No Flight of Fancy

Fort Worth, TX-During a flight at 35,000 ft, the pilot of an F-16 puts his aircraft into a high angle of attack, greatly reducing the flow of air over the wings-a stall. Without that critical airflow, the pilot loses control of the plane and it drops into a flat spin, falling to earth like a rock. The pilot works quickly to adjust the plane's control surfaces and force the nose down. With the plane now falling nose first, air will again flow over the wings and the pilot will gain control of the F-16. Time to panic? Not quite: it's all part of a carefully scripted and monitored test flight to put the aircraft through its paces.

When people think of test flights, they may imagine a pilot taking off just to see how a plane flies, but there's much more to testing a plane. So says Darrel Russell, a senior staff engineer in the Flight Test Instrumentation group at Lockheed Martin Aeronautics in Ft. Worth, TX. This group handles all the flight-test instrumentation for the F-16, F/A-22, and F-35 program at the aerospace company.

Russell explains that test requirements originate from two groups-customers and 'discipline' engineers. Customers-the armed services that buy tactical aircraft-fighters-want to measure specific parameters that will verify the performance of an aircraft to agreed-upon specifications.

On the other hand, discipline engineers cover specific technical areas such as cockpit environment, acoustics, aerodynamic loads, and so on. The environmental engineers, for example, must provide a good operating environment for a pilot. To prove their system works, they tell the flight-test instrumentation group the parameters they must monitor during test flights.

In turn, the flight-test engineers tell the discipline engineers what they can actually deliver in terms of numbers and types of sensors, sensor resolutions, and other factors. 'If engineers have unrealistic requirements,' says Russell, 'we explain what we can fit on an airplane already crammed full of electronics and fuel.' Then Russell and his colleagues help the discipline engineers decide on appropriate sensors and measurements.

In addition to instrumenting an aircraft with the usual pressure, strain, flow, load, and temperature sensors, Russell's group can include more exotic 'sensors,' such as high-speed digital cameras. They put them under an aircraft's wings and on its wing tips to monitor bombs, missiles, and other ordnance, called 'stores' in military jargon. When a weapon separates from the aircraft, these high-resolution cameras can capture events at up to 500 images/sec.

Each sensor on an aircraft sends its information to a data-acquisition module. The outputs from the data-acquisition modules go to a special device that merges the data and produces a pulse-code modulated (PCM) signal. In addition to sensors, the instrumentation system includes a GPS receiver that provides timing information. The ability to synchronize timing with GPS signals means engineers don't have to reload timing information into an aircraft's instrumentation package before each test flight.

Real Estate Squeeze

For Russell and his colleagues, the biggest instrumentation challenge is size. Everything must fit into a compact shell that wasn't designed to accommodate any added test instruments. Russell notes that although many companies offer useful commercial off-the-shelf (COTS) products, his group can look only at a small number due to space and weight limits. 'As technology advances we can shrink the data-acquisition systems,' says Russell. 'But the engineers always want more data, so we use the same space but we monitor more sensors.'

Although flight tests on older generations of aircraft used magnetic tape to store test data, the instrumentation engineers now use a FLASH-memory recorder that stores 138 Gbytes. Lockheed Martin had a supplier significantly modify a COTS solid state recording system to include a Fibrechannel card and a standard fiber-optic connector. (Fibrechannel, a fiber-optic communication link, operates at 100 Mbytes/sec.) The recorder also stores data from eight PCM inputs, which carry information from the data-acquisition modules at up to 20 Mbit/sec each. Lastly, the recorder stores information from eight MIL STD 1553 data buses that each operate at up to 1 Mbit/sec.

But engineers don't have to wait until the end of a flight to obtain test results from the FLASH memory. The instrumentation package includes radio transmitters that beam data to ground stations for real-time monitoring. Each test aircraft includes antennas added specifically for test-data telemetry. Two antennas, one on the top and one on the bottom of an aircraft, ensure a ground station will receive sensor data no matter the orientation of the test plane.

Preparing for Take-off

All of the instrumentation wiring and almost all of the instrument modules are what Russell calls instrumentation orange. 'Anything meant for flight test is orange, so if people see orange wires coming out of an aircraft section on the factory floor, everyone knows they connect to sensors or instrumentation cables. A few blue wires connect to MIL STD 1553 buses, but everything else is orange.'

To eliminate the possibility of failures in the test instruments-which would ruin an expensive test flight-prior to placing instruments in an aircraft, the instrumentation group thoroughly tests individual components. Then it tests the complete instrumentation system using wiring harnesses and cables that simulate the actual aircraft wiring. Russell says, 'After the instruments get on an airplane, troubleshooting time increases by a factor of 10, and accessibility is difficult. We can't easily get to instruments because everything is buried inside an airplane.' In addition, many sensors get built-in during production, so some, like those in a wing assembly, remain inaccessible in a finished aircraft.

