This I Know is True, Maybe..

This I Know is True, Maybe..

(Some Ideas About Single Line Kite Stability)

Kite stability is enabled by contriving a kite's C of G  ( the notional point at which weight forces act to pull it downwards) to be below it's C of L (the notional point at which aerodynamic lift forces act to lift it upwards).

This is the primary requirement, in the absence of which sustained flight is not possible.

There are, however, many other relationships which to varying degrees contribute to or diminish kite stability. 

A useful way of thinking about kite stability is to divide instability into two broad types.

The first can be called:

Volatile Instability.

This is the well known tendency for a kite to oscillate from side to side in a figure eight pattern of ever increasing amplitude until it dips low enough to hit the ground or begins looping as a prelude to ditto.

The driving mechanism behind volatile instability is simply that aerodynamic forces increase with the square of the wind speed rather than directly proportionally.  When a kite rotates around it's A of R , by even just a degree or two (as will happen often in any wind that is less than 100% steady),  the extra lift forces generated by the side of the kite that is moving faster than the actual wind speed far exceed the loss of lift from the side that is momentarily moving slower than the actual wind.  Once initiated this rotation will continue, driven by this aerodynamic effect, until something intervenes to stop it.

One such intervention is the primary kite stability pendulum effect of having the kite's C of G  acting at a point below it's C of L.  Unfortunately the resistance offered by this pendulum effect is very small until a significant amount of lean has occurred so the rotation can become well established before much resistance is offered by the pendulum effect alone.  By the time the pendulum has developed sufficient moment to slow and stop the aerodynamically driven rotation the kite will have rotated further than would have been necessary to adjust to whatever minor change in wind initiated the rotation.  In the process of correcting from this over-rotation, aerodynamic forces will again force the rotation past it's central position and so on in what can become destabilising oscillations of ever increasing magnitude.

But worse, rotational oscillations initiate sideways oscillations that can be even more destructive of stability. 

Whenever a kite's C of G does not lie in a vertical plane through it's C of L, that is when it is leaning a few degrees, the kite will begin to move sideways.  Once begun this movement will continue until one or both of two effects act to limit it.  If the kite's lean reverses- that is it gets itself pointed back in the opposite direction, sideways movement will slow, stop and then also reverse direction. Eventually however if the kite moves a long way to the side, the kite's line will develop such an angle to the true wind direction (when viewed from above or below) that a component of line tension will act to resist further sideways movement.  This effect will occur sooner (that is after less sideways movement has occured) if the kite's line is shorter.

Unfortunately transverse movement causes the apparent wind speed to increase, in turn increasing the lift force acting on the kite- and not just directly proportionally but by the square of any apparent wind speed increase.  As for the primary oscillating rotations, this amplified increase in lift forces can energise side to side translational oscillation and will do so unless elements of the kite's mass are positioned so as to minimise the development of destructive harmonics, and there is sufficient damping to control any remainder. 

.

It might be thought that increasing the spacing between a kite's C of G and C of L (by, for example, adding weight at the tail end) should improve a kite's resistance to volatile instability. Making the pendulum longer (by, for example shortening the front bridle lines to move the kite's C of L  upwards) does make available a larger force moment to correct any rotational perturbation but this also makes correction from any angular disturbance faster. Unfortunately, faster angular rotation generates aerodynamic forces that will tend to cause any such rotation once started to over-correct- because these forces increase with the square of the velocity as described above.

Those with knowledge of applied mathematics might suggest here that the rate of correction won't actually change by increasing the distance between a kite's C of L and C of P because the natural period of such a pendulum is fairly independent of length (the moment of weight force increases but inertial resistance to this also increases).  However, inertial resistance and aerodynamic (usually drag ) forces both act to slow the rate at which the kite operating as a pendulum can correct any angular disturbance.  The aerodynamic effects do not increase proportionally when the distance between the kite's C of L and C of G is increased and therefore the kite's response rate to angular disturbances will be faster when the C of L, C of G spacing is increased (the restoring force increases more than resistance to it).

