COTS devices find their way into long-life, high-reliability military applications
Use of commercial off-the-shelf (COTS) components in military and aerospace applications is an accepted fact. The move encompasses not just individual components, but sometimes entire subassemblies based on COTS devices. As a result, military and aerospace grades have shrunk from 9% of the overall electronics market in 1984, to just 0.9% in 2005. COTS is here to stay.
The impact of this change is already evident, particularly in terms of obsolescence. Projects with development phases that last up to 20 years and in-service lifetimes of 40 years are being built using components with life cycles measured in months. The scale of the problem is illustrated by recent estimates suggesting that electronic components are now being obsoleted at a rate of over 13,000 per month.
The effects in terms of reliability, whilst equally important, are less clear at this point. That's simply because insufficient in-service data exists to assess real failure rates and lifetimes.
Against this background, it remains imperative that military and aerospace engineers ensure the reliability, safety integrity, and lifetime of their designs. In particular, whole-life component management planning is essential to ensure adequate supplies of fit-for-purpose components and subassemblies throughout the CADMID (concept, assessment, development, manufacture, in-service, disposal) cycle.
A whole-life component management plan has three overarching goals. The first is to ensure that COTS parts selected can satisfy the performance requirements of the military and aerospace environment. This might be in terms of rad-hardness, resistance to shock and vibration, or temperature range; and it means not just "survival," but also remaining fit-for-purpose under any conditions.
This front-end process impacts design, as much as procurement and testing. For example, programmable logic and memory are both following a trend of shrinking device geometry, bringing higher capacity and lower supply voltages as time goes by. The designer should be planning for this development to allow for technology insertion during subsequent design and production phases.
The second aim is to ensure continuity of component supply (and, in the end, security of production and maintenance). In practice this means implementing many of the obsolescence management best practices that have become established in recent years. Given the need to provide decades of product life with components whose life cycles run into months, this may include strategies such as advance procurement and long-term storage of components.
Finally, the plan needs to compensate for the lack of provenance of COTS parts. Unlike MIL SPEC components, COTS devices don't come with a certificate of conformity. There's no guarantee that, from batch to batch, components will be manufactured or assembled in the same facility, let alone with the same process. This requires a rigorous testing and inspection regime.
The whole-life plan commences at the concept phase with an evaluation of failure methods in current components, a process that extends well into the assessment stage. This establishes base criteria for the performance of available devices and can be structured into a formal quality criteria plan for the components.
The next stage is a top-down selection process, yielding a design parts list (DPL) that gives design teams a choice of components. Component engineering professionals must begin to monitor the health of supply of devices as soon as they're added to the DPL. This stricture may sound overly cautious. But, with military and aerospace design cycles routinely lasting from five to eight years, it's likely that a good proportion of the COTS components on the DPL will not just suffer supply problems—they'll become obsolete altogether, before production even begins.
Whatever happens, the design will inevitably need to accommodate technology insertions—package changes, speed enhancements, and so on—over a period of years. The substitute components chosen will need to be put through the evaluation and qualification process in the same way as any new addition to the DPL.
The lifetime procurement plan continues with the development of a suitable component qualification plan or specification. Its aim is to ensure compliance against performance and durability criteria of the specific application.
Traditionally, this is done with the help of a reliability handbook, the most widely used in the past being MIL-HBDB-217. But the publication itself is now obsolete, since component assembly techniques have moved on. Instead, engineers rely on a range of resources, including Def Stan 0041, MIL SPL, and NASA specifications.
However, undoubtedly the main (and most reliable) sources of data today are likely to be the manufacturer itself, or independent component test houses such as IGG, which conduct accelerated life testing to provide accurate statistics and qualitative information on failure rates and modes. It should be noted that for COTS components, the resulting data may well be batch-dependent.
Using COTS components for harsh environs and demanding applications almost inherently means that this qualification process is, in fact, an upscreening—testing lower-grade devices to elevated specifications. The standard procedure is for an initial visual and mechanical screening and dc electrical check, followed by a detailed constructional analysis, involving x-ray inspection, electron microscopy, C-scanning acoustic microscopy (C-SAM), and microsectioning (see Figure). There will also be a hermeticity test.
The accelerated life-testing process follows, with devices being monitored during the test for critical ac parameters to identify out-of-spec parts, as well as total functional failures. There may also be a radiation test or other tests specific to the devices' environment.
Rather soberingly, it's important to remember that the end result of this process is a qualified batch, not a "qualified component." After qualification, a lot should be pulled from the qualified batch and placed in long-term storage, with safeguards and watchdogs.
As the project progresses, acceptance testing of further batches of components is another a key part of the process. The degree of testing and inspection depends largely on the source of procurement. However, even parts from trusted sources need visual inspection and comparison with engineering drawings to verify markings, packaging, and overall condition against any traceability data that's available.
Overall, the specification, qualification, and procurement process is more complex than ever. At the same time, non-availability of MIL SPEC components shifts the burden of qualification and on-going acceptance away from the component manufacturers and toward the component purchaser.
Increasingly, manufacturers must look to specialist service providers for component engineering and obsolescence-management skills. As procurement authorities move whole-life reliability down the supply chain, the move to COTS will mean that handling component procurement will become central to a project's success.
Lloyd Francis is aerospace and defence Manager at IGG.