Experimental results and interpretation: yaw experiments with a controllable rudder and wingtip-mounted drogue chutes on flex-wing hang gliders

Experimental results and interpretation: yaw experiments with a controllable rudder and wingtip-mounted drogue chutes on flex-wing hang gliders

August 6, 2007 edition
Steve Seibel
www.aeroexperiments.org

General:

Beginning in approximately 2002, I've been carrying out some experiments with yaw control devices on flex-wing hang gliders. Some of these experiments involved a controllable rudder mounted on the keel. Others of these experiments involved small drogue chutes deployed in flight from the wingtips.

The controllable rudder was flown on a Wills Wing Spectrum, a Wills Wing Skyhawk, and an Airborne Blade. For some photos, see www.aeroexperiments.org/galleryexperiments.

The small drogue chutes were deployed from the wingtips of a Wills Wing Spectrum, a Wills Wing Skyhawk, a Wills Wing Raven, a Wills Wing Falcon, an Airborne Blade, a Icaro Laminar R-12, and an Aeros Stealth KPL. For some photos, see the link given above.

The small drogue chutes, and the rudder, all had the effect of yawing the nose of the glider to point to one side in relation to the actual direction of the flight path and relative wind and any given moment. For example, when I deflected the rudder to the left, or deployed a drogue chute from the left wingtip, the nose yawed visibly to the left and the yaw strings streamed to the left, showing that the nose was pointing to the left of the actual direction of the flight path and relative wind, so that the airflow over the glider had a right-to-left component. The yaw strings continued to stream to the left until the rudder was brought back to center or the drogue chute was jettisoned, regardless of any subsequent changes in the bank angle or the direction of the flight path.

Roll torque due to yaw or sideslip:

In general, the rudder created a "wrong-way" roll torque. In other words, the roll torque created by the rudder acted in the opposite direction from the roll torque that we would see when deflecting a rudder in most "conventional" airplanes or sailplanes with dihedral. In general, when I deflected the rudder to the left and locked it in that position, this created a continual right roll torque that made the glider try to roll (bank) toward the right. If I did nothing to counteract this roll torque, the glider would end up in a right turn with an increasing bank angle, even though the yaw strings demonstrated that the nose of the glider was pointing to the left of the actual direction of the flight path and relative wind at any given moment. To neutralize this right roll torque and make the glider fly in a straight line, I had to keep my body shifted to the left of the centerline of the control frame.

In general, a wingtip-mounted drogue chute also created a "wrong-way" roll torque. In other words, the roll torque created by a wingtip-mounted drogue chute acted in the opposite direction from the roll torque that we would see when making a yaw input in most "conventional" airplanes or sailplanes with dihedral. In general, when I deployed a drogue chute from the left wingtip, this created a continual right roll torque that made the glider try to roll (bank) toward the right. If I did nothing to counteract this roll torque, the glider would end up in a right turn with an increasing bank angle, even though the yaw strings demonstrated that the nose of the glider was pointing to the left of the actual direction of the flight path and relative wind at any given moment. To neutralize this right roll torque and make the glider fly in a straight line, I had to keep my body shifted to the left of the centerline of the control frame.

In the experiments with the wingtip drogue chutes as well as the experiments with the rudder, this "wrong-way roll torque" effect was always much stronger at low angles-of-attack (high airspeed) than at high angles-of-attack (low airspeed). This is not surprising in light of the following facts:

1) Sweep creates a dihedral-like effect. Flex-wing hang gliders have both sweep and anhedral. Therefore the roll torque that arises when we yaw a flex-wing hang glider will be determined by the competing effects of sweep and anhedral.

2) The dihedral-like roll torque created by sweep is known to be strongest at high angles-of-attack.

3) The roll torque created by anhedral is known to be relatively independent of angle-of-attack.

4) Therefore the competing effects of sweep and anhedral will create a stronger anhedral effect at low angles-of-attack (high airspeeds), and a weaker anhedral effect or stronger dihedral-like effect at high angles-of-attack (low airspeeds).

