for 15 years, but now military program managers want to extend them as long as 18 years. That puts pressure on the component and system suppliers to guarantee technology can reach those expectations, particularly for military systems looking for an extension of legacy designs. The component manufacturers need to provide a higher level of radiation guarantee and prove it.”
“Sequestration is still in place and funding for many programs is being delayed,” says Doug Patterson, Vice President of Military & Aerospace Business Sector at Aitech Defense Systems in Chatsworth, CA. “Many of the big primes are buying components for prototyping designs via their internal research and development dollars. Classified programs are still steady, with many programs let years ago moving into production. Full production essentially means one or two satellites.”
NASA funding also is not expected to increase for some time. The President’s Fiscal Year 2015 total budget request for NASA – $17.460 billion – is essentially flat from last year’s request of $17.646 billion in 2014. It is not forecasted to exceed $18 billion until 2019 when NASA expects it to be $18.169 billion.
Engineers at Aeroflex Microelectronics Solutions also are seeing activity in the classified markets as well as in commercial telecommunications satellites and civil platforms, says Tony Jordan, Vice President of Product Marketing and Applications Engineering at Aeroflex Microelectronics Solutions in Colorado Springs, CO. “Classified programs are moving along with good activity in hosted payloads, where a box rides on the satellite as a secondary payload. The hottest area for us is still our standard products such as microprocessors, memory, serial bus products, voltage supervisors, power conversion, etc.
“Commercial weather needs seem to be heading down the host payload approach as well, leveraging commercial satellites as host for weather instruments that will deliver data for forecasting models,” he continues. “These instruments are built by commercial companies leveraging technology, developed by the government, to gather weather forecasting data. There are a couple initiatives in this area regarding instruments and hosted payloads.”
“While [Low Earth Orbit] LEO missions may have more latitude in being able to accept lower level radiation performance, they typically have shorter lifetimes,” Dolan says. “We see most of our high-tech military customers and Geosynchronous Earth Orbit (GEO) telecommunication missions require minimum 100 krad-type radiation performance guaranteed as mission life is typically getting longer.”
Figure 1: The GPS III satellite from Lockheed Martin makes uses of radiation-hardened electronics for its payload and control systems. Photo courtesy of Northrop Grumman, whose Astro Aerospace business unit supplies self-deploying, monopole JIB antennas for the satellite.
(Click graphic to zoom)
“A communications satellite being sent up for 10 to 15 years still needs full space-qualified parts, but for shorter life platforms even NASA is looking to do more with less,” Patterson says. NASA Jet Propulsion Laboratory (JPL) has been flying 30 krad Commercial Off-The-Shelf (COTS) boards in satellite experimental platforms for years, he adds.
Reducing costs while maintaining reliability
“The budget-constrained environment is forcing program managers to look closely at each program and try to be smarter about which really needs megarad or 100 krad protection and which ones might get by with, say, 30 krad protection such as unmanned platforms with shorter mission times,” Patterson continues. “This can result in huge savings.”
“I think there are two changes that are occurring in the market,” Longden says. “First, programs are faced with reduced budgets which has led to many customers reducing the quality level of products that they are procuring for space systems. Second, due to reduced program budgets and the need to perform more data processing in space there has been an increase in the number of programs requiring higher processing performance such as is available from our SCS750 Single Board Computer (SBC).”
“I think the most demand is for lower radiation [like VPT’s] SV series because of the cost differentials,” says Leonard Leslie, Manager of Space Programs at VPT in Blacksburg, VA. “When some customers really sit down and run the numbers they discover they don’t need higher radiation levels for total dose and SEE performance in certain applications. However, not every application or program falls in one camp or the other. For example, unlike general satellite applications, launch vehicles have no total dose requirements, but have a high single event requirement. Each application and mission is different.
