A few years ago my friend Javier Arango and I got to talking about certain oft-repeated statements about the airplanes of World War I. Javier has a collection of very accurate reproductions of World War I airplanes (as well as a genuine Camel and Bleriot 11) and happens to be a scholarly fellow with a Harvard bachelor’s degree in the history of science. He is accordingly interested in teasing out the threads of truth from the tapestry of memory and invention that is the past. We thought it would be interesting to do some flight tests of his airplanes using modern measuring equipment, and to compare the results with contemporary accounts. It was — and continues to be — widely believed, for instance, that the Fokker triplane of Manfred von Richthofen was a superior climber, but theoretically it should not have been. Was it, in fact? Is it true that rotary-engine airplanes, because of the huge gyroscopes on their noses — the entire engine spun, and with it a huge propeller — turned more readily in one direction than the other? These were the kinds of questions that we hoped to resolve with numbers rather than anecdotes.
I learned when visiting the Naval Test Pilot School at Patuxent River, Maryland, a couple of years ago that relatively inexpensive flight-data retrieval systems are now available. They contain solid-state gyros and accelerometers and a GPS unit, and record a dozen or so parameters — angles, speeds, accelerations and so on — at intervals of a quarter of a second. Unlike the large built-in “black boxes” (actually orange) beloved of the NTSB, these are little, self-contained battery-powered cubes that you stick somewhere in the cabin with tape or Velcro. The stored data can be played back on a computer, either as graphs of selected parameters or as a flight simulator-style rerun of the flight, seen from outside the airplane or from inside the cockpit, with photographically accurate scenery and all the other miraculous stuff we have come to expect from Google Earth.
Javier bought one of these gadgets from Appareo Systems. It worked beautifully, but of course it lacked any provision for recording mechanical parameters such as control surface positions and stick forces. So Javier also got a stick-force digitizing device from Futek Advanced Sensor Technology, and a Graphtec data logger to record the Futek output as well as stick and rudder movements, which we obtained from several strategically placed slider potentiometers. I wrote a utility to synchronize the data streams from the Appareo and the Graphtec and combine them in a single Excel spreadsheet.
One thing we have no measurement of is power. But it happens that rotary engines do not have throttles; they have just four or five power settings, which are selected by shutting off the spark to different numbers of cylinders, and so the pilot’s kneepad notes are sufficient to record which power level is being used for each maneuver.
We decided to make the first tests in Javier’s Sopwith Strutter, a two-seat biplane powered by a 160 hp rotary engine and named not for its proud gait but for having “one and a half struts” between the wings. We chose it because, even though it was not in the same class as the Camel or Fokker triplane and its performance was of less historical interest, it is a well-behaved airplane with roomy cockpits, and, unlike the fighters, it has a rear seat that allows testing it at two different CG positions.
Being able to move the CG is desirable, because it is by testing speed and stick force at two or more CG locations that you assess longitudinal stability. Longitudinal stability is the quality in an airplane that makes it hold speed and return to its trimmed speed after a disturbance; it can also be thought of as speed stability. As the CG moves farther aft, it takes less and less stick force to hold the airplane at a speed other than its trimmed speed. When the CG gets far enough aft, the stick force ceases to change with speed; at that point the airplane no longer possesses any longitudinal stability — it is just as content to fly at one speed as at another.
The Camel has little or no speed stability to begin with; its CG is normally at the extreme aft end of the stable range. In order to measure its inherent stability, you would have to move its CG forward, because moving it any farther aft would make the plane dangerously unstable. But the Camel is a tiny, pug-nosed airplane with firewall and engine crammed right back against the pilot’s feet; just in front of the firewall, the engine is spinning around; and so the prospects of moving the CG forward are not good.
Javier made the first test flights in spring 2009. These were tests of tests, so to speak; we wanted to see whether the data collection system was working, what we could make of the results, and what we should be looking for.
Everything did work, and we emerged with thousands upon thousands of data points. The big disappointment was that our attempt to measure longitudinal stability by the stick-force-per-knot method failed, as Javier had suspected it might. Although the Strutter flies pretty much like a modern airplane except that its rudder is about 10 times more important, its control feel is so vague, control-system friction so great and speed range so narrow that it turned out to be impossible to identify a clear variation of stick force with speed.
But here are some other things we found out during the test of tests: The maximum rate of climb was 830 fpm at 43 knots — an impressively steep angle. Maximum level speed was a less-impressive 77 knots; minimum speed power-off was 41.5 knots and, with 50 percent of power, 37.4 knots. The takeoff roll took 10 seconds, liftoff in ground effect occurring at 30 knots. Aileron and rudder seemed to want to have as little as possible to do with each other. A rudder pulse in either direction without a corresponding aileron input produced a large yaw angle and a small bank angle — smaller with left rudder than with right. The Strutter, like all airplanes of the period, is as happy to fly somewhat sideways as it is to fly straight ahead. Javier handled the airplane gingerly; he never exceeded 20 degrees of bank or a roll rate of 20 degrees a second. If the airplane would rather turn one way than the other, it was not obvious. The Strutter is a big airplane, however, and, although its engine is quite powerful and its prop quite large (11 feet in diameter, to be exact), we really won’t have good data of the gyroscopic effects of the rotary until we have instrumented the Camel and the triplane. Before we do that, however, we will make one more flight with the Strutter. One thing we still need to do is to calibrate the stick forces with a spring scale of some sort; so far, all we have from the Futek is raw voltages.
Oh, I almost forgot! The phugoid period is 30 seconds. I knew you were wondering about that.
Fumes
On Nov. 29 of last year, a number of newspapers throughout the country picked up an AP wire-service story out of Des Moines, Iowa. The headline was “Full gas tanks could stop many small plane crashes” — certainly true, if the cause of the accidents were fuel exhaustion. The story reported that out of 8,016 crashes of civilian noncommercial aircraft during a five-year period going to the end of 2008, 238 were due to fuel exhaustion. In those accidents, 29 people died, as compared with 2,640 people in all.
The story does not quite bear out the alarmist headline. Only 3 percent of accidents were due to fuel exhaustion; comparatively speaking, they do not seem to merit the word many, though 238 is certainly many more than the ideal of none.
It seems, however, that if you must get into an accident, it had better be due to fuel exhaustion. It takes more than eight fuel-exhaustion accidents to kill one person, but only three of all the other kinds together. The main reason for the comparatively high survivability of fuel-exhaustion accidents is probably that the power loss tends to occur at altitude, giving pilots time to plan their unscheduled landings; and it may also be that, when there is no fuel, there is no fire, or that pilots are more likely to rashly stretch their fuel when they are alone. I would suppose, too, that people on IFR flight plans are much less likely to run out of fuel than those flying VFR, either because of a higher average level of professionalism or because IFR flight planning requires more careful planning and greater fuel reserves; and it is certainly easier to make a successful emergency landing in VMC.
As Tom Haueter, director of the NTSB’s Office of Aviation Safety, pointed out, fuel-exhaustion accidents, as a class, seem inexplicable. “You just go, ‘What are you thinking? What are you doing?'” The AP story cites a number of cases in which a pilot believed, for one reason or another, that he had more fuel aboard than he really did when he took off; but that is a poor excuse, because the bottom line in almost every case is that, when fuel gauges get close to the end of the scale, a cautious pilot lands whether he thinks they’re wrong or not. An Ottumwa, Iowa, college flight instructor, Jane Berg, is quoted as calling running out of fuel “probably the silliest mistake that a pilot can make.”
Well, there’s always trying to taxi while still tied down.
