Subject: NAS2-99090 October monthly report

Date: Nov 1, 2000

FLIGHT TESTS RESUME

Most of the work effort for October was directed at flight tests and all the preparations required for flight tests (i.e. recheck weight and balance, calibration, control setup, recheck and test all systems including data collection and video, etc.). Flight plans, test results and changes made as a result of flight tests are described below.

The CarterCopter (CC) returned to Olney, Texas, on Friday, October 6th for resumption of flight tests. High-speed taxi runs were made Saturday to explore the idea of doing a conventional autogyro rolling takeoff. During one of these runs the aircraft lifted into the air approximately one foot. The DV camera located on the tail recorded the event. The picture quality of the DV camera and the recorded cockpit conversation are broadcast quality. The first goal is for the four new CC pilots to gain the necessary flight experience using conservative procedures to protect both the pilots and aircraft.

The rolling takeoff idea presented several problems. We decided to initially keep the rotor pitch constant to reduce the number of variables. When the rotor was tilted back far enough to catch the air needed to maintain the necessary rotor rpm, the CC would tilt back onto its training wheels, which are designed for a max of 25 mph. We next tried tilting the rotor forward to keep the training wheels off the ground until reaching (approx.) 50 mph -- at which time there was enough lift from the horizontal stab to keep the tail up so the rotor could be tilted back. During the period of acceleration to 50 mph, the rotor produced sufficient lift to raise the CC up on its main gear but not enough to provide stability. Without stability, cyclic control was very sluggish and rudder inputs caused the CC to roll due to its (now extended) soft landing gear -- making the CC feel unstable.

The pilots felt uncomfortable with the situation so it was decided to change the takeoff procedure. They would try lifting off at a lower speed by increasing collective. During the rotor over-speed run-up, the pilot experimented with pulling collective to get its feel. The plan was to limit the torque to 1800 ft-lbs for the initial tests but the torque display was hard to see and he accidentally torqued the drive system to 2400 ft-lbs. The gearbox failed. Later examination showed the ring gear support was not stiff enough to keep the ring gear from pushing away from the pinion gear, which moved its contact point out to the edge of the gear and greatly increased its bending moment and stress level. This deflection aspect had not been considered in the original design.

The following two weeks were spent designing, building, installing and testing a modification to the gearbox ring gear support. The engine water temperature got hot at Olney during low speed taxies, so we added another fan to uniformly pull air through both radiators and the oil cooler. The additional fan draws fresh air into the engine compartment and keeps it cooler also.

We returned to Olney Friday, October 20th and accomplished a few more items on our checklists despite having to dodge rainstorms and fix minor mechanical problems. The pilots decided to begin this flight-test program by exploring hover type jump-takeoffs. We spun the rotor up 104 times in the first 3 days of testing.

All day Friday and Saturday morning, we repeatedly over-speeded the rotor, reduced throttle and pulled collective until the aircraft just cleared the ground. There is a fair amount of stick friction/force so there was a tendency to over control. This was uncomfortable for the pilots, so the numerous repetitions were made in the hope the pilots would develop a feel for controlling the aircraft.

The additional cooling fan worked fine, although after several rotor over-speeds the engine needed to be idled for a couple of minutes to let the temperatures drop. A few times we did not wait long enough and boiled the water.

For Sunday flight-tests, the pilots decided to let the aircraft roll forward a little before pulling collective. They did this by releasing the brakes and not reducing the throttle back to idle. They hoped it would give them a little more time in the air before the rotor rpm dropped too low -- and possibly help reduce the lateral sensitivity. When they performed this procedure, they found it made the lateral stability worst. As the aircraft moved forward the downwash over the rear of the rotor increased faster than at the front of the rotor. This caused an unbalance in lift on the rotor and made the rotor plane of rotation dip to the right, which had to be countered with a cyclic input to the left. The unbalance in rotor lift changed with speed -- forcing them to constantly change the cyclic input. Not having had any real experience at flying the aircraft, and not knowing what to expect, the pilots were reluctant to climb up more than a few feet or accelerate past 20 mph where the rotor control should become easier. As a result we had a couple of hard landings (no damage) that once again caused us to rethink the flight- testing approach.

