By: Peter Lynn, February 2021

The first kites were most likely leaves, that, quite by chance, fly as kites when a line is attached to them in the right place. The discovery could easily have happened when such leaves were used to float fishing lines out to where the fish were- most likely in Indonesia (where Lokoloko leaves are still used for kite fishing), and possibly more than 10,000 years ago.

Leaf kites are a type of framed kite: Their structure is the spine and ribs, the 'covering' being the leaf's membrane.
From this beginning, framed kites then developed to use wood or bamboo structures (now carbon fibre) with a covering material - now commonly paper or fabric. Framed kites are made in a huge range of shapes and sizes, with the traditional 'two sticks and a tail' 'kite' shape still being what most people immediately imagine when they hear the word kite.

In the 1950's, another genre- the ram air inflated kite- was developed by Canadian born, Florida domiciled Domina Jalbert. Ram air kites have no sticks or rigid structures. They use air pressurised by the 'stagnation pressure' of wind flow to inflate internal spaces and resist compressive forces. Ram air style single line kites include 'parafoils' used as lifting kites through to the huge range of 'themed' designs that now populate international kite festivals. Ram air inflated kites can also be made in very large sizes.

There are also rigid kites- built of balsa wood like model aeroplanes, or more often now from foam and carbon fibre. In the 1970's I did a lot of experimentation with kites made of expanded polystyrene- they were a great way to quickly try out different shapes.

There is a fourth genre of single line kites, those that have a single soft skin, no sticks or other structural members and no internal spaces. Examples of such kites have been known since before Jalbert: Parasails (towed parachutes) are of single skin construction and fly on a single line in the same way as a kite. But parasails depend on having a person weighing 30 or more times the weight of the canopy hanging below to keep them pointed upwards, which is not practical for most kite applications.

In 2013, looking for a new challenge (that is, bored), and with a big heap of left-over fabric in unsaleable colours to get through before I cark, I set out to see whether single skin single line (SSSL) kites could be developed that would fly satisfactorily. In this I'm indulging a desire to get back to kites that fly alone on a single line in the classical sense. I'm in awe of the graphic creativity expressed in the amazing range of creatures that now populate the sky at kite events. They have bought a new purpose to the kite flying world- and require a considerable understanding of aerodynamics and kite lore in their making. But there's something special about 'back to the origins' kites that fly by themselves away up there, and I've always been a sucker for ideas that are new and unexplored. From the start of this new project, I added two additional criteria to the single skin single line principle: No added weights and no stiffeners of any kind, not even stiff fabric. Although realising that weights and stiffeners could improve reliability in gusty conditions, I felt they would interfere with my learning how to understand and control the tiny pressure differences that are all that is otherwise available to resist compressive and buckling loads- if it can be established in fact that they can be adequate for this purpose. Fabric stiffness can be a significant structural element in very small kites, and I also wanted to avoid dependence on this in any way so as to make designs fully scalable. The following is what I have learnt so far: Not what you could call resoundingly successful or breakthrough developments, but some theories and a few basic techniques and systems that can be built on should anyone else ever become similarly afflicted with the challenge of single skin single line kites.

Lift: Aerodynamic lift forces that enable kites to fly.
Drag: Aerodynamic drag forces that pull a kite in the wind direction
Lift/drag ratio (L/D); Aerodynamic efficiency; tangent of the flying line angle.
Centre of gravity (C of G): Where the sum of a kite's weight forces act.
Centre of lift (C of L): Where the resultant of lift forces act.
Centre of pressure (C of P): Where the resultant of lift and drag forces act
Recovery: How a kite recovers from a lean or lateral displacement to return to stable central flying. Weaving: Repeating lateral movements that sometimes builds to looping out.
Slow Recovery (SR): When a kite recovers too slowly or not at all from an angular displacement.
Leaning: Leaning off to one side or the other- sometimes to the edge of the wind window.
Diving over: sudden onset dive over to left or right.
Lateral area: Keels, flares and other vertical (as opposed to horizontal) area.
Span: the kite's width
Chord: the kite's length (excluding tail).
Aspect ratio: Width relative to length, technically; span squared divided by area.
Angle of attack (A of A): angle that the wind strikes a kite's lifting surfaces at.
Leading edge (LE): A kite's front edge.
Trailing edge (TE): A kite's rear edge.
Luff: When a kite's leading-edge dips under the wind, causing the kite's line to lose tension.
Stall: When a kite falls back, will not climb to its apex, or won't launch.
SSSL: Single Skin, Single Line kite.

