Experimental results and interpretation: using yaw inputs for roll control while ground-handling flex-wing hang gliders
August 8, 2007 edition
Early in 2007, I visited Joe Greblo of Windsports at Dockweiler beach in the Los Angeles area. Joe demonstrated to me that yaw inputs in the "normal" direction work quite well for roll control in some flex-wing hang gliders whenever the pilot's feet are touching the ground-- i.e. during ground handling, launch runs, landing run-outs, etc. By yaw inputs in the "normal" direction, we mean that when the pilot creates a yaw torque on the control frame that yaws the nose of the glider to the left, this creates a left roll torque that lowers the left wingtip and raises the right wingtip. Controlling roll through yaw in this way works best when the glider is allowed to fly at a relatively high angle-of-attack, i.e. when the nose of the glider is not pulled down too aggressively.
During my visit to Dockweiler I saw that these yaw control inputs were very effective on a Wills Wing Falcon and a Wills Wing Condor. I convinced myself that these yaw inputs worked largely through the interaction between the sideways component in the relative wind and the 3-dimensional shape of the wing, and that a sideways airflow over the glider could create the desired roll torque even when the glider's heading was constant. In other words, the roll torque created by the pilot's yaw inputs were not solely or primarily dependent upon the difference in airspeed between the two wingtips that exists whenever the glider's heading is changing. Joe is convinced that these yaw inputs work well for roll control on many other types of flex-wing hang gliders, including high-performance wings.
I've not yet convinced myself that these yaw control inputs provide an effective means of roll control on some of the gliders I've owned or had access too, such as the Airborne Blade, Icaro Laminar R-12, or Aeros Stealth KPL, especially on flat ground. I'm definitely going to continue to pay attention to this in the near future.
Joe Greblo has also noted to me that to help his students learn to ground-handle gliders in wind, he stands at the back of the keel as the student balances the glider in the wind. When the instructor pushes left on the back of the keel so that the nose of the glider is displaced to point to the right of the actual direction from which the wind is blowing, this makes a roll torque to the right. Again we are seeing a dihedral-like effect, in spite of the fact that we are also undoubtedly creating some movement of the keel with respect to the crossbar, in a manner that should contribute an anhedral-like effect.
In any glider where yaw inputs do provide an effective means of controlling roll, this generally only works when the wing is allowed to fly at a moderate to high angle-of-attack. When the bar is pulled well in to keep the wing at a low angle-of-attack, a yaw input can be felt to make a roll input in the "backwards" direction. In other words when the bar is well pulled in, yawing the nose of the glider to the left can make the glider roll to the right. 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).
I don't know how to fully explain the apparent contradiction between the dihedral-like handling characteristics that I experienced while ground-handling the Wills Wing Condor and Falcon, and the anhedral-like roll responses I observed when I experimented with a controllable rudder and wingtip-mounted drogue chutes on several different flex-wing hang gliders (including a Wills Wing Falcon). (For more on these experiments, see the related article on this website entitled "Experimental results: yaw experiments with a controllable rudder and wingtip-mounted drogue chutes on flex-wing hang gliders", or for an expended version, see "Experimental results and interpretation: yaw experiments with a controllable rudder and wingtip-mounted drogue chutes on flex-wing hang gliders".) In general in these experiments, if the VG was loose or absent, a left yaw torque from the rudder or from a drogue chute deployed from the left wingtip created a right roll torque even when the wing was flying near min. sink, i.e. even at high angles-of-attack. In other words the wing showed at least a mild anhedral-like roll response to yaw inputs even at high angles-of-attack (low airspeeds), as well as a much stronger anhedral-like roll response to yaw inputs at low angles-of-attack (high airspeeds). On the other hand, during the ground handling experiments and techniques described above, the gliders showed a dihedral-like response to yaw inputs over a significant range of moderate to high angles-of-attack. Here are some ideas to explain this apparent contradiction:
1) It seems that by definition, when a pilot is running with a glider -- i.e. when the pilot's feet are still touching the ground -- the glider is not bearing the pilot's full weight. A glider may behave as if it has less anhedral when the glider is not bearing the pilot's full weight, possibly because the glider has less sail billow.
2) On launch, sloped ground tends to create a dihedral-like effect because as one wing is yawed forward, it becomes exposed to stronger ridge lift. This doesn't explain why these yaw techniques work during a run-out landing on flat ground.
