BD-5 FLIGHT TEST PROGRAM REPORT

LESTER H. BERVEN (AM)
CHIEF TEST PILOT
BEDE AIRCRAFT, INC.
NEWTON, KANSAS

(Published in the journal of the Society Of Experimental Test Pilots)

Introduction

The BD-5 is an amateur-built general aviation aircraft which represents a new concept in sport aviation. The aircraft is purchased as a complete materials package and is assembled entirely by the owner/pilot. Upon completion, the BD-5 is licensed by the FAA in the Experimental-Amateur Built Category and is restricted to a local area defined in the aircraft operating limitations until it has flown 75 hours, at which time the local area restriction is removed, and it is operated in accordance with FAR 91 for homebuilts.

The BD-5 is basically a single-place, low wing monoplane pusher with several unique design characteristics. It has removable wings with a constant diameter tube spar which fits over the center section, a two cylinder, two cycle, dual ignition, internally mounted mid-engine which drives the wooden, fixed pitch prop through a Gilmore belt reduction drive system with a 1.6 ratio. It has manually retractable landing gear, and a combination push-rod/cable control system which is actuated with a side stick on the right console. Both a long wing (21.5 ft. span) and a short wing (14.3 ft. span) are available.

The BD-5 was first conceived in 1967 as an ultralight, self-launched glider, but in searching for an acceptable powerplant, it was found that the high power to weight ratio of the snowmobile type 2-cycle engine would allow a very efficient powered aircraft and development was channeled along these lines.

Development Flight Tests

The first prototype BD-5, N500BD was built in 1970. The fuselage was bolted aluminum angle framework covered with a molded fiberglass shell. The wing structure was an aluminum tube-spar with aluminum ribs and wing skins. This aircraft had the original V-tail and was powered by a 36 BHP Polaris snowmobile engine.

This fiberglass V-tailed prototype was first flown in September 1971 by Jim Bede. It made a total of 2 flights, both in ground effect just above the runway. These two flights showed a serious deficiency in both directional and longitudinal stability and control and a redesign of the tail was begun immediately. The next configuration tried was a swept conventional vertical fin with a highly swept (60 degrees at the L.E.) stabilizer/elevator. High-speed taxi tests with this configuration showed very limited elevator power for rotation and a large trim change with power due to the induced flow from the propeller.

Because of the great interest in the BD-5 and the large number of orders he was receiving, Jim had previously decided to pay for the tooling necessary to produce a metal fuselage, and the first all-metal BD-5, N501BD, was being built at this time. Because of the limited manpower available and the difficulty of making modifications to the fiberglass BD-5, Jim decided to stop development and testing of the fiberglass prototype and concentrate on getting N501BD in the air. It was also at this time that he decided to add some more manpower, and he hired Burt Rutan as Director of Development and myself as Test Pilot in early 1972.

The first all metal prototype, N50lBD, differed from ship #1 not only in the fuselage design but it also had an all-flying stabilator, an American made Kiekhaefer 440cc snowmobile engine, electrically retractable landing gear, and a variable speed drive system. The cooling system for this engine installation was two NACA flush scoops (one on each side of the fuselage) from which the cooling air was ducted down over the cylinder fins and out the rear of the fuselage in an exhaust/ejector duct. The initial ground testing of this aircraft began in May 1972 with an evaluation of engine starting and cooling characteristics and with low speed taxi tests. These tests showed that the engine cooling was marginal on the ground and that the rudder was effective for taxiing above 25 mph IAS and that the ailerons were noticeable above 20 mph IAS.

High speed taxi tests were begun on 31 May and control effectiveness was evaluated in 5 mph increments up to 80 mph IAS. We found, as before, that the ailerons and rudder were very effective, but that the stabilator was not powerful enough to lift the nose, even at 80 mph (cg was 24% mac). We then reballasted the aircraft to 27% cg and tried it again. This time I could rotate by accelerating to 70, pulling the power back to idle and applying full aft stick. Once the nose came up I could add full power and hold it there. By using this technique, I got the airspeed up to 80 mph IAS, made the first liftoff, flew down the runway at a height of about 5 feet and landed after about 10 second flight. Both trim and controllability seemed acceptable at 80 mph.
Before the next flight, the main gear was moved forward one inch and the cg was moved aft to 28% in an effort to lower the rotation speeds.

The next flight was made on 7 June 1972 at a gross weight of 690 lb. Some more high speed taxi tests showed that the power-off rotation speeds had been lowered to 65 mph but that it was still impossible to rotate power-on. Using the old "throttle back to rotate" trick I accelerated to 75 mph and lifted off for an intended flight around the pattern. Acceleration to 80 mph was normal and I added a little back pressure to begin a gentle climb. The nose pitched up about twice as much as I wanted so I added forward pressure, with no effect.  I pushed even harder, and the nose came back to where I wanted it. I was then at 95 mph IAS and about 20 feet above the runway. At this point the nose began an uncommanded pitch down and it took considerable aft stick for recovery. This was followed immediately by a more violent pitch up to about 2g and a pitch down to slightly negative g. At this point I decided that this wasn't a very good day for the first flight around the pattern.