I have previously used volatile instability to describe weaving, have changed this because weaving is more descriptive- and shorter. Snaking and sashaying were on the short list. I've also changed from using superstability to slow recovery (SR), this is to get more clarity between leaning and slow recovery (although there is still some overlap). Superstability attempted to cover both in one definition.

Single line kites can only fly if their centre of gravity (where the weight forces act) is below their centre of lift (where the lift forces act). The weight forces pulling downwards against the lift forces pushing up, points the kite upwards. If the centre of lift (C of L) is below the centre of gravity (C of G), the kite will turn over and fly down instead. This 'First Law' is a necessary condition, in the absence of which a kite can't fly. It is not a sufficient condition however, because it doesn't by itself ensure stable flying.

Single line kite stability is a complex dynamic relationship between line, aerodynamic, weight and inertial forces with the goal being to find combinations that recover from changes in wind speed and direction while keeping sideways movements to a minimum and not getting into any positive feedback instabilities. There are no formulas for this, just kites that fly and others that don't to learn from, a few qualitative principles, and lots of trial and error.

By observation, there are two main types of instability: Weaving (repeating lateral movement that can sometimes build to looping out) and Slow Recovery (SR). Weaving is caused by dynamic interactions between a kite's weight force moment acting to correct lateral displacement (from a wind shift for example), line pull, inertial reactions, and aerodynamic forces. It establishes when weight-moment initiated recovery and/or the lateral component of line tension sends a kite back past centre in a repeating sequence of increasing amplitude. Slow Recovery (SR) happens when the weight moment isn't sufficient to correct an angular displacement - that is, to point the kite upwards quickly enough to prevent it from moving off to one side or diving to the ground. Leaning can have causes such as significant asymmetry in the kite, or too much lateral area in front of the C of L, or can be a symptom of SR.

The effect of aspect ratio (AR) on stability is profound: The wider a kite is relative to its length the more slowly it recovers from being knocked off straight for any reason. This is because the restoring moment is the kite's weight x's sine of the angle of lean x's a function of the distance between the kites C of L (where an extrapolation of the kite line intersects the kite) and its C of G. Resisting this is the kite's rotational moment of inertia- a flywheel effect. The higher a kite's AR is, the smaller it's restoring moment is and the greater it's resisting inertia. Higher AR is therefore associated with slower recovery.

AR has another significant influence on kite behaviour: When a kite is correcting from a lean, its faster wing generates more extra lift than its slower wing loses (because aerodynamic forces scale with the square of windspeed). This can drive weaving as it causes the kite to accelerate and turn tighter. The kites that tend to do this most are those with aspect ratios around 1.0- I surmise because for both higher and lower ARs, the kite's rotational moment of inertia is greater by comparison to its weight moment and resists the accelerations that are driving this instability. Drag forces can damp this movement but scale with the lift forces so will tend not to provide additional damping as apparent wind speed increases.

The experience of most framed kite makers is that kites don't scale. A design that flies well with a 1m wingspan, will not fly well, or at all, if scaled down to 100mm or up to 10m. But this is because the weight /area ratio of framed kites increases with size, and changes in weight/area profoundly affect stability. Larger framed kites also flex, bend and distort proportionally a lot more than small ones, which effects flying behaviour. Above about 10m wingspan, even with carbon fibre, framed kites become either too heavy to fly in a useful wind range, or very fragile.

My experience is that single line kites do generally scale and over a very wide range- provided their weight/area ratio stays relatively constant.

The smallest ram air flag kite is 0.75 m span, 0.35sq.m, 0.1kg. The largest is 45m span, 1250 sq.m, 360kg. The mini's weight/area is around 0.28kg/sq.m, the mega's is 0.29kg/sq.m. They both fly well, and very similarly, with good recovery and no weaving. The mega is more than 3,500 x's the mini by area.

