This little RC model trike was the only flex-wing aircraft I tested that ended up having a net positive coupling between yaw and roll at most angles-of-attack--so that left rudder created a left bank and a left turn--indicating that the dihedral-like effect created by the swept or delta wing shape was actually dominating over the large amount of net geometric anhedral (created by sail billow) that is clearly visible in this photo. See theory section of this website for more. This aircraft's yaw-roll coupling was explored both with a rudder (ground-adjustable only, and positioned both above and below the keel in different trials), and with a ribbon streamer to increase the drag of one wingtip, with consistent results in all cases.
The rudder in hand--experimental rudder for hang gliders. The horizontal member that will rest on the glider's keel is a piece of angled aluminum; the hose clamps will securely attach the whole assembly to the keel. When the cord that passes through the eyelet at the tail end of this piece of aluminum is taut, the rudder is locked at neutral. When the cord that passes through the eyelet at the far end of the cross-arm is taut and at the full limit of its travel, the rudder is fully deflected to the left. The two rudder cords are operated in flight by giving one slack and pulling the other through a jam cleat to tension it as desired. The two jam cleats (one for each of the two rudder cords) were mounted on the glider's left down tube. All launches and landings were made with the rudder locked at the neutral position. A couple of sharp knives were readily at hand in case the cords needed to be cut for any reason. The principle followed in designing the rudder was "KISS"--the initial goal was simply the safe collection of some interesting data. The rudder was tested by mounting it on a car, deflecting it fully, and driving at 70 mph before it was attached to any hang glider. In flight, the effects of the rudder were carefully explored by locking the rudder at a very small deflection, before increasing the deflection. Objects in my right arm are counterweights-see below. The rudder experiments were very succesful in the sense that they led the author toward some new insights about the amount of net geometric anhedral that is contained in the wing of a typical flex-wing hang glider. However, because of the "negative coupling between yaw and roll" that occurs at most angles-of-attack, a rudder would not be a useful addition to a modern, off-the-shelf, flex-wing hang glider.
This is a link to a friend's photo album on a commercial site. Includes great views of the Spectrum with rudder, and the snapshot album in the lower left corner includes a photo with a good view of the turn probe with slip-skid bubbles etc stored on the launch-and-land position on the lower left rear wire. For flight this probe inserts into the socket visible at the center of the base bar. In the same photo, the 2 cleats on the left down tube are used to lock the rudder cords into place to hold the rudder either fixed at neutral, or in a fixed deflection to the left. In the same photo the 2 rudder cords are barely visible, running almost parallel to the left lower rear wire. A casette recorder is on the right down tube for recording observations in flight. Thank you Walt Stickel for the nice photos!
Rudders large and small: experimental rudder for hang gliders, miniature rudder (ground-adjustable only) on RC model trike. The objects on the left of the photo (water bottle with string, railroad spikes wrapped in duct tape) are counterweights for the rudder--the rudder was fairly heavy, so careful balance calculations were carried out before attaching it to any hang glider. Counterweights were normally hidden within the sail but are visible in the Raven photograph above.
Wingtip drogue chute--This is the largest wingtip-deployed drogue chute that I used on any hang glider. Ruler on photo #2 is 18 inches or 46 cm for scale. Much smaller sizes were tested first and this large one was only flown at low airspeeds; at high airspeeds (low angles-of-attack) the roll torque (acting to roll the glider toward the wingtip without the drogue chute) due to the glider's "negative coupling between yaw (slip) and roll" was uncomfortably large. The purpose of deploying a drogue chute from the wingtip of a hang glider was to investigate whether the "wrong-way roll torque" observed when flying with the rudder was primarily due to the fact that the rudder was exerting a sideways force on the keel and distorting the airframe, or primarily due to the way that the wing's anhedral geometry was interacting with the sideways airflow component that was created by yawing the nose of the glider to the side (in relation to the actual direction of the flight path and airflow or relative wind). The latter was shown to be the case as similar results were obtained with the wingtip drogue chutes, which would have distorted the airframe in a very different way. Space doesn't permit a full description of the methods and back-up systems used to safely deploy the chute in flight from the wingtip, and then cut it away instantly when desired. Wingtip drogue chutes were flown on all the gliders that are pictured above with rudders. #2, #3