As counter-intuitive as it might seem therefore, kites will generally be less volatile unstable if the distance between their C of L and C of P is reduced because this causes them to correct more slowly from any angular displacement that occurs

The effect is best known to kite makers by way of shortening the back bridles as a cure for minor volatile instability.  This shifts the kite's C of L rearward while leaving the C of G unchanged hence shortening the distance between them and reducing the available correcting moment. 

Unfortunately this change, if used to extreme, can cause other undesirable changes in a kite's characteristic so is rarely a total answer when volatile instability is severe.  Increasing a kite's angle of attack (by changing relative bridle lengths) sufficiently to slow the rate of correction from minor disturbances down to where it doesn't set off aerodynamically driven over-correction will often reduce light wind performance to an unacceptable extent, cause stalling or cause the kite to generate more line pull than is acceptable (big kites).

Another possible solution is to shorten the distance between the C of L and C of G  by adding weight to the nose of the kite, (thus moving the kite's C of G nose-ward without changing the C of L).  Although this will reduce volatile instability, most kite makers are not willing to accept the increase in overall weight of the kite that this causes.  To be sufficient, the extra required may be as much as 30% of the kite's weight.

For some styles of framed kites it is possible to re-arrange the frame placing so as to move the kite's C of G noseward without increasing the total weight at all.   For example moving a delta kite's spreader forward does exactly this and is the primary means of eliminating volatile instability for this style of kite.  No such change without increasing overall weight is usually possible for soft kites unfortunately.

The above can be stated as:

The first principle of volatile instability:

To the extent that it can be contrived within the constraints of whatever shape is desired, for resistance to volatile instability, the kite's C of G should be as close as possible to it's C of L

It is true because it reduces the speed of recovery from any angular disturbance and hence minimises the possibility of aerodynamically driven oscillations developing.

There is one exception to this general rule (that lengthening the distance between the kite's C of L and C of G will cause volatile instability to get worse) but it is extreme and never entirely satisfactory.

It is to add weight to a kite's trailing edge, increasing the spacing between a kite's C of L and it's C of G and therefore exactly the reverse of what is usually tried.   It can work to stop volatile instability if the weight added is so massive as to overwhelm any potential destabilising forces. 

Usually when there is sufficient weight to accomplish this,  a kite will be hanging so low to the ground that it can scarcely be considered as a kite any more, but at best it will have a very narrow wind range- more wind and volatile instability will reassert itself, any less and the kite will stall.

Another approach which might also be considered opposite is to move the kite's C of L forward by removing all bridling from the rear parts of the kite.  In the extreme, only the front part of the kite then generates any lift while the bulk of the body contributes only drag - that is most of the body becomes a tail to stabilise the remaining small lifting area. It is therefore not really an exception to the principle highlighted above.

Although this approach will reduce the kite's efficiency, that is it will fly at a lower angle than if fully bridled, efficiency is rarely the aim for decorative kites- what matters is that they fly well even if only at a moderate angle.  It is therefore a useful technique to make theme type soft kites fly satisfactorily.  

To work best the connection between the front, lifting part of the body and the rear part should be  at least a bit flexible.  As a minimum condition, the after-body's angle of attack must be much less than that of the front or lift generating part.  If the after-body has appreciable angle of attack, the condition above (that only the front part of the body generates appreciable lift) is not fulfilled and the rear part will sometimes try to over-fly the front resulting in a rolling sort of instability, discussed later,  that is not the same as volatile instability.

The Second Principle of Volatile Instability.

Whenever there is a destabilising change in wind direction, a kite will respond by a lesser angular deviation if it's rotational inertia is greater- and  the smaller the angular deviation that occurs in any destabilising event, the easier it will be for the kite to regain equilibrium.

Rotational inertia is the "flywheel" effect- when a spinning ice skater pulls her arms in, she rotates faster because rotaional inertia has decreased.    For a kite, the further it's mass is from the axis of rotation, the more it will resist any force that tries to cause it to rotate.

"Resistance to volatile instability is enhanced by maximising a kite's rotational inertia".

It is true because greater rotational inertia reduces the angular extent by which the kite will be disturbed by any particular  incidence of  disturbed airflow and because, as for the first principle, it slows down the rate of recovery hence reducing the possibility of aerodynamically driven over correction.