Single skin kites can also be scaled up until fabric and corded reinforcements reach their strength limit without any significant increases in weight/area- but in very large sizes would have far too much pull to be practical for kite event flying.

Apart from weight/area, there are some other properties of kites that don't scale:

As referred to above, there is a non-scaling effect in deflections- that fabric and structures are subject to higher overall loads in larger kites, so will deflect/distort/bend proportionally more. This is significant for rigid and framed kites. but isn't generally noticeable for ram air or single skin kites within fabric and cording strength limits.

For ram air kites, the mass of enclosed air increases with the cube of dimension while their area and weight scale with the square. Air doesn't have weight, but it has mass. To mitigate this, in very large sizes, ram air kites are made with no internal partitions so that their enclosed air mass ( > 5 tonnes for a 1250 sq.m flag kite) can rotate somewhat independently when the kite is correcting - otherwise its inertia would slow correction to a problematical extent.

Reynolds number (Re) effects. Re is an indicator of the onset of turbulent flow, which definitely doesn't scale- but is also only significant for very tiny kites in our case.

Surface roughness doesn't scale (a 20mm bump is significant for a small kite but less relevant on a larger one) - and has some effect I expect- has been posited as the reason that very large leading-edge inflatable (LEI) kite surfing kites fly in lighter winds than smaller ones.

And entrained air mass? When a kite moves it carries some mass of air with it (not just the boundary layer) and this mass is a cube function of dimension, not square - how significant is this for scaling?

Lateral area (keels, flares etc) is useful for stability, and a small amount is necessary to prevent side slip, but if most of the lateral area is forward of a kite's centre of lift (defined by where an extension of the flying line would intersect the kite) then obviously the kite will not fly straight because as soon as it leans off slightly one way or the other, wind forces striking this forward area will push it even further off straight. This is easy to understand by thinking of a kite's lateral area as a wind vane pivoted around the flying line - if all the area is forward of the line the kite will attempt to turn around and fly backwards.

All lateral area acts to slow angular recovery, and the further it is from the kite's rotational centre (usually somewhere between its C of G and C of L) the greater resistance it will offer to the weight-moment attempting to cause a correction.

But kites with lateral area that is large by comparison to their lifting area- like box kites, snowflake kites and tetrahedrals for example- can still fly straight. These styles do sometimes find a stable position leaning off to one side or other in light conditions, but soon come into line when there's appreciable wind. There are clearly other factors at play here and I surmise that one is rigidity- soft kites seem much more susceptible to this problem than more rigid kites. Another might be bridle point. Perhaps lateral area as a cause of leaning is less of a problem when the bridle point is well forward?

In light winds, when a kite develops a lean for any reason, wind no longer flows parallel to its keels or flares, and if these have more area to the rear than forward, wind striking these areas from one side can stop the rear of the kite from dropping back down to allow the kite to point upwards again. The above photo is of a ram air pilot kite hanging off to the right with the rear parts of both flares holding the kite's trailing edge up- it would hang off to the left just as happily. This is only a problem for these Pilots in very light winds.

When lateral area is large, and mainly or solely behind a kite's centre of lift, the effect can be extreme and prevent recovery from any tilting at all. A single skin Serpent's head with no tail and all its lateral area towards the rear will not fly straight for even a second; it dives over to one side or the other immediately on launching, requiring quick reactions even to get a photo. With these rear flares removed, the same head will fly stably, and centrally, even without a tail. SSSL sleds, with lateral area that is full length and nearly twice their lifting area, do stay up, but tend not to fly straight. For lateral area not to be a cause of leaning and slow recovery, it should be kept as close to the kite's centre of lift as is practical and be no larger than is necessary to prevent weaving instability.

Kite tails are very interesting devices. Considering only ribbon tails for now, they are self-supporting in all except the lightest winds (and will usually pull up on the rear of a kite in stronger winds). They add drag and shift the kite's C of G rearward (by some function of their weight and an inverse function of their length, I suspect). Depending on how they are attached (see discussion of this in the Serpents section), tails can usefully improve the stability and reliability of SSSLs- though "just add more tail' is not the answer to all stability problems, because very long tails will often cause slow recovery (SR), and on occasion I have seen tails cause weaving rather than damping it (which is more usual).

