By J.R. Wilson
The U.S. Air Force's new F-22 Raptor advanced tactical fighter is finally preparing to move into production after more than a decade of development. In the process its avionics architecture has passed through at least three cycles of obsolescence and relies on an Intel microprocessor — the i960MX — that went out of production four years ago.
There is nothing unusual about that, of course; the time frame for developing, testing, and producing a new military aircraft far exceeds the confines of Moore's Law, as well as the confines of commercial development, production, and eventual obsolescence of microprocessors.
For the F-22, an upgrade to a new PowerPC processor already is on the drawing board, beginning with Lot 5 production of the aircraft around 2004. However, the aircraft's builders must freeze the jet's design baseline at the end of this year, which means a fifth generation (G-5) chip, at best, which will undoubtedly have long passed its own moment of obsolescence by 2004.
The avionics suite is the responsibility of engineers at the Boeing Military Aircraft & Missiles Systems in Seattle, with Mike Harris as avionics production team leader. The overall avionics production team manager is Tom McDermott of Lockheed Martin Aeronautics in Marietta, Ga. The two companies team with Pratt & Whitney in East Hartford, Conn., to produce the F-22, which will replace the aging F-15 as the Air Force's primary air-superiority jet fighter.
The F-22's integrated avionics suite features extensive use of very high-speed integrated circuit (VHSIC) technology, common modules, and high-speed data buses to manage its sensors. This approach frees the pilot to concentrate on his mission.
The technologies applied to that task include a Common Integrated Processor (CIP), the system's "brain", which has been described as having the equivalent power of two Cray supercomputers; shared low-observable antennas; Ada software; expert systems; advanced data fusion cockpit displays; integrated electronic warfare system (INEWS) technology; integrated communications, navigation and identification (CNI) avionics technology; and optical fiber data transmission.
Along with the U.S. Army's RAH-66 Comanche scout-attack helicopter and the still-in-development Joint Strike Fighter (JSF), the F-22 represents the most advanced electronic technology in military aviation.
Originally, it was to have shared much of its avionics technology with the Comanche under standards established in the late 1980s by the Joint Integrated Avionics Working Group (JIAWG). That included a standard backplane interface, a test and maintenance interface, and two external fiber optic serial databuses — a high-speed databus and a sensor data distribution network.
"All of those initially were common with the Comanche, but the helicopter has since moved on," McDermott says. "We're using that baseline today. We intend to carry that baseline out through at least production Lot 4, through about 2004," which represents about 55 aircraft out of 339 currently approved by Congress.
"Right now the F-22 program has a funding cap for both development and production, so being stable is good," McDermott says. In the future, a combination of DMS [diminishing manufacturing sources]) and trying to be more compatible with other programs will bring us to a more commercial architecture. What will really do that is if we get a new derivative of the aircraft in the program, such as an air-to-ground version. Right now we are basically an A-model throughout the entire production run and funding is capped. But if there is a new version, as with the F-15E, that probably would be a big driver to bring up the COTS content."
Lot 1 production is to begin this year and run through 2003, using the avionics baseline developed in 1991. To keep that configuration stable for the past 10 years, program managers lined up parts early on. They made several "bridge" buys to procure enough components for production and spares — including the central microprocessor — until the next logical redesign point for each subsystem.
In addition to the Intel i960MX-based multiprocessor (a cluster of 35 processors), the suite uses an F-22-unique signal processor from Raytheon derived from the radar processor on the F-15.
"That signal processing element does around 200 million operations per second," McDermott says. "The i960 is about 30 million operations per second. What we get in the end is about one billion instructions per second for data processing and about 3 billion per second for signal processing. It was considered to have plenty of margin when we set that baseline in 1991 and for what we have on the airplane today, it has held up. It's a modular system, so we can increase capacity by adding more modules to the box. But right now we haven't identified any functions to put on the airplane that it can't handle."
When the time comes, designers say they expect to replace the signal processor with a PowerPC using AltiVec technology, Motorola's high-performance vector parallel processing expansion to the PowerPC RISC processor architecture. AltiVec adds a 128-bit vector execution unit operating in concert with the PowerPC's existing integer and floating point units to provide highly parallel operations, as many as 16 simultaneously in one clock cycle.
"It really depends on what the overall change plans for the aircraft are and what funding is available, but eventually, I imagine we will go that way," McDermott says.
Optical fiber interconnects
Most of the aircraft's box-to-box interfaces are fiber optic. "From the sensors to the common integrated processors is a point-to-point integrated fiber optic link, very wide band, which is quite fast and easily takes care of any bandwidth problems we might have for those sensors, both in terms of speed and range," says Boeing's Harris. "That was designed up front quite well with plenty of bandwidth. In terms of studying what future F-22 avionics will look like, we are looking into the RapidIO and InfiniBand standards, but we really don't have a requirement right now to go that way.
"The air-to-air mission requires some throughputs and speed within the processors and signal processors to handle the sensors, but not much different from the requirements for the F-15E, except it is a lot newer and a lot easier to handle the requirement," Harris continues. "The same is true for the Super Hornet."
Despite the advance technology aboard the aircraft, future growth is still an important consideration. While the F-22 already has two common integrated processors, it also was designed with space for a third, which has not been used yet. "I'm sure that extra space was put in to handle future derivative missions beyond the air-to-air requirements," Harris says.