I pulled the power back to slow down to a previously flyable air-speed. As the airspeed decreased, the controllability improved and I made a reasonably normal landing in the remaining runway. After much discussion and some more ground tests we decided that the problem could have been caused by having the aerodynamic center of the stabilator forward of the pivot.  Calculation showed it to be right on, but with the highly swept stabilator the inflow from the prop would have moved the center of pressure inboard, and therefore forward. We then changed the stabilator geometry to increase the area 15% and moved the pivot 1" forward to eliminate the possibility of any instability. With the new stabilator, the next taxi tests showed that the rotation speeds had not improved, and that with full aft stick, the stretch in the longitudinal control system would allow the stabilator to back off a full 10 degrees from its static maximum of 15 degrees.  This control system deflection plus an aerodynamically unstable stabilator was undoubtedly the problem on the second flight.

Because of these problems and the fact that the stabilator with a sweep of 60 degrees was only giving us a CLmax of 0.6, we decided to completely redesign the longitudinal control system and change the stabilator to an unswept trapezoidal planform with a thicker airfoil section, 20% more area, and a larger anti-servo tab.

During this period of redesign, the Kiekhaefer 440 cc engine was replaced with a Hirth 650 cc and a series of airframe/engine integration tests were run.

The new stabilator was ready for test on 8 July 1972. To check the stick free stability of the stabilator, we disconnected and locked the anti-servo tab, and at 40 mph IAS measured the stick force at full aft stick. This technique indicated a very slightly stable tail, which was right where we wanted it. High speed taxi tests showed a power-on rotation speed of 50 mph IAS, and a power-off of 40 mph at a cg of 25% mac. Once the nose-wheel was off, I could hold it off down to 30 mph.

The next flight, on 11 July 1972, was the first real flight of the BD-5, when it flew away from the runway for the first time. Acceleration to 80 and lift-off were normal. I held the aircraft level and let it accelerate to 100 mph while I made some quick stability checks and leaned the mixture slightly. Everything appeared normal, so I raised the nose to hold 100 mph. Initial rate of climb was about 600 fpm, with gear down and gw 700 lb. Cylinder head temperature at lift-off was about 380 degrees F (450 degrees F is redline). Passing the end of the runway at 500 feet, CHT was 440 degrees F and increasing, so I started a left turn back to key position, still climbing.

As I turned downwind, CHT was 550 degrees so I leveled off at 800 ft., throttled back and adjusted the mixture to cool down the engine. Half throttle gave about 100 mph IAS but the CHT stayed at 550 degrees so I decided to return for landing. As I turned base, I pulled the throttle back to idle and set up a 500 fpm descent at 100 mph. About halfway across base leg just to keep it interesting, the cockpit started to fill with smoke, so I lowered the nose to increase R/D and turned final. The smoke was getting worse and I wasn't coming down very fast, so at about 100 feet and just off the end of the runway, I shut the engine down and turned off the master switch, neither of which seemed to affect airspeed, R/D or the amount of smoke. Touchdown was at about 80 mph and the smoke was really bad. As soon as I lowered the nosewheel at about 60 mph, I opened the canopy to clear out the cockpit and did our first high speed canopy opening tests. At 50-60 mph the canopy floated directly overhead and could be pushed back to full open with very little force on it. I taxied up to the hangar and reported the smoke, but nothing could be found. It turned out that the smoke was exhaust recirculating into the cockpit which we fixed by installing a fresh air vent and sealing the firewall better. Ground tests showed that the cooling problem was a lack of adequate air exit at the exhaust ejector and we wound up an adjustable cowl flap on the bottom of the fuselage. With this fix we started flying the aircraft regularly and were able to get the first flying qualities data on the next several flights.

Instrumentation

Before I describe the results of the flying qualities tests, I'd like to show you our instrumentation system.  For documentary shots, roll rate, and dynamic stability data we use an over-the-shoulder super-8 movie camera, some film from which you have already seen. For stick force and position data we have a homebuilt force grip designed by Burt Rutan which is very simple and surprisingly accurate. Stick position is read off a flexible tape threaded through the top of the stick. Admittedly the system isn't very sophisticated, but it does allow us to see the variation in significant stability parameters with cg position, which is all we wanted anyway.