So just where does all the instrumentation fit in a tightly packed fighter? Russell's group uses the space set aside for the gun-yes, they still put them in many fighters. Out comes the ammunition drum and in goes an instrumentation package, specifically designed to fit into that space.

Real Time Data Collection

As an aircraft flies through test maneuvers, its transmitted data gets picked up at a special Lockheed Martin facility in Ft. Worth. This facility includes radio receivers, recording equipment, and decoders that convert the PCM signals from the sensors into data expressed in engineering units such as degrees, pounds/in², and gallons/min.

This information then appears on video strip charts that discipline engineers carefully monitor for unusual results that could indicate a failure or other problems. Each discipline engineer can monitor as many as 16 data channels related to his or her area of interest. Touch-sensitive screens on the monitors let the engineers annotate the recordings for later reference, and engineers also may opt to produce a paper chart of the data.

Because engineers who 'know' the equipment they designed monitor the aircraft's data in real time, they watch for the unexpected. In the event they detect anomalous data as a pilot goes through critical maneuvers, they can immediately alert a control engineer and have the pilot stop the maneuver and return to base. Each monitor station includes a removable hard drive, so at the end of a test, engineers can take captured data to a computer for further analysis.

'The digital cameras under the wing let us track exactly what happens in real time, image by image,' says Russell. 'If we see a problem, we can immediately modify or end a test. Before, if a problem arose, the pilot would have to land so we could process and analyze films. Then the pilot might have to make an additional flight based on our analyses.'

The discipline engineers also use cameras in the cockpit to monitor different types of displays undergoing testing, or to see what the test pilot is doing. If a test involves a live missile test, for example, the engineers want to observe what the missile 'sees' and reports on the pilot's display.

Russell stressed that Lockheed Martin can take its flight testing wherever clients need testing done. A special mobile van can receive telemetry information and retransmit it through a satellite link back to the company's Ft. Worth, TX facility. So to the engineers in Texas, an aircraft undergoing testing in California or over the Gulf of Mexico might just as well be operating locally. Instead of sending engineers and test equipment all around the US for tests, which could take weeks or months, the company now sends the van and two people to run it.

After a test, engineers often need to further analyze data from a flight. But how can they tell when separate events occurred? Like many military and aerospace test ranges, the Lockheed Martin facility complies with a standard that gives each data value a time stamp. This 'stamp,' which is accurate to a millisecond, makes it easy for a computer to properly align the acquisition times of stored data values. The standard, IRIG Standard 200-98, arose from the Range Commander's Council (http://jcs.mil/rcc), which oversees the work of the Inter Range Instrumentation Group (IRIG). One of that group's latest standards, IRIG 106 Chapter 10, specifies a standard that lets test ranges 'packetize' all forms of test data for easy access across hardware from many vendors. Accuracy is 0.1 microsecond.

Wind Tunnel Testing

The Flight Test Instrumentation group at Lockheed Martin also gets involved with wind-tunnel testing of inlets for jet engines. These engines require a constant stream of air to operate properly. During the design of an aircraft, engineers come up with many shapes and forms for air inlets and ducts. Some work and some don't. To find the best configurations, the design engineers run computational fluid dynamic (CFD) programs that test thousands of possible designs. Then they settle on a few for wind-tunnel testing.

The testing involves a scale model of the air inlet and the ductwork, in which sensors at the aerospace inlet plane (AIP)-the entrance to the jet engine-record pressures for later analysis. During a test run, the wind-tunnel operators move the model so air enters it at various angles and at various flow rates. (There's no jet engine in the wind-tunnel model.)

Not only do the test engineers want to measure pressures, and thus flow rates, they also want to watch for 'buzz,' a dangerous condition that prevents air from reaching an engine. You can think of buzz-a resonance at the air inlet-as an effect like the whistle you get when you blow over the mouth of an empty soda bottle.

Russell's team uses a series of vanes-what the engineers call rakes-fitted with pressure sensors to measure airflow. Those vanes position the sensors in a radial pattern at the model's AIP. A typical test may include 40 or more sensors. When testing F-35 models, the engineers use two rakes-one on the engine AIP, and one on the inlet to the vertical lift fan used for hovering. 'We found distortion in the airflow coming over the F-35's canopy,' says Russell. 'It turned out the distortion took place between several of the pressure sensors, so we had to add extra rakes to get better resolution.'

Several years ago, Russell and his colleagues built an analog recording system to capture wind-tunnel test data. The system relied on analog magnetic tapes, and the tapes alone for a single series of tests could cost $250,000. Recently the engineers redesigned the system to record data digitally using new PXI multi-channel data-acquisition and signal-conditioning cards. The entire system cost $250,000. And now tests take less time to run. The customers used to pay for everything, says Russell. 'But now we're on fixed-cost contracts, so we must save every dollar we can.' That quarter million-dollar saving per test is no flight of fancy.

Contributing writer Jon Titus can be reached at[email protected].