“We still get orders for programs with high radiation requirements, but the majority fall typically in the 30 krad(Si) range. Our SV series devices are generally qualified to 30 krad(Si) and we do offer models that are not rad-tested for prototyping. For full up requirements we have the SVR families qualified to 100 krad(Si) and 85 MeV-cm2/mg SEE performance with displacement damage performance. Those parts are tested for high radiation environments and designed with more stringent worst-case analysis because in general they are used for longer missions in high-radiation environments.” (See Figure 2.)
Figure 2: VPT’s SVR series products are qualified to 100 krad(Si) and 85 MeV-cm2/mg SEE performance with displacement damage performance.
(Click graphic to zoom by 1.9x)
“Our customers are putting intense pressure on us to find innovative ways to deliver space-quality product at lower costs,” Patterson says. “We answer this by offering three different grades of product. For labs and for prototyping purposes we have a 100 percent software compatible product, followed by an intermediate degree conduction-cooled series of 3U CompactPCI qualified for extreme temperatures via upscreening methods. For those users who want all the bells and whistles and need a fully rad-hard product, our flight-qualified series is available at a much higher cost than the intermediate option. We offer the intermediate grade for customers who want to save money but still want reliability.”
Aitech’s next generation SP0 3U CompactPCI product now features a PowerPC 8548 processor. Radiation testing on the board has been completed and the SP0 is now qualified for 100 krad(Si) LEO, Medium Earth Orbit (MEO), and GEO. Testing has begun for deep space mega-rad environments, Patterson says. “The next generation of the company’s computer for space applications will have a serial VPX architecture with a processor that is yet to be determined. It will be a multicore processor, but there is still work to do as well as some preliminary testing. Patterson says he expects the product to be available sometime in 2015. The 3U form factor is the perfect size for the growing number of small satellite applications and their payloads, he adds.
For the next generation of Maxwell’s space SBC product “we are currently evaluating market needs for high-speed processing, high-speed data throughput, and standardization efforts to select the next generation of high-speed interfaces,” Longden says. (See Figure 3.) “We are currently developing a new model of our SCS750 that will include two full speed SpaceWire ports capable of simultaneous, bi-directional communication. In the component area we have just introduced new products” in Maxwell’s Rad-Pak technology: the 512Mb NOR Flash; a 16-bit, 200kSPS, analog-to-digital converter; a 16-bit, 30MSPS, digital-to-analog converter; and later this year will have prototypes ready for the company’s new 256Gb NAND Flash product. A flight-level NAND product will be available early in 2015.
Figure 3: Maxwell’s SCS750 single board computer will include two full speed SpaceWire ports capable of simultaneous, bi-directional communication.
(Click graphic to zoom by 1.9x)
Constrained economics combined with upcoming changes in political leadership and philosophy for investment in space technology also makes it hard to forecast where the opportunities will be beyond the next couple years.
“This market definitely is not real-time tied to the budgets because often the programs are funded well before any political or economic changes,” VPT’s Leslie says. “On the other hand, that also makes it hard to predict long term on whether it will move up or down. While we are more military focused at this time, we see a lot of potential in commercial spaceflight down the road.”
“There is a great deal of uncertainty from the political leadership – a new administration is only two years away – as to what the budget level and the resulting procurement path for space classified programs will be,” Dolan says. “The uncertainty of the customer about what they want makes it hard for industry to forecast beyond the next couple years. We have good visibility on system requirements for programs funded and launches scheduled in the next 1-2 years. After 2017 there is some uncertainty on how the military will go about procuring payloads. There is a lot of dialogue now about the potential to pursue a policy of disaggregation whereby the military may use hosted payloads on commercial systems for less critical aspects of the mission.
“Therefore, in the meantime, the military leadership wants their money to go further and we are doing different things to make that happen,” Dolan continues. “One is to help them reduce their total cost of ownership by increasing levels of radiation testing during the design process, which also helps them reduce risk in the long term. We have invested in ELDRS [Enhanced Low Dose Rate Sensitivity] testing capability on products such as our space grade BiPolar Small Signal transistors as well as our mixed signal integrated circuit devices to better indicate how they will perform in extended missions. In fact, we have redesigned some of our older technologies to significantly improve the radiation performance for ELDRS. At Microsemi we spend a considerable amount of money characterizing our standard products for these radiation effects. Without this available characterization data there would be no market for us. Specific orbits (LEO, MEO, and GEO) require different levels of data that must be supplied to the customer before they can design a part.