Flying Qualities

Both the looks and the performance of the BD-5 are excellent, but the thing that really makes this little airplane great is the flying qualities. Almost invariably the first question everyone asks is "Isn't that fast little short-coupled airplane going to be awful sensitive and hard to fly?" The answer is absolutely not! It was recognized in the early design stages that very small airplanes are sometime overly sensitive and have low control forces, and that was the major factor in choosing a wrist controlled side-stick for the longitudinal and lateral control. By decreasing the pilot's force capability to that available from his wrist, the aircraft and pilot forces are matched; and although the stick forces are very light, they feel absolutely normal. There is no tendency toward P10 or overcontrolling in any flight condition we've investigated so far, and we've been to 240 mph IAS and 5 g's, including most normal aerobatics. I've found the same thing the Air Force studies on the side stick show; that compared to the center stick, the side stick offers more precise control, a more natural seated position, more useable instrument panel space and more natural control movements.  With both forearms supported, turbulence or aerobatic accelerations do not feed back into the controls, and fatigue is reduced on long flights. I really believe the side stick is the optimum system for very small aircraft, and I don't think the BD-5 would be a success without it.

Our early tests with N501BD did show that the stick force per g was too light-only about 3/4 lb/g. On the next several flights we increased both the deflection and area of the stabilator anti-servo tab and arrived at our present optimum value of about 3 lb/g. Now, that may sound a little light, but it fits perfectly from a control harmony standpoint, and it's purely wrist action. In fact after about 30 minutes of aerobatics it feels like it may be too much. This does bring up an interesting point, however; if you raise your forearm off the armrest, the longitudinal control dynamics change considerably, and the stick force per g becomes too light for an inexperienced pilot, who would then be easily capable of overstressing the aircraft. I'm not sure what the solution to that situation is, other than possibly some kind of forearm restraint. The results of our longitudinal static stability tests are shown in figure 1 for two cg positions. Stick forces and deflections are both stable even back at 27.5% mac. Maneuvering stick forces are shown in figure 2, and so far have set the aft cg limit at 28.5% which gives a 1.5 lb/g gradient.

At normal cg (25%) the longitudinal short period is deadbeat at all speeds.  The phugoid, however, has a period of only 15 seconds and is neutral to very slightly damped. This is probably due to the fixed pitch prop/high thrust line combination, which causes an unstable pitching moment with speed changes. This theory is supported by the jet BD-5 which has a much more well damped phugoid.

Lateral/Directional Stability

Lateral directional stability is good, with positive dihedral effect and light but well harmonized rudder forces. At 120 mph IAS a full rudder sideslip requires about 15 degrees of bank for straight flight. Full rudder sideslips are possible from 120 mph IAS down to stall with no adverse characteristics in either CR or PA configuration. Rudder fixed, the directional short period has a frequency of about 0.8 cps and a damping ratio of about 0.5. There is no tendency toward dutch roll and spiral stability is neutral to very slightly stable. Adverse yaw is quite low at cruise speed, and at approach speed, a rudder fixed roll at about 30 degrees/sec will result in about a one-ball sideslip on the slip indicator. As a result, rudder/aileron coordination is very easy and makes it a pleasant airplane to fly. Maximum roll rates are about 120 degrees/sec with the long wings, and 200 degrees/sec with the short wings. Rudder power is s6mewhat limited with the long wing and below 80 mph it takes full rudder for coordinating a full deflection aileron roll.

Stalls

Stall testing has uncovered two significant features-one good, and one bad.  The good part is the stall characteristics. There is a very definite buffet 3 to 5 mph before the break, and no tendency to roll off until you have a moderate amount of sideslip. Up to a full ball sideslip, the aircraft will still break straight ahead, and recovery is immediate if you relax any back pressure. The not-so-good part is that the stall speeds are significantly higher than we had expected. Data show a maximum lift coefficient of only 1.06 clean and 1.45 with full flaps. This works out to gross weight stall speeds of 72 mph clean and 61 with flaps for the long wing and 86 mph clean and 74 dirty with the short wings. We feel this poor CLmax is due to a combination of low Reynolds number, laminar separation and a relatively large tail download (15% of gw).

Current Flight Tests

We managed to get 33 flights and 25 hours on N501BD and were able to complete most of our flying qualities optimization. Both systems and performance were non-representative, so the best we have there is preliminary data. In those 33 flights we had two inflight engine seizures due to lack of adequate cooling and wound up cooling the engine from a scoop on the bottom of the fuselage. It seems that the airflow has to go from the exhaust side of the cylinders to the intake side, or you get hot spots around the ex-haust ports.