Some kitefliers appear to believe that drogues (buckets of air), which have negligible weight by comparison to their drag, are the equivalent of tails. They are not. Drogues are only suitable stabilisers for kites that are strongly weaving unstable. Any kite that is even slightly inclined to slow recovery will be worse with a drogue added.

Attaching tails (or drogues) by a "Y" bridle can also bring on SR- because Y bridles limit the ability of a kite to correct and fly straight. With this correction slowed, kites will tend to move a long way sideways before they get around to pointing upwards again.

Long tails can also have a bad effect on kite behaviour. I've done quite extensive testing with tails on some SSSLs and have found, invariably I think, that any longish ribbon tail has a better calming effect on a kite when it is attached doubled than as one length- and this is true whether a kite is inherently weaving unstable or SR unstable (which supports the above theory that tails' C of G contribution is proportional to their weight x's 1/length). If tails are VERY long in relation to a kite's size, they function rather like their end is staked to the ground- which is OK while they are exactly in line with the kite's line. But even a tiny wind shift causes a component of the tail's drag to pull the kite sideways, which makes that component larger and so on until the kite is lying on the ground at the side of the wind.

By a similar mechanism, tails that are heavy by comparison to the kite's head and attached so as to prevent the head's independent rotation, cause a kite to fall right off to one side or other of the wind window when the wind drops. As the kite stalls, the tail will always be slightly to one side or the other, which will cause the head to lean over and slide off to one side as it descends. Big SSSL serpents are a devil for this, only partly mitigated by automatic bridles that make stalling less likely.

Automatic bridles save having to bring kites down for adjustment when the wind changes. For most framed kites they are also used to reduce kite pull in stronger winds. They usually do this by spring loading the rear bridles, though the system developed by Rudolph Grund for Lindenburg's box kites in the early 20th century was more sophisticated; it changed the angle of incidence between front and rear cells. Soft kites (ram air and single skin) use auto bridles for a different reason; to reduce the kite's angle of attack in light winds so as to extend its wind range. For this, the simple approach is to spring load the front bridles, which need to be shorter when the wind is light and longer for stronger winds to prevent leading edge collapse.

Unfortunately, although this approach works satisfactorily for kites with substantial form rigidity- like Peter Lynn Kites OL Octopuses, for SSSLs with single skin fabric leading edges, it's usually unsatisfactory, because in gusty conditions the leading edges bridles can pull in when you don't want them to, causing a luff, and collapsing the kite.

Simon Freidin (Melbourne), who has made a series of 1Skin style SSSLs in sub 1sq.m sizes, has been able to use this approach successfully though. He basically just attaches the centre leading-edge bridles to light bungy cords (with a stretch limiter), so that in strong winds the bungies are fully extended and as the wind becomes lighter the leading-edge bridles progressively shorten. Perhaps the difference is in the size of the kites- for 0.75sq.m I Skins, fabric stiffness provides appreciable chordwise stiffness, but doesn't for the much larger sizes I generally work with (up to 30sq.m lifting area so far).

The first system for SSSLs I tried used a rigid bar with bridles spaced along it and the flying line attached to one point only: (I know, I know, cheating again by using rigid elements- but this was just to check the idea). The principle behind this is that when a kite is flying at high A of A (while being launched for example) its C of L will be more to the rear than when it is flying at a low angle of attack. With the correct line attachment point, this C of L migration should be able to be used as an automatic adjuster. I also tried angle limiters and bungies between different points on the lever and 200mm or so down the flying line to get different effects. While in theory, practice and theory are the same, in practice they're not. I'm not sure that any of these lever systems actually provided useful automatic adjustment- maybe a little for Serpents.

The next approach I tried uses a pulley, a bungy, and extension limiters in such a way that when the main kite pull extends the bungies, this progressively lets out the front bridles. By using the total pull as the actuator, the luffing collapse problem in gusty winds is avoided. This style of adjuster has worked successfully for every style of SSSL I've used them on.

A later system (early 2020) separates out the leading edge and trailing edge bridles and attaches them together to a bungy from the main line attachment point, while leaving all the other bridles connected directly to the main line attachment point. This system retains the above pulley system's anti-luffing characteristics by using the rear bridle tension as a proxy for the total kite pull, while being much simpler and using lighter bungies. It works particularly well when the rear bridles are of the aeolian type (described in the Serpent section)- because the tension in aeolian bridles increases at a faster rate than wind speed squared and increases as A of A decreases.