Common  examples of the operation of this principle are the impressive resistance to volatile instability inherent in long narrow framed kites (although some of this may derive from lateral area effects)  and for very high aspect ratio kites of the delta and genki type (although extra drag effects are often contrived at the wingtips of these styles of kites and this may also reduce volatile instability.   A more direct way of testing this theory is to add small weights to the wing tips of a kite, placed in such a way as to not move it's C of G and observe the improvement to volatile instability that results.

These two principles are not just abstract theories, they signpost the physical characteristics of stable kites and the corrections that are necessary for those that are unstable in the volatile sense.

To give practical effect to these principles:

Firstly, the distance between a kite's C of G and C of L should be as short as possible, while still being sufficient to ensure that the C of L ( which can move up and down a bit as the kite's angle of attack changes) never moves to a point below the C of G  as this would cause the kite to tip over.

Secondly, the kite's mass should be disposed as far away from it's rotational axis as possible.

The greater extent to which a kite can be designed to meet these two requirements, the less other stabilising devices such as keels drogues or tails will be required to prevent volatile instability.        

For soft kites, that is,  kites without structural members and especially for theme kites (that is kites that are required to look like something whether or not this shape is stability friendly), there may be limited opportunities for changing the rotational inertia and the C of G and C of L relationship. Nevertheless, these should be considered first, notwithstanding that even if they are arranged ideally, in the absence of extra damping effects, volatile instability will still get the upper hand eventually as the wind speed increases.  This is because the underlying stabilising effects they apply are proportional to the kite's mass, which doesn't change, while destabilising effects are driven by aerodynamic forces which increase with the square of the wind speed.

There are two categories of things we can do to combat remaining volatile instability.

The first is to contrive a contervailing force by way of lateral area, (usually fins or keels).  Correcting forces generated by lateral area increase approximately linearly with the angle of swing and with the square of apparent wind velocity so can be well matched with the forces they are required to oppose. Clearly lateral area should not, at least on average, lie ahead of the kite's axis of rotation or else it will act to amplify rather than oppose any rotation that occurs.

Even then however, in acting to restore any errant rotational displacements of the kite, the aerodynamic forces generated by the angle that the keel or etc presents to the wind when doing so will also cause undesirable sideways movement.  There are two common solutions to this difficulty.  One is to keep the lateral area small (so that it's effect in pushing the kite sideways will be minimised) but to position it as far rearward as possible so as to have the greatest possible moment or leverage against the angular deviation it is required to oppose.  A risk with this solution is that the aerodynamic restoring moment generated by such lateral area can cause over correction and hence lead to destructive harmonic oscillations.

The other solution is to dispose the lateral area in an approximately balanced arrangement around the kite's axis of rotation  and make it so large in proportion to the size of the kite that major angular deviations don't have any chance to get established.  Sled kites use this approach.

Very often for  theme or decorative kites there are visual constraints on how or even if keels or other forms of lateral area can be used.

The second effect that can be used to control remaining volatile instability is damping..

For example, by way of a tail or drogue pulling on the rear of the kite.  Drag forces are ideal for this task because they increase with the square of the wind velocity, just as do the destabilising forces they are required to oppose.  In practise, all volatile instability involves some elements of both rotational and translational oscillations because any rotational deviation will cause some sideways movement-but  kites very rarely stop flying solely because of rotational oscillations -as unattractive to purists as these can be- while translational (that is, side to side) oscillations are the dominant  form of kite instability and almost always get worse as the wind speed increases.  As used for combating volatile instability, drag forces have a great advantage over lateral area, that in acting to correct rotational deviations, they need not also cause the kite to move sideways.

Just adding sufficient drag (by drogues or etc) to the rear of a kite is no cure-all for volatile instability however.  Clearly, to stop any rotational movement that occurs, drag devices have to apply their correcting force at some distance from the kite's axis of rotation because if applied exactly at the axis, rotation could occur unimpeded.  The greater the leverage drag forces have, the more effective they can be in damping out volatile instability.  The problem is that, by their very nature, drag forces oppose change- this is advantageous when the change they oppose is from steady flying to rotational or translational deviation but it becomes disadvantageous when the kite leans to one side or the other causing sideways movement.  If not correctly disposed, drag forces can then restrain the kite's ability to straighten up. 