3) For some reason perhaps the glider has a stronger dihedral-like effect, or a weaker anhedral-like effect, in ground effect than out of ground effect.
4) Perhaps some aspect of the experiments with the wingtip-mounted drogue chute, as well as some aspect of the experiments with the controllable rudder, is making the glider behave as if it has more anhedral than it really does, so that neither the experiments with the wingtip-mounted drogue chutes nor the experiments with the keel-mounted rudder are offer a good model of how the glider behaves in free flight when there is a temporary sideways component in the relative wind.
(I have a hard time seeing how this could be the case as the drogue chutes tended to trail slightly downward in the glider's downwash which should oppose, not enhance, any anhedral-like effect that is present.)
(The experiments with the keel-mounted rudder are not a perfect model of how the glider behaves in free flight when there is a temporary sideways component in the relative wind, because the rudder must create some movement of the keel with respect to the crossbar, which would tend to create an anhedral-like effect that would not be present in normal flight. The purpose of the experiments with the drogue chutes was to avoid this problem. However the same type of movement of the keel must also be present when an instructor or launch assistant pushes to the left or right on the aft end of the keel of a glider, and here we see a dihedral-like roll response to the yaw input. So once again we are seeing different results during ground handling and during flight.)
5) In general, the more weight a flex-wing hang glider is loaded with, the lower the angle-of-attack it tends to trim at. (This is why light pilots generally have to hook in further aft than heavy pilots, on a given wing.) Therefore during ground handling in moderate winds a glider may tend to trim at a higher angle-of-attack than in flight. As noted above we expect to see a weaker anhedral effect, or a stronger dihedral effect, as angle-of-attack is increased.
(I have a hard time seeing how this difference in trim angle-of-attack could really explain the fundamental difference between the results I saw during in my in-flight experiments and the results obtained while yawing a glider during ground-handling. In some of my in-flight experiments I pushed out on the control bar to bring the angle-of-attack to near stall. My recollection is that in some cases I found a very weak dihedral effect and in other cases I still found an anhedral effect, or could not determine the direction of the effect. Meanwhile the dihedral-like response that we can observe during ground-handling is, at least on some gliders, not limited to the trim angle-of-attack or higher angles-of-attack-- I believe that we can also see it on some gliders even if we are pulling the bar slightly in from trim. Assuming that we are not holding the glider in a stalled angle-of-attack during ground-handling, I think it's fair to say that for some given range of angles-of-attack we can see a dihedral-like response to yaw during ground handling and an anhedral-like response to yaw in flight.)
Here is an experiment that would resolve whether hypothesis #1 is accurate and significant, or not!
In one of the experiments described above, I flew a Falcon 140 with a wingtip-mounted drogue chute and observed an anhedral-like roll response to the chute's yaw input, even at an angle-of-attack near min. sink. I was near the middle of the weight range for this glider-- my hook-in weight was probably around 155 pounds. The same type of glider carrying the same or more weight could be tested to see if the upwind or the downwind wing tends to rise when the glider is held in a yawed attitude on flat ground. It would be important that the ground be flat and the wings be initially completely level at the start of the test. The wind needs to be strong enough that the glider can lift a full 155 pounds while riding at the trim angle-of-attack. If the glider is lifting the full weight of a 155-pound pilot, the pilot's feet will be off the ground and wire assistants will be needed to hold the glider in a yawed attitude. The wire assistants need to be careful that they are not pitching the glider up or down; the glider should be flying at trim in these tests. Another way to do these tests would be to use a heavier pilot and do the tests in wind conditions where the glider at trim will lift 155 pounds of the pilot's weight-- the pilot could stand on a scale to confirm this. In this case since the pilot's feet will touch the ground, the pilot could be creating the yaw torque and the wire assistants might not need to exert any force at all on the glider. If through experiments like these, it can be demonstrated that the "upwind" wing tends to rise even on flat ground and even when the glider is lifting the fill weight of a mid-range or heavier pilot, we can reject hypothesis #1. On the other hand if the upwind wing tends to rise when the glider is lifting less weight but the downwind wing tends to rise when the glider is lifting more weight, and the glider is at the trim angle-of-attack in both tests, this would seem to support hypothesis #1.