Another possibility (Feb 2021) is to attach all the bridle pairs to the flying line via a catenary loop, with a bungy on its forward leg. This should have the advantage that it won't only change the relative lengths of the front and rear bridles but have some effect on adjacent ones as well. It will also allow all primary bridles to be cut to the same length, with required differences set by secondaries connecting the primaries to the catenary loop.

I could get single skin single line kites (SSSL's) to fly almost from the start, but "fly' at first meant staying up briefly and only in a very narrow wind range. For many of the first 20 or so prototypes, the window between stalling (when the kite falls back, won't rise) and luffing collapse (when the kite's leading-edge buckles under and won't hold its shape) was vanishingly narrow. One reason for this was that I had based their leading edges on single skin traction kite practice- which is to have a lower leading edge, typically supported by partial ribs and projecting back as far as 10% or more of chord- as shown in the photo of SSSL10 above (LE to the right), which has a lower leading edge of around 5% of chord. This type of leading edge is very angle of attack sensitive.

Another was that to keep the leading edge from buckling in, I had found it necessary to hook the kite's trailing edge downwards- also visible in the above photo. This can be thought of as 'capturing' enough pressure to keep the leading edge 'inflated'. Unfortunately, it also moves the kite's centre of lift rearward, which causes stalling and stability problems.

But even with these two constraints it was possible to make useable SSSL's. The SSSL 10 Sled style was developed further by adding a central rib and flown at various events in Asia and Europe during 2014.

Robert van Weers had also made at least one successful SSSL by this time, also incorporating a lower leading edge and hooked trailing edge.

But these early efforts had narrow wind ranges and even then, required multiple bridle settings. They were also subject to 'diving over'- suddenly diving to the ground from what had been stable flight with little or no warning, a phenomenon that was to plague my next 4 years of development.

Later in 2014 I began to work on eliminating the trailing edge hook-down that was the prime cause of the narrow wind range problem. That 'open' trailing edges are possible for single skin kites was by then known from single skin traction kites. A way they do this is to use more chordwise camber than the radius described by their bridles. Say the leading and trailing edge bridles are 3m long but the kite's skin is cambered to a 2.5m radius, then some (small) component of these bridle's tension will stretch the fabric out chordwise and counteract fabric compression. In practice, even bridling of the same length as the camber is effective at resisting skin compression. Reasons for this will be addressed later in Pocketing.

Using the uniform camber principle (no trailing edge pull down), the 7 cell Boomer style SSSL was soon flying- followed by 5 cell 1Skins and 4 cell Singers.

While retaining the same leading-edge construction as earlier designs, the use of chordwise camber instead of trailing edge hook-down provided a much wider wind range. Their excellent lift to drag ratio (L/D) gave them a 60degree plus flying angle and the usual pull/size advantage that single skin kites have over ram air designs gave a 3sq.m Boomer lift equivalent to a 10sq.m or larger ram air pilot. At this stage it seemed that intensive detail development would eliminate their remaining problems, which were:

• - A tendency to lean to one side or the other rather than flying straight.
• - Requiring multiple bridle settings for different wind speeds.
• - And sometimes diving over unpredictably when flying in winds above about 30km/hr.

I spent much of the next 3 years developing and refining the Boomer/1Skin/Singer style SSSL's, building close to 100 prototypes.

The leaning problem was addressed by reducing aspect ratio. 5 cell ISkins are much less inclined to leaning and 4cell Singers have the opposite problem- weaving instability, requiring a tail to quieten them down. The relationship between aspect ratio and leaning, slow recovery and weaving is explained more fully in the section on Aspect Ratio (AR).

The need for multiple bridle settings was largely solved by the development of automatic bridle adjusters. This is covered in Automatic Bridles.