The most extreme example of this occurs when a drogue is attached to the rear of a kite by a long line.  If the kite deviates to either side by any substantial degree, the drogue then tends to hold the rear of the kite back forcing the kite's head more and more to that side until the kite is lying on the ground at one edge of the wind.  From the kite's point of view, a big drogue attached on a long line to it's rear is no different to having the line to the drogue tethered directly to the ground some distance downwind.  In this case, while the kite does not move out of line, it will fly reliably because the restraint will limit rotational oscillations but if it does gets out of line for any reason it will then continue to move sideways and will not self correct.

A principle for drag devices on kites that tend to move from side to side is therefore that they should be as close as possible to the kite.

For kites that tend to oscillate rotationally without substantial sideways movement, drogues can be on longer lines and in some cases this is actually advantageous because the drogue then works better by being in smoother wind.

Of course, all parts of a kite generate drag forces whether we want them to or not and, it is these rather than separate external drag devices that are the most useful, and the primary line of defence against residual volatile instability.

For kites that are designed without visual constraints, the drag bi-product of lift generation can be sufficient to control any residual volatile instability without any lateral area or extra drag devices being required up to an L/D of 3.0 or so (that is, providing the inherent drag forces are no less than about 1/3 of the total lift forces).  Above this figure, even if great care is taken with mass disposition, lateral area or extra drag devices will be required to  ensure stability.

The second broad type of kite instability can be called:

"Superstability"

This is because it is really an excess of stability, notwithstanding that it can cause a kite to stop flying just as surely as volatile instability can.

Superstability is when a kite flys without any signs of nervousness or volatile instability but then leans over and arcs off to one side or the other, often very slowly, and usually without recovering from the initial lean- although it may drift back to the centre a few times and even exhibit the same behavior to the opposite side.  It is quite distinct from volatile instability in there

 being no appearance of rotational oscillations or transverse "figure eighting". 

Like volatile instability, superstability also generally gets more severe as wind speed increases.  In lighter winds and in it's minor forms it will appear as just a tendency to fly a bit to one side of the true wind direction. 

Most kitefliers think that superstability is caused by some assymetry in the kite as superstable kites will generally always go just one way or the other, not both.  In the limit, all kites are of course assymetrical but assymetry is not the prime cause of superstability.  Rather if the conditions are appropriate for the onset of superstability some minor assymetry will then determine in which direction it will occur.

There are different types of superstability, in fact the only feature all  types have in common is the pattern of miss-flying that results.

Categories of Superstability

The first, C of L position superstability is the least common type and occurs when a kite's centre of lift is at or below it's centre of gravity.  Whether they know why they are doing so or not, designers tend to position their kite's C of G and the C of L very close together because of the improvement in resistance to volatile instability that results.  The C of L is not a fixed position but moves up and down a bit as a kite's angle of attack changes.  The C of L will generally be at it's lowest point when the angle of attack is very high (that is when the kite is stalled or nearly so) and when it is very low, that is when the kite is flying at an angle above, say, 85 degrees (very unlikely for decorative kites).  In  extreme cases of this type of superstability, a kite can tip over but more usually it's pendulum effect diminshes below some minimum magnitude necessary to keep the nose pointing upwards and the kite becomes superstable.  Because it is angle of attack dependent, often this type of superstability will occur erratically and possibly only in mid range winds rather than when the wind is strong.

The second type can be called Tail Drag Superstability and has already been described in the discussion on drogues as drag devices above.  It occurs when too much drag is applied to the rear of a kite in an attempt to combat volatile instability and/or when a drogue or other drag device is connected to the rear of a kite by a longish line rather than directly.  In effect it is the same as what happens when the tail of a kite catches on something causing the kite to lean over to one side until it is lying on the ground straining at the snag in much the same way that a cat pulls away when it's tail is held.