The last flight of N501BD occurred on 8 Oct 1972 and was the fourth flight with the short wings. Take-off was normal up until I started leaning the mixture at about 50 feet.  The engine didn't respond normally and sounded richer as speed increased and unloaded the prop. I continued to lean until the mixture was full out, with no apparent effect. I was then about 300 feet over the end of the runway with no rate of climb. I made a left turn to line up on a convenient road and lost about 150 feet in the turn. Thinking I had possibly leaned the mixture too much, I tried richening it slightly, whereupon the engine quit completely.  I dumped the nose to maintain as much airspeed as possible and put the flaps down as I flared. It was at this point I discovered that the short wing BD-5 at 720 lb. and a rate of descent would not flare completely, even starting at best rate of climb speed (105 mph IAS). I touched down on the road at about 85 mph and an 800 fpm descent folded the main gear aft and ran off the left side of the road down into a 10 foot ditch, up the other side and came to a stop in a small cloud of dust. The airplane was in surprisingly good shape. The major damage was to the leading edges of the wings, wrinkled fuselage bottom, and collapsed landing gear. The only injury sustained was to a field mouse I mashed at the bottom of the ditch. Examination of the engine showed that the mixture control cable had broken on the take-off roll and I was unable to compensate for the richer condition of the engine at higher speed. It appeared that it would take about a month to fix the air-craft, and since we had all the aerodynamic data on ship 2 we needed, and all the systems were non-representative, it was decided to repair it only for looks as a display and press on with the next airplane, N502BD.

N502BD differs from N501BD in that it has manually retractable landing gear, plain flaps, longer gear struts, an integral blower fan to cool the engine, a fixed ratio drive system with a toothed belt, and a new carburetor, calibrated for the prop load, so that there is no longer any requirement to change the mixture with throttle setting and airspeed. It also has a dual ignition engine with a capacitor-discharge ignition system.

The first flight of ship #3 was made on 26 March 1973 and the flight program has proceeded with very few problems since then. The new manual landing gear works very well-gear up or gear down takes about ½ second with very little pitch trim change. The landing gear lever moves nine inches and it requires only 10 lb. to actuate it. We have a total of about 60 hours on N502BD now and have completed most of our systems and flying qualities testing.

Since N502BD's performance is more representative of the final configuration, we've attempted to get some more accurate performance estimates. However, we have yet to get our gear doors to work properly, and we're flying without a nose gear door and the main gear doors stuck open about an inch to an inch and a half above 140 mph IAS.

The maximum power available from our present 650 cc engine is 50 BHP at S.L. and 6300 engine rpm because of our non-optimized exhaust system. With the configuration, the present long wing BD-5 will indicate 154 mph at full throttle and 7500 feet, (175 mph TAS) and about 180 mph at S.L. Rate of climb at S.L. and 660 lbs. is 900 fpm with a fixed pitch cruise prop at a best rate of climb speed of l00 mph. The short wing BD-5 with the same horsepower is about 15 mph faster and has 2/3 the climb rate. Take-off distance with the cruise prop is about 1800 feet for the long wing and 2500 feet for the short wing. These should be improved considerably when we get our CLmax up to a better value.

BD-5J

During our development program we came across a little turbojet engine that seemed to be just made for the BD-5, so we built another airframe and installed this Microturbo TRS-18 engine. It has 200 lb. thrust, and electron-ic fuel control with automatic starting and was designed to operate from a plenum chamber inlet. To get enough fuel capacity (50 gallons) we increased the gross weight of the aircraft to 950 lbs. and made an intermediate length (17 feet) fully wet wing. To take the higher gross weight landing loads, we also went to oil/air struts all around which we've tested to a 10 fps touch-down at 900 lb.

First flight of the jet was made on 20 July 1973 and lasted 30 minutes. The jet has all the same outstanding flying qualities of the prop BD-5 with a very smooth reliable jet engine and I doubt if there has ever been an airplane that's more fun to fly.

Stall speeds at 800 lb. are 76 knots clean and 65 knots with flaps. Take-off and landing distances are 1200 and 800 feet, respectively. Initial rate of climb at 130 KIAS is 1600 fpm and in our present configuration the service ceiling is 26,000 feet.

Because the oleo gear struts are significantly different from the fiber-glass ones, we are unable so far to completely retract the gear. So we not only don't have gear doors, but the main gear hangs down a little over 3 inches. In this configuration maximum speed at S.L. is 200 KIAS, decreasing to 130 KIAS, (185 KTAS) at 20,000 feet. Right now the performance is less than we'd hoped for but by cleaning up the gear and improving the engine thrust output we should have a pretty good performing airplane.

Conclusion

So far in our BD-5 test program we've opened up the airframe envelope to 220 KIAS and 5g and have completed our aerodynamic optimization except for cleaning up the landing gear and improving our CLmax. The remaining systems tests are engine power and cooling optimization and drive system development to decrease weight and improve the soaring capability.

The major flight tests remaining are the envelope expansion out to a Vdive of 300 mph and the spin tests. Our present schedule calls for completion of both of these during this fall, and well in advance of any of our customer's first flights.

 
Last Update: 10/21/99
Web Author: Juan Jimenez