RETURN TO THE ROLLING TAKEOFF

The following procedure was developed after consideration of everything learned to date:

1. Establish that the aircraft will not rock back on the training wheels during acceleration if the rotor plane of rotation is kept parallel with the ground.
2. Establish the aircraft is stable and controllable at taxi speeds up to 60 mph.
3. Determine a throttle setting at the prerotate clutch release point that can be held constant to provide the following:
   1. A comfortable acceleration whereby the rotor does not drop below 325 rpm by the time the aircraft reaches 30 mph.
   2. Sufficient HP to continue to accelerate in the relative high drag of autorotation at the slow speed of 30-40 mph.
4. Rapidly pull collective at 30 mph until the aircraft becomes airborne and then continue to pull collective to hold the aircraft in the air until the collective reaches a preset detent position of 5 degrees. At this point the pitch cyclic will be varied to hold the desired flight path.
5. At 50-60 mph reduce the throttle to maintain airspeed or close it entirely if the pilots become uncomfortable at any time with the way the aircraft is handling.

During a rotor over-speed run up prior to completing the above procedure, the gearbox again failed. The pilot did not add more torque than had already been used numerous times before. The failure ended this second round of flight-testing.

ROUND THREE

We repaired the gearbox and resumed flight-testing at Olney on Friday, October 27th. We pampered the gearbox by keeping the loads under 1500 ft-lbs. of torque. At 1500 ft-lbs. we can still over-speed the rotor to 350 rpm, which is sufficient to make short rolling takeoffs.

When we redesigned and rebuilt the gearbox ring gear support following its first failure, we proof tested it by running it over 2600 ft-lbs of torque for (approx.) 30 seconds and numerous additional times at over 2000 ft-lbs. The current gearbox problems are caused by the fact the gears we are using were designed to turn in the opposite direction. The loads trying to spread apart the gears are 4½ times higher than normal due to the higher pressure-angle of the gear. During this repair, we increased the preload on the tapered roller bearings to the maximum amount recommended. At the same time, we reduced the backlash to the minimum amount recommended. If the gearbox continues to fail, we will have to design and build another gearbox to resolve the problem. This will take approximately 3-4 weeks.

Fig. 1 - New configuration CC w/ larger diameter rotor


Fig. 2 - New configuration CC w/ extended stabilator, longer nose
gear and air scoop

Friday was spent trying to overcome the control problems that the new rotor hub design has presented to the pilots. One problem comes from the gyroscopic reaction of the spindle stabilizer bar to a cyclic input. The spar and spindle stabilizer bars have made the rotor very stable and eliminated any stick shake, but because of the gyroscopic precession, the spindle initially moves (when given a rapid position change) at an angle approximately 45-degrees to the stick direction. It eventually lines up with the stick, but both spindle movements (initial and delayed) are contrary to what the pilot needs or expects.

Another problem we addressed was determining the optimal position of the stick/spindle when the collective is pulled and the aircraft lifts off the ground. This position changes with forward speed and how fast collective is pulled. If the spindle angle does not have the rotor lift going through the aircraft center of gravity, then the CC will pitch or roll suddenly when lift/collective pitch is increased.

Added to this is the potential problem of throttle change, which produces a torque change and a rolling moment to the aircraft. We try to limit as many variables as possible when first lifting off the ground; including throttle, stick position, rudder, airspeed and collective -- which is positioned into a preset detent.

On the first flight Saturday morning, the numerous potential distractions/disturbances appear to have occurred at the same time. The pilots may have felt a little rushed because rain was predicted and I was urging them to stay in the air a little longer to give them and the aircraft time to settle down and stabilize. Added to this was the distraction caused by missing the collective detent several times while in the air. The result was that aircraft control was not smooth and the right tail boom slapped the ground. Enough damage was done to stop further testing.

The tail boom has been repaired. Before we fly again we will put the CC in the pit and see how much weight can be removed from the stabilizer bars. I did not expect the gyroscopic reaction to be significant and thus made no effort to reduce the weight from what we initially tried and found to make the rotor stable. We also plan to test the rotor with a reduced amount of teetering underslung to see if the spar stabilizer bar weight can be further reduced. Since our spindle tilt axis is located on the teetering axis, an out-of-balance due to teetering is not transferred into the cyclic or felt by the pilot.

We will make the following changes to improve control:

1. Preload the cyclic linkage to remove its slop or deadband.
2. Replace all control related Teflon-lined rod ends and spherical bearings with ones that have aluminum bronze inserts.

The Teflon-lined bearings are self-lubricating, but are so tight they have more friction. I suspect that half of the control friction is a result of the 26 rod ends and 16 spherical bearings.

Next flight tests are scheduled for November 10 - 13.