The third type of superstability is a rather unusual and fortunately rare type that I have only experienced  twice and don't have very good understanding of.  Both times it has happened to me has been with rigid structure kites with tails that were rigidly attached (in the lateral plane) to the front section of the kite.  On each of these occasions apparently appropriately designed and bridled single skin framed kite would sometimes dive off inexorably to one side in a clearly superstable mode.  I surmise that it was something to do with their bridling and structure not allowing the kite's head to correct itself after some minor rotation ocurred.  The first time it happened was about 1980 with a Thai serpent style kite that had a fully battened tail and a multi point bridle.  The second time was about 1984 with a fully bridled Edo style kite that had a long full width battened tail.  The fix for the serpent was to change the bridling so that while still controlling the kite's angle of attack longitudinally, it did not constrain the kite's angle laterally.  For the Edo, the cure was to change the way the tail was attached to the head so as to allow  free pivoting (laterally) between them.

I often wonder why some of the soft kites we now build don't suffer from the same problem as they appear to have the required preconditions.  The reason is probably that ram air inflated structures, no matter how extensively bridled or closely coupled to similarly inflated tails, are still flexible enough to allow minor angular corrections of the head part without the entire tail having to simultaneously swing into line also.

The fourth type of superstability can be called  "frontal area instability".  A form of this was also touched on above in the section on lateral area.  If a kite is built with a large keel or keels so that there is more moment area of keel surface upwind of it's A of R than downwind, as soon as any minor rotation occurs the extra area of keel at the front will force the kite's nose around even further and superstable behaviour will result.  Another form of frontal area instability occurs when kites are built with large wide blunt noses- Parafoils are susceptible to this.  In this situation,  when a small rotation occurs around the kite's A of R the front or nose of such a kite will then apply it's drag force to amplifying this movement, the opposite of a self correcting effect. In such a case drag applied to the  rear end of a kite would assist the kite's pendulum response to correct the initial angular rotation and destructive oscillations could possibly  build up but drag forces applied off centre in front of the kite's A of R work against correction.  This discourages the type of harmonic responses that cause volatile instability.  The initial minor rotation amplified by assymetric drag on the kites nose will instead cause the kite to traverse off to the side, usually until the centreing component of line pull is sufficient to stop it going further.  The kite's pendulum response will also oppose the assymetric drag from the kite's nose that keeps the kite tilted over but the drag forces are often so relatively large as to be overwhelming.  In minor cases the kite will just annoyingly  hang off a bit to one side, but remain quite steady.  In average cases the kite will go so far to the side as to lose it's lift, drift back to the middle and then do it again- but this is not at all like volatile instability though violent things can happen as it re-establishes lift and flies up each time it returns to the middle of the wind.  In extreme cases the kite will traverse so far sideways that it will contact the ground.

Frontal area instability of the blunt nose type will be more severe if the kite's nose is wider than the remainder of the kite's body and less severe if the kite gets wider towards it's tail- this is obviously true because in the case where the front is wider, a greater angular deviation is required before drag forces generated by any part of the kite's afterbody can act to resist further angular deviation.  Conversely for wedge shaped kites that are narrower towards their leading edges, even a small angular displacement will cause significant increase in the corrective drag forces generated by the afterbody.  Of course the reverse is true for volatile instability so it can be a choice of one evil or the other. 

The fifth type of superstability is caused by structural irrigidity.

For soft kites this generally appears as progressive loss of inflation as wind speed increases and is caused either by inappropriately placed inflation points or by leakage.

When inflation points are placed too far from the leading edges, as the kite climbs to a higher angle of attack with increasing wind speed, air bleeding into the kite will no longer be at maximum or stagnation pressure.  The appearance of this will be that the kite's leading edges will indent and this can cause the kite to lean over to one side because the front of the kite then generates too much drag.  In the limit it will cause the kite to descend to one edge of the wind.  However, sometimes inflation points on soft kites are deliberately placed back a bit from the leading edge to cause just this drag inducing indentation at higher windspeeds as a counter to volatile instability.

Leakage induced deflation will have the same appearance and result and occurs progressively as wind speed increases because this increases the pressure difference through the kite's skin causing it to leak more.  This happens noticeably more as fabric gets worn and is therefore also the reason why soft kites that tend to volatile instability when new behave better as they get older.