In the experiments with the wingtip drogue chutes as well as the experiments with the rudder, the "wrong-way roll torque" effect was always much stronger with the VG loose or absent than with the VG tight. With the VG fully tight, at the min. sink airspeed--i.e. in the high-angle-of-attack (low airspeed) part of the flight envelope--some of the gliders actually showed a dihedral-like response to a yaw input rather than an anhedral-like response. As the bar was pulled the bar in to decrease the angle-of-attack, the coupling between yaw and roll progressively changed from positive, to neutral, to negative-- i.e. the glider changed from showing a dihedral-like roll response to a yaw input, to showing no response to a yaw input, to showing an anhedral-like response to a yaw input. Here are some thoughts on why the "wrong-way roll torque" effect was always much stronger with the VG loose or absent, than with the VG tight:

1) A flex-wing hang glider has a complex wing shape and it is not adequate to quantify anhedral by measuring the droop in the leading edge in relation to the line of the keel tube or the wing root chord line. The billowed shape of the trailing edge creates a dihedral geometry in the inboard parts of the wing and an anhedral geometry in the outboard parts of the wing. The geometry of the outboard part of the wing is most important to the overall balance of roll torques, because the outboard part of the wing lies furthest from the CG. Therefore, from an aerodynamic point of view, loosening the VG to increase the sail billow increases the wing's overall anhedral geometry, and tightening the VG to decrease the sail billow decreases the wing's overall anhedral geometry. This can be seen simply by looking at the wing from the side.

2) Therefore the conventional idea that flex-wing hang gliders with pulley VG systems have more anhedral when the VG is tight than when the VG is loose needs to be re-examined. When we look at anhedral in a holistic aerodynamic sense, rather than simply by measuring the droop of the leading edge tubes in relation to the keel tubes, the reverse appears to be the case: flex wing hang gliders (especially with cam VG systems, but also with pulley VG systems) can be seen to have more anhedral with the VG loose than with the VG tight.

3) Since the keel on a flex-wing hang glider can shift in relation to the rest of the airframe, the experiments with the keel-mounted rudder were not a perfect model of how flex-wing gliders respond to a sideways component in normal flight. When deflected, the keel-mounted rudder undoubtedly shifted the keel to one side in relation to the rest of the airframe. This would shift the sail billow in a way that would contribute to the "wrong-way roll response". With the VG tight, this billow shift would be diminished. How much of the "wrong-way roll response" to the rudder was due to the shifting keel? It's unfortunate that the Wills Wing Raven was not flown with the controllable rudder, as the keel on this glider was not free to float in relation to the crossbar. It would have been interesting to see whether this glider showed a "wrong-way" roll response to rudder inputs over the whole airspeed range.

4) It doesn't appear that deploying a drogue chute from the left wingtip would directly distort the airframe in any way that would contribute to a right roll torque, regardless of the VG setting. Therefore the experiments with the wingtip drogue chutes do seem likely to give a "true" indication of how flex-wing gliders respond to a sideways component in the relative wind. The fact that in the drogue chute experiments the "wrong-way roll response" was stronger with the VG loose than with the VG tight does support the idea that flex-wing hang gliders have more anhedral, in an overall aerodynamic sense, when the VG is loose than when the VG is tight.

The Airborne Blade was the only glider with a VG that was flown with both a rudder and a wingtip drogue chute (in separate flights.) Flying the glider with VG tight and noting the "neutral" airspeed where a left rudder deflection created neither a left nor right roll torque, and also noting the "neutral" airspeed where a drogue chute deployed from the left wingtip created neither a left nor a right roll torque, gave a way to compare the overall dynamics at play in these two experiments. The "neutral airspeed" with the drogue chute was only slightly higher than the "neutral airspeed" with the deflected rudder. This suggests that the two types of experiments were roughly equivalent. This suggests that in the rudder experiments, only a small part of the "wrong-way roll torque" was due to the fact that the rudder shifted the keel to one side in relation to the rest of the airframe. Nonetheless I still feel that the experiments with the wingtip drogue chutes offer a more valid representation of how a flex-wing hang glider responds to a sideways component in the relative wind, in normal flight.

Aerodynamic sideforce:

Note: the aerodynamic sideforce effects we'll discuss below are rather insignificant in comparison to the roll torque effects we discussed above. The aerodynamic sideforce effects are theoretically interesting, but play only a very minor roll in the glider's flight dynamics.

In addition to looking at the direction and magnitude of the roll torque that arose when I deflected a keel-mounted rudder or deployed a drogue chute from a wingtip of a flex-wing hang glider, I also used a slip-skid bubble to measure the direction and approximate magnitude of the aerodynamic sideforce that existed when the gliders flew in yawed attitude with the rudder deflected or with a drogue chute deployed from a wingtip. This sideforce was quite small.

Interestingly, in the experiments where a keel-mounted rudder was deflected to the left, causing the nose of the glider to point to the left of the actual direction of the flight path and relative wind, so that a yaw string streamed to the left, the slip-skid bubble actually deflected very slightly to the right. A conventional slip-skid ball would have deflected very slightly to the left. This is the opposite deflection of what we see when we deflect the rudder to the left in a "conventional" airplane or sailplane with a fuselage and fixed vertical tail. This indicates that the aerodynamic sideforce created by the rudder itself was actually stronger than the aerodynamic sideforce created by the relative wind striking the various surfaces of the hang glider in a sideways manner. This is not surprising in an aircraft with no fuselage or fixed vertical tail. Flying-wing aircraft are known to experience unusually small aerodynamic sideforce components when there is a sideways component in the relative wind.