Here are some photos of hang gliders with the wingtip-deployed drogue chutes in place, before flight. This set-up was difficult to photograph, but if you look closely, you can see the wingtip-deployed drogue chutes enclosed in small containers (socks), one or more on each down tube. The pink strings run from the drogue chutes out to to an eyelet on the wingtip and back to the down tube. When a drogue chute was pulled from its container and dropped into the airflow, it ended up trailing behind the wingtip, exerting on aftwards force on the wingtip-mounted eyelet, which created a yaw torque. To jettison the chute, the string was cut or pulled loose at the the point where it was attached to the down tube. Note that the strings are taped to the flying wires to keep them from snagging in brush during launch; it took some experimentation to find a way to do this that would hold firm, yet would release cleanly when I applied some tension to the strings before deploying the drogue chutes. (I always checked that the strings were unentangled and free of foreign debris before I deployed the chutes, or else there would be a risk that the string would not feed cleanly through the eyelet when I was ready to jettison the chutes). On many of these flights I carried 1 or 2 chutes on each side so the amount of yaw torque could be incrementally increased, and to allow multiple deployments in case a chute tended "fly" unusually low or high rather than directly behind the wingtip or in case a chute had to be cut away before meaningful data could be collected. Please contact the author for more safety-related information before attempting to replicate this experiment! #2, #3, #4, #5, #6
Turn probe (bank-angle reference wires, yaw string,
slip-skid bubbles; markings on lower wire provide a reference guide for the angular deflection of the yaw string): pilot's-eye view, side view #1, side view #2, stowed on front side wire, stowed #2
Early version of turn probe (yaw string and
base-bar-mounted slip-skid bubbles), #2, #3 (mirror was installed to view a yaw string mounted on the rear of the keel for observations of the "airflow curvature" effect).
Hang glider with "bowspirit" extension to mount forward yawstring for observations of the "airflow curvature" effect in turning flight. The "crossbar" near the aft edge of the yaw string is a reference guide to allow the pilot to better judge the angular deflection of the yaw string. #2, #3
View from cockpit of Vim Toutenhoofd's Swift on tow. Faintly visible in the top of the picture are the slip-skid bubbles I temporarily installed to explore the aerodynamic sideforces created in a slip in this unusual sailplane with minimal cross-sectional surface area (especially without the fuselage fairing). A slip-skid bubble is like a slip-skid ball (standard instrumentation on most airplanes) but moves in the opposite direction. And easier to find at the local hardware store. The funny round thing that appears to be a floating spherical object is actually the tail end of a yaw string--Vim has found that mounting a little paper cone to the tail of the yaw string results in more stable readings in this application. The yaw string itself is not visible in this photo as it is hidden behind the cone. Thank you Vim for the opportunity to fly the Swift! (Better view of slip-skid bubbles on Swift.)
Zagi RC model glider with rudder and adjustable anhedral/dihedral. In this photo, the Zagi has the approximate amount of anhedral that was needed to replicate the "negative coupling between yaw and roll" that I saw while flying the various hang gliders with the rudders and wingtip drogue chutes. At first glance this appears to be far more anhedral than was present on some of the hang gliders such as the Wills Wing Spectrum and Wills Wing Raven. See the article in the theory section about how washout and billow contribute to the net geometric anehedral of a swept wing.
Note that in this configuration, I've tried to avoid having all of the rudder's surface area above the glider's CG, which would tend to increase the Zagi's "negative coupling between yaw and roll" by a slight amount. In actual practice I found it very difficult to detect any difference in the Zagi's flight characteristics when the rudder projected entirely above the tail boom, or when the rudder had surface area projecting both above and below the tail boom--the roll torque created by the sideways rudder force acting above or below the Zagi's CG was not very significant in comparison to the roll torque created by the Zagi's coupling between yaw and roll, except at anhedral angles and angles-of-attack where the yaw-roll coupling was very close to neutral.
Variable-geometry Zagi with large amount of anhedral. In this configuration a right yaw input created a left roll response at all but the highest airspeeds (lowest angles-of-attack)
Variable-geometry Zagi with dihedral. In this configuration the Zagi could easily flown with rudder inputs alone, as per a Gentle Lady or other similar "floater" sailplane.
Planform view of Zagi--the Zagi's sweep was slightly modified (increased) to approximate the sweep (as measured at the quarter-chord line) of the Wills Wing Spectrum on which I began the rudder experiments. Most of the other hang gliders that I tested with the rudder had a similar amount of sweep as measured at the quarter-chord line--the gliders with wider nose angles also had more swept trailing edges, and the gliders with smaller nose angles also had less swept trailing edges, so the sweep as measured at the quarter-chord line stayed very roughly constant. (Compare to planform view of hang gliders: Raven)
This photo shows how the variable-geometry Zagi's wing spars are slightly offset, overlapped, and joined by a single bolt running parallel to the chord of the wing, allowing the Zagi's anhedral or dihedral angle to be easily changed.