DESIGNERS OF TOMORROW'S fighters are already wrestling with an unusual problem - obsolescence. Not of the aircraft as weapon systems, but of key components, principally in the avionics. The problem is being made worse by the protracted development and production timescales caused by reduced defence budgets, and by the decline in the military's importance as a market because of its reduced spending power.
"Every platform fielded today has a significant problem with parts obsolescence, and it is getting more pronounced as aircraft are getting older and there is no longer the same level of industrial base," says Reg Varga, director of open systems at McDonnell Douglas Aerospace (MDC). The use of commercial "open-systems" hardware and software in military avionics is seen as the key to tackling the growing problem of parts obsolescence.
Introducing commercial products into the existing combat-aircraft fleet is one of five areas being addressed by US Air Force researchers under an "ageing-avionics" effort, says John Ostgaard, branch chief at Wright Laboratories' Avionics Directorate at Wright-Patterson AFB, Ohio. The others address today's high software-development and aircraft-modification costs; the high maintenance, low reliability and poor diagnostic capability of in-service avionics; and the design problems which result from using rapidly changing commercial technology.
In parallel with its research effort, the US Air Force is working with industry on a prototyping programme intended to transition technology to the fleet within three to five years, says Ostgaard. Examples include a Lockheed Martin project to develop commercial replacements for out-of-production electronics cards in the F-16's stores-management system.
The Air Force is not alone in its concerns about ageing avionics. MDC's Varga believes the US Marine Corps is "leading the pack" with its Open Systems Commercial Avionics Requirements (OSCAR) programme to replace the mission computer in the MDC AV-8B with a commercial processor. The US Defense Advanced Research Projects Agency, meanwhile, has launched a pilot programme to introduce commercial products into today's weapon systems, and has invited proposals from the services and industry on possible lead applications.
Although the US Department of Defense has a joint task-force pursuing the use of open-systems technology in weapon systems, both Ostgaard and Varga say that there is no overall initiative to tackle the problem of ageing avionics. Instead, they expect commercial technology to be adopted within product lines, with manufacturers applying their particular choice of the open-system standards across their range of aircraft. "Working across platforms, companies and services is pretty tough," argues Varga.
The problem, as always these days, is lack of funds. The services have to address the issue of ageing avionics, because "-maintenance costs are eating into the modernisation budget", says Ostgaard, but they do not have the funds to tackle the problem across the fleet. Instead, says Varga, they have to deal with the problem piecemeal, comparing repair costs with replacement prices to build a business case for introducing commercial technology, product by product.
"The problem is not an easy one. The main issue is the dollars available in the acquisition budget to support the change [to commercial technology]. Industry is trying to offer options to make it easier," says Varga. "MDC's thrust is to develop an [avionics] architecture for the future, and to get early elements [of that architecture] into existing products," he explains.
EMPHASIS ON AGEING
Ostgaard describes ageing avionics as an "emphasis area" within the USAF's science and technology programme. Areas being addressed include software costs. "The Air Force investment in its current generation of software is more than $3 billion, and we upgrade major weapon-systems every seven years on average, so we face the cost of upgrading $3 billion-worth of software every seven years," he says.
A major contributor to the cost of developing software is the exhaustive testing required. This could be automated, Ostgaard suggests. "The cost per line of code over 20 years is about $400-450. We could halve that cost if we can address testing through emulation and automation," he says, as well as validation of the software during development, rather than after, as occurs now.
Software costs can also be reduced by the re-use of code, made possible by using commercial programming languages which are independent of the processing hardware used. MDC used this approach in its 1996 Common Operational Flight Program demonstration. Under this project, the same navigation-software module was flight-tested in three different aircraft using two different commercial processors. The first flight was in an AV-8B, using a Motorola PowerPC processor; the second was in an MDC F-15, using a MIPS R4400 processor; and the third was in an MDC F-18, again using the PowerPC. The software module behaved identically in each case.
MDC was subsequently appointed prime contractor for the Marine Corps' OSCAR programme, under which subcontractor CDI is developing a replacement for the AV-8B's AYK-14 mission computer using the PowerPC commercial processor. Varga says that development units will be delivered for laboratory testing in early March, with production to start by year-end and upgrading of AV-8Bs to begin early in 1998. The commercial processor offers "-a two orders of magnitude increase in performance [over the AYK-14]". The AYK-14 is also used in the F-18, and the Navy is in the early stages of looking for a commercial replacement.
The high cost of modifying aircraft wiring and cooling provisions when upgrading avionics is another area being addressed under the banner of ageing avionics, says Ostgaard. "Rewiring costs can be 25% or more of modification costs," he says. Solutions being evaluated include installing a "bridging network" to connect the military-standard 1553 databuses in today's aircraft with a commercial network which would allow use of open-system interfaces. "This would work with larger aircraft, but there is not much room in smaller fighters," Ostgaard says, so another solution being considered is to run high-speed data protocols over existing wiring. "1553 works fine, but we need more performance," he says.
Cooling is also an issue, as high-performance commercial processors generate substantially more heat than those used in today's military avionics. "We are looking at ways to provide distributed cooling, so that we don't have to rework the entire cooling system," says Ostgaard. Power is also being addressed, as the latest commercial electronics require a 3V electrical supply, rather than the 5V now used in avionics.