For framed kites the main cause of structural irrigidity is frame deflection.  A secondary cause is fabric stretching.   The appearance of superstability caused by either of these will be as for loss of inflation with soft kites- gradual loss of flying angle and arcing descent to one edge of the wind.  Very rarely, frame deflection with increasing wind strength will cause volatile instability instead but this is all to do with which part of the frame deflects.  Progressive frame deflection with increasing wind is often deliberately used to offset tendencies to volatile instability as wind speed increases and it is sometimes a fine judgement to find the balance between this and the onset of deflection induced superstability.

Other Types of  Instability.

There are three types of single line kite instability that are neither volatile instability nor any form of superstability.  They are not related to each other..

Tail Push Instability

This happens when a kite's after body tries to fly past the front section. If a kite is thought of as being a front section and a rear section and the lift and drag forces on each are considered separately, this type of instability occurs if the rear section is flying more efficiently (that is attempting to fly at a higher angle) than the front section. It occurs when a kite's front section is shaped and bridled so as it is almost stalling in normal flight (usually to generate frontal area drag, one of the major techniques which can be used to reduce volatile instability in soft kites that don't have lateral area to assist stability).  If the kite's rear section is then shaped or bridled so that it has moderate angle of attack, say 15 degrees or so, as the wind speed increases there will almost certainly be some set of conditions by wind speed and flying angle when the front section will stall and the rear section will attempt to push forward.  

For kites with adequate compressive rigidity between the front and rear sections (whether ram air inflated or  framed) this will present as a characteristic rolling motion when the rear section attempts to shunt the front section to either side as it attempts to get past.  If present to only a mild degree, this may not be much of a problem because the structure will just re-distribute the extra lift but it may still show as a skewing type of oscillation that is often seen on long narrow framed kites.

An example of this type of tail push instability  occured in our Trilobite kites.  Their bodies are laterally rigid enough in compression to resist buckling under this action but in attempting to push forward, the rear body would force the front section alternately to one side then the other.  In some ways this effect is similar to frontal area instability, one of the forms of superstability described above, but it is different because, as the driving force is from the rear section pushing forward rather than the front section being dragged sideways, once the sideways displacement reaches a sufficient angle, the rear section lift decreases and side forces on it to increase, causing a correction which can then start a roll to the opposite side.

For soft kites without adequate rigidity in the "neck" area, tail push instability  can be a major problem because the rear section in attempting to shunt the front section forward can cause compressive buckling of the body between the front and rear sections.  An example of this form occured in a Sumo wrestler form soft kite I built in 1998.  This persistently buckled up in the neck area causing the head to loll over to one side or the other, forcing the the kite into a relentless dive to that side.

Various possible solutions to the different forms of tail push instability.

For the Trilobites I first tried increasing the angle of attack of the rear section and decreasing the angle of attack of the front section by taking rocker out of the body until they were not so different.  No matter what bridling was then used, this caused extreme volatile instability- so extreme that a drogue large enough to control it also caused tail drag superstability. By using a set of three large "bucket" tails instead of a single drogue it proved possible to control this volatility but the tails had to be so large as to be visually intrusive and the angle of flying was barely 20degrees.  For kite shapes that are inherently more stability friendly this solution could be more satisfactory.

Then I tried decreasing the angle of attack of the rear section to nearly zero by building even more rocker into the body. The rear section then generated only drag and almost no lift and did not try to fly past the front.  Of course the kite's overall efficiency and pull was reduced a bit but this can be an advantage not a disadvantage for primarily decorative kites.  Remaining volatile instability was controlled by the addition of three moderate sized bucket tails that generally visually enhance the kite anyway.  This approach worked for Trilobites and gave them the best wind range of any designs we had produced up to that time.

For the Sumo, the answer was first to increase rigidity in the neck area by improving inflation and "bulking up" the neck dimensions- not so inappropriate for a Sumo anyway!-and then to add progressively bigger drogues until the body was held back enough to stop it trying to overfly the head.  During this process there had been a time when I thought the problem was poor inflation in the head because of leakage from the body so I had also fitted a fabric septum with controllable permeability between the neck and body sections.  Inflation was through the head so this also allowed the body to be flown with any desired degree of deflation- as a volatility stabiliser, because partly deflated sections in the windstream generate substantially more drag than tight sections.