Diving over proved to be a much more intractable problem. The following are the possible causes of 1Skin diving over that I considered (in no particular order):

1. That the C of G is too close to the C of L to provide sufficient restorative moment as the wind strength increases: SSSL's are very light. A 3 sq.m 1Skin weighs just 300gm- which is all the restoring force that is then available when the kite leans over- and in a decent breeze will pull 40kg and more on the line. Additionally, SSSLs generally have their C of L's at 35% chord or more (by comparison with sticked and ram air kites at 15% to 25%), which doesn't give the small weight force much leverage to pull the kite straight when it leans a bit. It's easy to imagine the tiny restorative moment available being overwhelmed by aerodynamic forces that can be 200 times as large.
2. Bridle stretch. When a kite flies on any sort of a curve, its outer wing tip generates more lift than its inner tip by a square relationship. This creates more pull on the outer bridles and if they stretch significantly will cause the kite to dive over even more because of positive feedback.
3. Asymmetric fabric stretch. Kite fabric when heavily loaded, as it is by SSSL's in strong wind, often does not stretch symmetrically because fabric's warp threads don't have the same spacing as the fill threads and they are often not at right angles to each other.
4. Migration of the kite's centre of lift towards the rear as it's A of A reduces in stronger winds. This is a speculation that as the kite's A of A approached zero (high flying angle), flow was attaching progressively to the upper surface, rather than separating in a turbulent wake, causing the kite's centre of lift to shift back behind the kite's C of G and causing it to dive over and fly downwards.
5. Chordwise buckling. When flying, 1Skins often show spanwise crease lines (see Boomer photo). These are sites where further buckling is likely to occur under compressive loads. If these develop at all asymmetrically then they can, and eventually will cause diving over.
6. Fabric creases. For strong wind flying, 1Skins generally require re-tuning for straight flying after being packed away. This seems likely to be because the creases from tight packing are in different places each time and have an initiating effect on where compressive buckling occurs when flying.
7. Leading edge buckling (indenting) in stronger winds and at lower A of A's

Addressing these possible causes of 1Skin diving over in order:

1. Insufficient restoring moment because their C of L is too close to their C of G.
   That they are so close together is a consequence of the camber that SSSLs have to use in order to prevent compression buckling. That this does contribute to diving over is undoubted. A first attempt to get the C of L further forward was to let off the rear bridles by 30mm or so, and this definitely helped, though it also allowed the trailing edge (TE) to buckle upwards in stronger winds (visible in the Boomer photo) which caused a lot of flapping and wrecked the fabric after a few hours flying. The next approach was to reduce aspect ratio (AR)- a move from the 7 cell Boomer to the 5 cell 1Skin. Reducing AR has the effect of moving the C of L and C of G relatively further apart, giving the restoring moment of the kite's weight more leverage. Taking this further with the 4 cell Singer proved to be a step too far, Singers are weaving unstable without a tail. Singers still crashed out in strong winds too, though at a higher average windspeed than 1Skins (which are better than Boomers). The next attempt was to try kites with porous rear skins. The idea behind this is that leakage through the rear section would shift the centre of lift forward while keeping the C of G in the same place. A disadvantage was that the effect seemed to be greater in lighter winds- the opposite of what was optimal. At a small cost in light wind flying, they did seem to be a marginally better in strong winds, perhaps because of having more drag at the rear keeping the skin tight, or it could have been the desired C of L shifting effect- or maybe I was just being hopeful.
    
   

   1Skin 77 with porous rear section, November 2016
    
   An obvious way to increase the distance between the C of L and C of G is to add weight to the rear, but doing this increases the kite's A of A, which moves the C of L rearward, so there isn't as much gain as may be expected- and besides, I've foresworn adding weight (except for testing theories). A better way to shift the C of G rearward is to add a tail (see Tails section), and ribbon tails do improve the strong wind reliability of 1Skin type SSSLs. Their disadvantage is that they are additional weight to pack (if air travel is necessary), can catch on things (like little children and adjacent kites) and make launching over water more difficult- which limits the usefulness for towed displays.
2. Bridle stretch.
   I tried different weight polyester bridles (stretchy) and Dyneema bridles (not so stretchy) on various prototypes but couldn't be sure of what effect this was having. Diving over didn't decrease dramatically, even with hard finish Dyneema. There are also the chordwise effects to consider. Because there is more tension on each flare bridle than on any leading-edge bridle, bridle elasticity could be expected to exacerbate any luffing tendency in strong winds, by pulling the leading edge down (relatively). I have seen the deleterious effects of stretchy bridles on larger ram air kites but not yet unequivocally on 1Skins - but still tend to use Dyneema bridles, just in case there is something disadvantageous going on.
3. Asymmetric fabric stretch.
   Every 1Skin I made flew differently and required some tuning for straightness, even after bridles had been re-set to exact lengths after a settling in period. This can only be because of sewing asymmetry and fabric asymmetry. These could be contributing to diving over, but my sense of this is that other things are causing the diving over while the kite's basic asymmetries just decide which side the kite is eventually going to dive off to. Even perfectly tuned 1Skins can still misbehave in strong winds.
4. C of L migration at low A of A's.  To test this, I tried flow separators at around 1/3rd chord on the upper surface. There was no noticeable improvement.
    