Maintenance costs for ageing avionics are "in a spiral", says Ostgaard. While commercial technology offers a solution, improved avionics packaging will be required "-to prevent COTS[commercial off-the-shelf] becoming a four-letter word," he says. Commercial electronics will have to be protected from the harsh environment inside military aircraft if high reliability and low maintenance is to be achieved.
According to Ostgaard, improved packaging, diagnostics and embedded fault-tolerance will be required if the military-avionics support system is to move, as planned, to the concept of discarding, rather than repairing, faulty line- replaceable modules. "If you go to throw-away modules, you have to have high confidence that you have found the bad module," he says. Current built-in test routines can isolate failures to "ambiguity groups" of two or three modules; "We have to get to one for throw-away," he says.
Developers of the Lockheed Martin/Boeing F-22 are already dealing with the issue of parts obsolescence, although the aircraft is not scheduled to enter service with the US Air Force until 2004. The team has compiled a catalogue of more than 400 components which are likely to be out of production by the time the F-22 becomes operational. This database is used to determine how to deal with obsolescence, ranging from buying the entire production run of parts at the start of the programme to targeting when to insert new technology based on when parts go out of production.
The team has developed a "roadmap" for improvement of the F-22's Hughes-produced common integrated processor (CIP), which addresses the need to introduce commercial technology to replace out-of-production parts and to tap into the rapid advances in commercial electronics which will inevitably occur over the planned production run of 438 aircraft. The first step, designated CIP 2000, will increase memory and reliability while reducing weight and cost. The current i960 processor will be retained, says Marty Broadwell, avionics integrated product-team manager, but an open-system databus and 3V power-supply will be introduced to allow the later use of commercial processors.
The CIP 2000 is scheduled to be available by the end of the F-22 development programme and to be installed in the first aircraft to enter service, replacing the current processor. The processor is expected to have a five-year useful life and to be replaced by the CIP 2005. Broadwell says that the intention is for this latter upgrade to be achieved by simply replacing plug-in modules in the F-22's CIP racks. He says that the plan is to make the hardware change transparent to the software.
Ostgaard says that an acquisition strategy has to be put in place to deal with parts obsolescence and the introduction of commercial technology. "We have to be aware that commercial products change constantly, often without us knowing," he says, and that there could therefore be functional changes of which both manufacturer and customer are unaware. "If we move to COTS, products will not stay the same. Processor chips change every 18 months and military systems last for decades. We need a strategy on how to drop in new technology at low cost," he says.
The military used to be a major market for the electronics industry, but now it accounts for only around 1% of sales because of declining defence budgets and booming commercial markets. As a result, the military can no longer dictate the pace and direction of development, particularly in digital devices, but increasingly in communications and electro-optics technology. The advantage for the military, if it can tap into the commercial "fast-track", is that it can then concentrate its diminished resources on those areas of technology where are unlikely to benefit much from commercial development, such as radar and electronic warfare.
The military will have to be isolated from the adverse effects of rapidly changing commercial technology. "We are looking at [avionics] maintenance and configuration-control by the manufacturer, rather than the military," says Ostgaard. "We would go to a strategy where the manufacturer can modify a system when it makes sense economically. As long as the user gets the performance, he won't care what's in the box," he says. This is the philosophy behind Lockheed Martin/Boeing's plan to provide contractor logistic-support of F-22, with the team offering to maintain the aircraft for a fixed cost per flying hour based on its vendors undertaking to repair, upgrade, or replace modules to maintain spares stocks at agreed levels.
Preparing for JSF
Looking even further ahead, Boeing and Lockheed Martin have been tasked with developing processes to deal with parts obsolescence over the 20-year production life of the Joint Strike Fighter (JSF), which is not scheduled to enter service until 2008 at the earliest. The two companies are developing JSF designs under competing concept-demonstration contracts. "We have to design in the ability to deal with obsolescence throughout the life of the JSF," says Dennis West, senior avionics technology-development specialist at Lockheed Martin Tactical Aircraft Systems.
Lockheed Martin plans to demonstrate its process for implementing an open-systems avionics architecture on the current F-16, at the same time addressing problems with out-of-production parts which now afflict the current aircraft. According to West, the process involves using a hardware-design language to describe exactly the form, fit and function of out-of-production cards in the F-16's stores-management system (SMS), which has been chosen for the demonstration.
Lockheed Martin subsidiary Sanders is developing a software compiler which will program the obsolete card's function into a field-programmable gate array. The original and replacement cards will then be bench-tested to compare how they react.
West says that the replacement card will be demonstrated in eight months' time, and that the company is looking for funds to redo the entire SMS. "Wright Laboratories wants money to go ahead and tackle obsolescence. Lockheed Martin has a vested interest because it has more aircraft out there than anybody else," he says, noting that the company's F-16s and C-130s account for about two-thirds of the USAF's aircraft, and that the ideas to be demonstrated on the F-16 "-are consistent with the C-130".
Ostgaard says that the Avionics Directorate is seeking to double its research spending on ageing avionics from the current $12 million a year and plans to begin the five-year prototyping programme in 1999, although industry is pushing for this to be brought forward.
The USAF plans to spend $40 million on prototyping, with industry also providing investment, and the service's avionics-upgrade plans are likely to influence the projects selected for this phase. "It's not a matter of if we will use commercial technology, it's a question of how to use it smartly," says Ostgaard.