These changes were sufficient to allowed the kite to be flown when the wind was mid range and there were no near neighbours but for festival situations a pilot kite was also necessary.  Now I know that what I should have done was to de-bridle the body and fly it with lift forces generated only by the head and shoulders, using the body as a stabilising tail rather than trying to use the entire body as active surface.  This would have fixed the tail push instability and the volatile instability in one stroke, oh well!

To summarise these solutions to tail push instability:  De-bridle the body, improve the neck rigidity, add a big drogue or, if all this is insufficient, uses a pilot kite.

Luffing Instability.

Luffing instability  is especially noticeable in parafoil style kites and others that have cambered top surfaces. It is caused by progressive attachment of flow over the top surface of a wing as it's angle of attack decreases.  For most wings, and almost all single line kites, flow becomes turbulent and detaches from the top surface somewhere before it reaches the trailing edge.  For wings and kites with cambered top surfaces, this separation point will move progressively closer to the trailing edge as angle of attack decreases.  The effect of this on a kite can be that as the kite flies to a higher angle this progressive attachment can cause the aerodynamic centre of lift to move rearward until, crudely, the rear part of the kite will lift above the leading edge and either cause the kite to luff or, in milder form cause an arcing apexing dive to one side or the other.  It is not the same as volatile instability because there is no oscillating or repetitive component (notwithstanding that if the dive arrests for some reason such as the kite flying into a low wind area, it may climb and repeat). Nor is it the same as any form of superstability because luffing unstable kites will generally accelerate as they dive while superstable sideways movements are slower and at fairly constant speed.

The usual solution to luffing instability is to move the point of maximum camber of the top surface close to the leading edge so as to discourage large migrations of the C of L with changing angle of attack. 15% (of chord dimension) is often taken as a suitably forward maximum camber point.   Open leading edges also have this beneficial effect for ram air inflated kites that exhibit luffing instability.

Rolling Instability.

This is a very specific type of instability that occurs when kites that have cylindrical bodies fly at moderate or higher angles of attack.  It is rather like volatile instability but is different in that the rotational oscillations don't degenerate into lateral figure of eighting to the same extent.  Kites with rolling instability show a distinctive rolling oscillation that can be violent enough to tip them upside down but the resulting lateral movements that these rotational displacements cause don't build up destructively as they do with pure volatile instability. I think it's caused by a phenomona called the von Karman vortex street which can be most easily observed by watching the alternate left and right handed vortices that break away from any round pole set in a river or stream.  The explanation is that as the water is split by the pole and flows around each side to join again on the downstream side, some initiating assymetry or turbulence causes the flow on one side or the other to stay attached past the 180degree point and creep up into the opposite side's territory until it goes too far and gets pushed off.  Flow from the other side then does the same thing and so on in an oscillating flow pattern that has a characteristic period.  For kites, as the separation  point migrates from side to side it causes aerodynamic forces to push the  kite rythmically to one side and then to the other.

Our Dolphin kites from around 1992 were a prime example of this.  Even in light winds the rolling became too violent to control even with large drogues.

If this instability is being caused by the von Karman effect it should be possible to cure it by fitting some sort of fence to stop the separation point migrating from side to side.  I've never tried this, possibly I'm not that keen to test the theory therefore!, but I'll stick to the rationalisation that such a fence would be visually offensive on a Dolphin.

A shape change that does help without causing visual offence is to build a lot of rocker into the body so that the nose part as far back as the side fins presents to the wind at much higher angle of attack than the after body.  If the main, (cylindrical) part of the Dolphin's body is at a very low angle of attack then airflow will pass along it rather than around it so the von Karman effect should be minimised. This is very similar to the Trilobite approach from above but for an extra reason. Trilobite "legs" are effective flow fences and anyway, their bodies aren't sufficiently symetrical for the von Karman effect to cause problems.  For the Dolphins, unfortunately, even this change does not entirely eliminate von Karman effects because no matter what minimal angle a kite is bridled to, the kite's weight alone will cause it to fly at a higher angle of attack until aerodynamic forces on the after body are sufficient to lift it to nearly horizontal.  This is a useful automatic angle of attack control system that adjusts the kites angle to near the optimum high angle of attack for light winds while allowing it to reduce in stronger conditions so that the kite doesn't become overpowering.  For Dolphin kites however, the von Karman effect is destabilising before there is sufficient wind to reduce the main body's angle of attack sufficiently to counter it. A large drogue can control the remaining instabilities but, as at 2001, only up to wind speeds of around 25km/hr.