   

   1Skin 67 with turbulators to cause flow separation, July 2016
    
5. Chordwise buckling.
   As explained earlier, provided a kite's camber is of lesser radius than the bridle lengths, some small component of bridle tension is available to keep the skin unwrinkled- in which condition it's best able to resist compressive (buckling) loads. But as using more camber causes the C of L to move too far towards the rear, there's a fine balance to be struck here (see 1 above). Another way to reduce chordwise buckling is to apply some chordwise tension to the fabric by adding drag at the rear of the kite- like by adding a tail or tails. Doing this by hooking the trailing edge has been ruled out but hooking the trailing edges of the outer flares could accomplish the same thing (and is used on later SSPs). However, 1Skins that have substantial tails still sometimes crash out in strong winds. So, though asymmetric chordwise buckling definitely is a cause of 1Skin diving over, it's not the only cause.
6. Fabric creasing.
   Softer (non-creasing) fabric would be an answer, but such fabrics are stretchier and less dimensionally stable, so would cause other problems. Basically, that this is a cause of 1Skin asymmetric flying, tells me that the design is too sensitive. Successful kite designs need to have a margin of reliability sufficient to cope with aging, minor damage and small left/right differences.
7. Leading Edge Buckling.
   Wind forces are inevitably going to push the leading edges of SSSLs in when they are flying at low A of A (high line angle). This is because the leading edges of SSSL's must extend down significantly or the kite will stall whenever it has to fly at high angles of attack - like in light winds. But this buckling-in doesn't seem to prevent 1Skins flying reliably in moderate winds, and even sometimes in very strong winds (I have at times managed to keep them up in winds of above 70km/hr for short periods). The question was therefore, why do 1Skins sometimes cope with stronger winds and sometimes don't- and is this a leading-edge problem or something else?

As a way to maybe answer this, I tried fitting double skin leading edges to 1Skins, with the space between ram-air inflated (I know, I know, cheating- and I tried the same check on an Octopus and Serpents too). The results were equivocal; definitely more reliable, but not a complete cure.

Simon Freidin (Melbourne) tried another idea to reduce leading edges indenting; Air dams. These are fabric dams across the kite's cells positioned so as to 'bounce' some pressure back to the kite's leading edge, in order to keep it 'inflated'. They do seem to do what they are supposed to do to an extent. This is something that could probably be usefully modelled in a virtual wind tunnel to optimise dam placements. 1Skins fitted with air dams still sometimes dive over when flying in strong winds.

Conclusion; Asymmetric leading-edge indenting IS a primary cause of 1Skin diving over, and nothing tried so far is a sufficient cure. Occasional satisfactory strong wind flying is almost certainly because on these occasions, LE indenting just happened to be symmetrical enough not to make problems.

When the wind becomes strong, the 1Skin style of kite will sometimes dive off to one side or the other, taking other kites in the vicinity down into a big tangle. The causes are a various but are mainly inherent in their rib supported leading edges, which generally don't 'pop out' again after they have indented. When this buckling-in happens asymmetrically the kite dives off to one side and likely crashes.

Although sometimes flying outstandingly well, especially in light steady winds when their high lift for size is useful for lifting line objects, Boomers, 1Skins and Singers remain generally unsuitable for use as pilot kites at busy kite events. But what I discovered during their development has been useful, especially the failures, a story of persistence unrewarded, but improved understanding.