“As far as the military is concerned, risk reduction also goes hand in hand with flight heritage,” she explains. “They want the extended radiation testing of products and components that have flown before on similar platforms and similar missions. This helps them speed up their choice of technology, which in turn enables shorter development times. We continue to see increasing adoption of our RT FPGAs for the full range of space applications. Also we are seeing increasing adoption of our first flash-based FPGAs for radiation applications, RT ProASIC3. RT ProASIC3 now has flight heritage on several science missions, including the International Space Station and two NASA missions: IRIS (LEO) and LADEE (Lunar orbit). RT ProASIC3 also has flight heritage on several international LEO remote sensing missions, and will be deployed in commercial communications missions in the near future. We expect relatively easy adoption of RTG4, our next generation FPGA family for radiation environments. It uses a 65nm low-power flash process, which is immune to changes in configuration due to radiation effects.” (See Figure 4.)
Figure 4: The RT ProASIC3 from Microsemi has flight heritage on several science missions, including the NASA IRIS mission (pictured) for low Earth orbit. Photo of Orbital Sciences team engineers monitoring the connection of the payload fairing over the IRIS spacecraft courtesy of NASA.
(Click graphic to zoom by 1.9x)
While the U.S. market is shrinking, export controls and improved competition overseas make it difficult for U.S. companies looking to make up for those losses in the international market. The U.S. government is bringing reforms to the International Traffic in Arms Regulations (ITAR) to make it easier for U.S. companies to compete in the satellite industry, but some feel it may not be enough to overcome damage already done.
“Over the last couple of years Europe has taken a stronger position related to removing all ITAR products from their systems,” Maxwell Technologies’ Longden says. “This has led to directives to eliminate U.S. ITAR products even if it means reducing performance and quality. Finally, due to the critical situation with Russia, one of our largest growing space market segments is no longer available due to export restrictions. The Russian space market has been very quick to react and is currently looking for non-U.S. alternatives, similar to what Europe is doing.”
“Export controls have always been there and have gotten tighter over the years,” Patterson says. “The U.S. used to import data into the country, but not export data outside of the U.S. Now we are seeing other countries putting in similar export practices, making it twice as complicated to do business internationally.
“Export reform has been long in coming to a horrendously restrictive and slow process,” he adds. “Even after reforms that are quite welcome, we still need to do a lot of paperwork and still need export licenses.” (For more on ITAR reforms see Special Report on page 16.)
Cost savings through manufacturing
System integrators and suppliers can also meet the cost cutting demands of customers by being more efficient on the manufacturing end.
Engineers at Northrop Grumman in Redondo Beach, CA implemented commercial best practices to generate cost savings in fabricating 36,000 ICs for the U.S. Air Force’s fifth and sixth Advanced Extremely High Frequency (AEHF) satellites, enabling production to ramp up on a broad scale for both payloads. Each payload has about 18,000 high-frequency Monolithic Microwave Integrated Circuits (MMICs) for frequency conversion, amplification, and switching. They are integrated throughout the AEHF’s major subsystems that enable real-time mobile, global access such as secure crosslinks, anti-jam uplinks and downlinks, and super high gain Earth coverage antennas.
“The Air Force procured these advanced, high-frequency MMICs through block buys early in the payload development cycle. Along with cost and schedule savings, the parts were more efficient to produce,” says Stuart Linsky, Vice President of Communication Programs, Northrop Grumman Aerospace Systems.
“While Northrop Grumman adopts commercial practices where helpful, its primary emphasis is to take the best practices from the commercial world and adapt them to meet the unique requirements of hi-rel military,” says Tom Block, Director of Microelectronics Products and Services at Northrop Grumman.