One other cure for this type of instability is to fly two identical kites side by side, touching. The combined kite's shape is then no longer a cylinder so von Karman is kept at bay. I've used this to keep Dolphins stable in stronger winds and I've seen Jurgen Ebinghaus fly his big fish this way, for presumably the same reason- although maybe he does it because it looks cute.  There could be other reasons why this way of flying is more stable however, so this is not a dependable proof of cause.

The Very Large and the Very Small.

Although they obey all the above principles, very large kites can appear to behave anonamously because the inertial effects of air mass in internal spaces become significant.  The MegaBite weighs just 200kgm but carries 2 tons  mass of air inside whereas a 9kgm standard large trilibite has only 4kgms.  A mini trilobite has only a trivial mass of internal air.

Very small kites can also behave differently but for a fundamental reason that puts them outside many of the above principles.

The boundary layer of air around objects, that is the layer of air that tends to stick to the surface rather than flow past has the same thickness irrespective of the object's size. For normal size kites, even quite small size kites, this layer is so thin as to be irrelevant.  For VERY small kites this sticky  layer of air changes it's effective shape (as experienced by air flowing past) and hence it's behaviour fundamentally.

Lighter than Air Components.

Kites have been built with lighter than air gases in their internal spaces, as a structural component and to assist light wind flying.  The effect this has on stability is likely to be complex because the lift force generated by the lighter than air component is independent of wind velocity.  If such a kite has some residual weight, when it is flying as a kite it will obey all the principles from above.  If it has residual bouyancy instead I can scarcely imagine how it will behave, especially during the transition from being a balloon to when aerodynamic forces become significant as wind speed increases from nothing. Such a device is not really a kite- it is an aerodynamically assisted balloon and stability relationships would therefore have to be developed as a separate study- notwithstanding that they may have features in common.

Combinations and Diagnosis.

Generally kites can exhibit more than one form of  instability at the same time.

This sometimes makes it very difficult to decide what forms are actually present so that corrective action can be taken.

Also, for small kites, sometimes things happen so fast that it is almost impossible to distinguish between different types of instability.  The best way to develop an understanding of the causes and cures of instability is to observe large kites flying because movements are slow enough to allow different patterns to be distinguished.

When uncertain, a useful technique can be to try a crude cure for whatever is suspected- say by adding weight somewhere. From the effect this has it is usually then possible to identify the type of instability that has to be dealt with.

While other forms of instability can be simultaneously present, volatile instability and some forms of superstability (C of L type and frontal area type) are opposite ends of the same spectrum.  As changes are made to control volatile instability therefore superstability may appear, and vice versa.    Because, if present, both tend to get worse as wind velocity increases there can be a situation for a given kite shape when there will be such a wind speed that anything done to reduce one will bring on the other. For a kite that is perfectly adjusted to maximise the wind strength that it will be capable of flying in, stability is therefore an open door which gradually closes from both sides as wind speed increases until it closes in the middle, at which point volatile instability  and superstability will both be occuring.

This would be expected as a consequence of the stabilising forces acting on a kite being proportional only to it's mass which is constant, while the destabilising forces are proportional to the wind speed squared.  Eventually there will be some wind speed at which every kite becomes unstable but this may be above the wind strength when it will fail structurally anyway or at least exceed when sensible kitefliers pack up and head for the pub

Although sometimes counterintuitive, the above observations and theories fit with the single line kite flying I have experience of so far- but kite behaviour is very complex so no doubt at least some of it is wrong, or at best describes special cases. 

Revisions to follow!

The descriptions of superstability should probably remain qualitative only because there is no great mystery in them nor advantage to be gained by putting numbers to their causes and effects unless this is to be done as part of some overall theory of why kites don't fly- but I think such a theory is not achievable, or at least is so far out of reach as to be an unrealistic dream.

The principle relationships of volatile instability could probably be quantified and it would be interesting and useful to do so, at least for some special cases.

                                       Peter Lynn, Den Hague, June 2001 

.