Since hang gliders lack a good horizon reference line, slight bank angles are a bit difficult for the pilot to detect. However, in the particular case where an aircraft is made to fly in a linear, non-curving, flight path, the slip-skid ball or bubble serves as a bank angle indicator. A ball deflects toward the low wing and a bubble deflects toward the high wing. Therefore the above results indicate that the when the keel-mounted rudder was deflected to the left, the hang glider had to be banked slightly to the left to yield a straight-line flight path. Again, this is opposite from what we see in a "conventional" airplane or sailplane with a fuselage and fixed vertical tail, and indicates that the aerodynamic sideforce created by the rudder itself is actually stronger than the aerodynamic sideforce created by the relative wind striking the various surfaces of the hang glider in a sideways manner.

The above results indicate that if the wings were kept exactly level, left rudder would have actually produced a slight right turn. When we look only at the actual turning force produced by the glider when the rudder is deflected-- i.e. when we look at the centripetal force which makes the flight path curve, not the yaw torque or roll torque-- we find that this acts in the opposite direction from what we normally see when we deflect the rudder on an aircraft. Again, this is consistent with the idea that the aerodynamic sideforce created by the rudder itself is actually stronger than the aerodynamic sideforce created by the relative wind striking the various surfaces of the hang glider in a sideways manner. This is not surprising on a flying-wing aircraft with no fuselage or fixed vertical fin. (A similar situation exists when the rudder of a twin-engined aircraft with one failed engine is deflected exactly enough to keep the nose of the aircraft pointing directly into the relative wind. Unlike most cases where a rudder is strongly deflected, in this particular case there is no sideways airflow over the fuselage, so the sideforce from the rudder itself becomes significant, and will cause a turn away from the direction of the deflected rudder, if the aircraft's wings are kept level. This turn can be prevented by increasing the rudder deflection, so that the airflow strikes the side of the fuselage and makes a sideforce in the opposite direction, or by banking the wing toward the direction that the rudder is deflected, i.e. toward the "good" engine. The latter is the most efficient because the fuselage remains streamlined. The fundamental purpose of the bank is not to help counteract the asymmetric torque from the working engine, but rather to cancel out the aerodynamic sideforce from the deflected rudder.)

In the experiments where a drogue chute was deployed from the left wingtip of a flex wing hang glider, causing the nose of the glider to point to the left of the actual direction of the flight path and relative wind, so that a yaw string streamed to the left, the slip-skid bubble deflected very slightly to the left. A conventional slip-skid ball would have deflected very slightly to the right. This deflection is in the same direction as we see when we deflect the rudder to the left in a "conventional" airplane or sailplane with a fuselage and fixed vertical tail. The very small magnitude of this deflection shows that the impact of the sideways component in the relative wind against the various surfaces of the hang glider created only a very small aerodynamic sideforce. Again, this is not surprising in an aircraft with no fuselage or fixed vertical tail. Flying-wing aircraft are known to experience unusually small aerodynamic sideforce components when there is a sideways component in the relative wind. This is consistent with the observation that a hang glider pilot never feels a strong sideways force pushing him away from the centerline of the control bar, regardless of how much sideslip the yaw strings are indicating.

Again, in the particular case where an aircraft is made to fly in a linear, non-curving, flight path, the slip-skid ball or bubble serves as a bank angle indicator. A ball deflects toward the low wing and a bubble deflects toward the high wing. Therefore the above results indicate that when a drogue chute was deployed from the left wingtip, the hang glider had to be banked slightly to the right to yield a straight-line flight path. This direction of bank is consistent with what we see when we make a left yaw input in a "conventional" airplane or sailplane with a fuselage and vertical tail, but the very small magnitude of the bank shows that the impact of the sideways component in the relative wind against the various surfaces of the hang glider created only a very small aerodynamic sideforce. Again, this not surprising in an aircraft with no fuselage or vertical fin.

The above results indicate that if the wings were kept exactly level, deploying a drogue chute from the left wingtip would have produced only a very slight left turn.

See John K Northup's 1947 Wright Memorial Lecture for another reference to the low aerodynamic sideforce arising from slipping flight in flying-wing aircraft, and also to the small drag penalty arising from slipping flight in flying-wing aircraft, and also to the objectionable (wrong-way) sideforce resulting from the use of a large short-coupled rudder.

Click here for a condensed version of this article, with most of the interpretation omitted.

Copyright © 2004 aeroexperiments.org