“[We develop] advanced military IC capabilities to [provide] space systems performance that is not achievable with commercially available technologies. While this initial cost may be initially higher, these technologies can be used, in turn, to address commercial needs after they have been developed and proven for military applications,” he continues. “The most prominent example of this is Northrop Grumman’s development of GaAs [Gallium Arsenide] transistors for its space systems that were, in turn, used to develop cellular phone power amplifiers. This technology was key to reducing the power consumption of the cellular phone, enabling it to be operated with much smaller batteries and thus be physically much smaller.” (For more on the AEHF MMICs see sidebar on page 30.)
Enhanced performance and thermal issues
Advanced electronics bring more performance and speed but also create engineering challenges as Size, Weight, and Power (SWaP) requirements get reduced. “Military users want next generation systems and capabilities, whether in big major programs or smaller disaggregated systems,” Aeroflex’s Jordan says. “We are seeing a demand for more processing capability, higher density in memories, and bigger pipes for data communications. Typical requirements include multi-gigabits per second of data transmission, gigaflops of processing capability, low power, and dense, high-speed memories. The challenge is to fit it all into a very small space, remove the heat, and/or manage the power.
“Heat removal has always been a problem,” Jordan continues. “But now as we pack more logic and memory into a unit area we have to remove that heat. That’s the big hurdle right now for the industry to deal with, not just Aeroflex. A lot of heat is generated with multiple lanes of +10 gigabits per second SerDes and 10s of millions of ASIC gates that need heat removal or power management.”
“There is a major Aeroflex initiative along these lines and it calls for maximizing thermal heat dissipation using flip chip assembly technology and heat sinks in an FPGA pin-compatible footprint. It will meet MIL-PRF-38535 Rev K requirements for Class Y non-hermetic packages space applications,” says Jay Johnson, Product Marketing Engineer at Aeroflex Microelectronics Solutions. “Aeroflex is currently using this technology to develop a 1752 I/O flip chip ceramic CGA package – the UT1752FC – for implementing a high performance UT90nHBD 90nm custom ASIC with a 3.125 Gbps High Speed SerDes IP. The UT1752FC package scheme uses a heat sink attached to the backside of the flip chip-assembled die that enables efficient heat transfer from the package. Aeroflex is currently working towards Class Y qualification for its flip chip assembly technology, which will in turn allow for qualification of the UT1752FC package.” (See Figure 5.)
Figure 5: The UT1752FC Class Y package scheme from Aeroflex enables more efficient heat removal.
(Click graphic to zoom by 1.9x)
“Aeroflex also released a new Bus Switch family targeted for high-speed, bi-directional Mux/Demux and bus isolation applications,” says Michelle Mundie, Product Marketing Engineer at Aeroflex Microelectronics Solutions. “It will increase bus speeds and data throughput for digital and analog applications. The technology can also be used for power management, cold sparing, and redundancy applications. Commercial bus switches exist on the market, but this is the only rad-hard one available and works quite well where we isolate a microprocessor from very large memory arrays, which have a tendency to load the microprocessor or FPGA down.”
(Click graphic to zoom)
Rad-hard power ICs
Reduced SWaP demands place pressure on all electronic component designers, but especially those who make power devices.
“There is still a great deal of demand for distributed architectures, which means more demand for Point Of Load (POL) regulators with high efficiency at low voltage,” says VPT’s Leslie. “We are also seeing a lot of big requests for low power converters at five watts and below. We receive a significant number of requests for low power converters for sensors and other analog and digital circuitry. Right now we are offering our SVSA 5 watt family and the SVHF and SVRHF 15 watt families for isolated converters to meet these demands.
“Processing power is going up, and as a result voltage requirements are going down. There is a big drive for high current at a low voltage,” he continues. “The challenge is providing a small package while keeping efficiency high at low voltage and high current. VPT offers the SVGA and SVRGA families of non-isolated point of load DC-DC converters to cover applications at low voltage and high current and high efficiency.”
(Click graphic to zoom)