While the Comanche helicopter has diverged from the original goal of commonality, one area in which the F-22 and other aircraft will share technology is flat-panel displays, which are from Kaiser Electronics in San Jose, Calif., now a part of Rockwell Collins.
"If there is any backward compatibility or commonality between airplanes, it would be in that area," says Harris. "The (F/A18-E/F) Super Hornet also uses the Kaiser flat-panel displays; they are different sizes, but the technology and most of the circuitry is the same."
The F-22 has an 8-by-8-inch display and a couple of smaller 6-by-6-inch displays; the Super Hornet has a couple of 6 by 6s and an 8 by 8 or 8 by 6 planned for the future.
"The technology has grown now such that the size of the display is irrelevant," Harris explains. "You get a lot of the efficiencies in having the same kind of display without forcing all cockpits to use the same size. Ten years ago the simple solution was to gain efficiencies by making all displays the same size, but that's not necessarily desirable today. Flat panels were expensive and hard to make and customized back then. That is no longer the case. That's a big success story."
The new generation of flat panel displays also is lighter and more durable, and better able to handle vibration and other environmental issues than earlier versions, Harris says.
"We don't have any problems on the Hornet or the F-22 in terms of temperature ranges and such. If the pilot is comfortable, the displays will work fine," he says. "That isn't saying there's no risk to large displays in military cockpits, but no more so than other things in the cockpit.
"The newer technology uses projection displays rather than liquid crystal, which takes care of any brightness issues you might ever have," Harris says. "You just turn up the power to increase brightness. So for military applications, I think that [sunlight readability] is solved."
The F-22 also includes some advanced design work to support maintenance by going one step beyond the modular line replaceable unit (LRU) design that set the F/A-18 apart from its predecessors.
"On the F-22, you can open a panel and replace the cards, which have less function attached to them, so you can sometimes replace one card with an adjacent card," Harris explains. "That makes the pipeline for spares less broad for that airplane. So as far as maintainability is concerned, that is a significant jump forward. And the lifecycle maintenance cost of F-22 avionics is much better because you don't have to replace the box, just the card."
Although the power requirements are basically 1990s technology, when the avionics system baseline was frozen, the 17 different power supply types incorporated in the avionics suite are fewer than necessary on other airplanes.
In January, flight testing for the F-22's Block 3.0 software components began. Block 3.0 — developed by 11 major subsystem suppliers — provides functions such as radar processing and sensor fusion, electronic warfare and countermeasures, communication, navigation, and identification and the pilot/vehicle interface. Two more upgrades are planned, but 80 percent of all functions that will be on the production airplane are now flying.
The pilot/vehicle interface and much of the operating system are either written or managed by Lockheed-Martin Marietta; the electronics warfare and CNI are managed or written by Lockheed Fort Worth and the sensor fusion and mission software are written at Boeing; the radar software is written by Northrop Grumman in Baltimore, and managed by Boeing in Seattle. Harris Corp. in Melbourne, Fla., is building the fiber optics data transmission and electronics interfaces, which McDermott manages. The integrated electronic warfare system (INEWS) technology comes from British Aerospace and managed out of Lockheed Fort Worth.
"That has always been a difficult technical problem and the F-22 has gone a long way toward solving those, but it is still a challenge technically to integrate pieces of electronic warfare that have always been separate and designed by different people into a single fused sensor on a small airplane, especially with apertures that normally counter the stealth requirements of the F-22," Harris says. "The CNI avionics technology has similar problems. The basic communications elements [developed by TRW and managed by Lockheed Fort Worth] are working fine, but again it is a technical challenge to integrate those."
A primary program goal of affordability also has met with what Harris terms "some huge success stories."
"Whenever there is an opportunity for redesign, improving quality or cost, we do that," Harris notes. "One such is the Northrop Grumman radar system. In order to build the stationary radar array, they have lots of transmit/receive modules fastened to a large backplane. Northrop Grumman has done a great job of using robotics in that production and saved a lot of cost. We're claiming something like a $300 million cost reduction just in the radar over the 339 airplane buy."
Prior to installation on the F-22 test aircraft, the systems were first integrated and tested at Boeing's Avionics Integration Lab (AIL) in Seattle and on the Flying Test Bed (FTB), a specially equipped Boeing 757. They are then shipped to Lockheed Marietta, where the aircraft are assembled.
Getting a Raptor into the air with combat-capable avionics was a major requirement that designers needed to meet before low-rate initial production (LRIP) could begin. Air Force Brig. Gen. Jay Jabour, the F-22 system program director, said flying Block 3.0 was "the program's current most technically demanding challenge."
The F-22 avionics suite represented by Block 3.0 provides and controls the aircraft's "first look, first shot, first kill" warfighting capability to accurately acquire, track, identify and engage multiple targets. It also enables the Raptor to automatically detect and defeat incoming missiles by initiating counter-measures.
"Block 3 is the first sensor fusion software, fusing all the sensors on the airplane to give the pilot the most information long before the bad guy sees him," Harris says. "That is the heart of F-22 avionics."
The Raptor program is managed by the F-22 System Program Office at the Air Force Aeronautical Systems Center (Wright-Patterson AFB, Ohio). A decision on LRIP is expected later this year, following an April Defense Acquisition Board (DAB) chaired by the U.S. Department Defense undersecretary for acquisition and technology.