Early in 2015 I started to try a few single skin ideas for themed kites. The first SSSL Ray has a bad case of SR (slow recovery), will barely stay in the sky long enough for a photograph- even with a heavy rope tail. By 2020, I did get Ray style SSSLs to fly relatively satisfactorily in mid-range winds, but only at the cost of reducing aspect ratio to the point that they look more like Indian fighter kites than Rays. (see section on Aspect Ratio (AR) for an explanation of its effects).

For SSSL Rays, the effect of AR is exacerbated by the difficulty of keeping their wing tips from buckling in. For this they require more depth and more bridles at the tips than is ideal, which adds drag out there, further slowing recovery. SSSL Rays also have the additional problem that when a lower wingtip collapses a bit, the opposite higher tip than drives the kite further over rather than helping recovery. But, having had such a long history with Ray kites (my first ram air inflated Ray kite design was in 1988), this is a style that I'm not giving up on just yet.

I had thoughts that the head would be the kite and that the body would function as a tail for flying alone. In practice, the bridles to the legs made the body buckle up rather than stream out flat. It's therefore just another wannabe kite that became a line object, having lift but no stability and requiring a pilot kite to keep it pointed up (a 1Skin in this picture).

Early in 2015 I tried an SSSL Octopus- Like Rays, a style that I've had a long association with, from small, framed octopuses which were the foundation of our business in the late 1970's, through to large ram air versions from 1990. It immediately flew well.

After a few more prototypes and a lot of fiddling with bridles, these SSSL Octopuses began to show the ability to recover from asymmetric leading-edge collapses- which are generally terminal for 1Skins. When the wind is too strong for a particular leading-edge bridle setting on these Octopuses, one side of the head collapses a little, then the opposite side and so on. These brief collapses, which happen on the right-hand side when the kite is tending right and on the left when it is diving left, cause the kite to immediately correct back to central flying. Of course, if the wind then becomes far too strong for wherever the front bridles are set at, the kite's leading edge will eventually collapse, taking the kite down- but this can be addressed by fitting an automatic bridle adjuster. The self-correcting response occurs because the highest apparent wind flow strikes the side that the kite is veering off to - which causes the head's leading edge to buckle in (indent) on that side. This buckling slows the kite's sideways movement but does not in itself correct the lean that has by then developed or send the kite back to the centre. Correction occurs because wind flow is not now parallel to the centre line of the kite but slightly transverse, favouring the side that isn't indented. This generates lift that sends the kite back towards the centre. Essential to this mechanism is that there are no keels or flares in the centre area of the kite to prevent the action of this transverse flow. As I've gradually learnt how this works and how to design for it, collapse resilience has become a key feature of later SSSL's.

SSSL Octopuses (and their Serpent siblings) have a lot of bridles- up to 60, but all of these except the ones around the leading edge and across the rear of the head are the same length, which simplifies bridling a lot. Given the "bridle lengths should be no shorter than the camber radius' theory described earlier, this is no great surprise, but it is surprising that it is so optimal- and believe me, I have tried many minor variations. That the bridles don't need to be longer than the camber radius to stop chordwise (and spanwise) buckling can be explained by the beneficial effects of pocketing, which is described later.

Unfortunately, single skin octopus tentacles tangle. After flying in mid to strong winds, sometimes they take hours to untangle. Many anti-tangling ideas have been tried: Various tapers, pointed ends, flat ends, stiff fabric, soft fabric, silicon coating. The tentacles of small, framed octopus kites don't usually have a tangling problem, so I presume that it's a function of size and relative fabric stiffness. Nothing (so far) works for the SSSL 20m Octopuses except linking the tentacles together with cross cords (as in the photo above)- leaving just enough free sideways movement so that they don't function as one wide tail (see later discussion about tail width effects on stability).

Doubly unfortunately, linked tails are a menace- they catch on every available obstruction, especially during launching, which can be dangerous. This has made SSSL Octopuses impractical for general festival flying except in very controlled conditions.

Apart from this, they have developed into what is still one of the best all round SSSLs I have yet made, with a wind range from less than 10km/hr to more than 70km/hr and stable flying even in gusty conditions.

To get around the Octopus tentacle tangling problem, from June 2015 I tried using the Octopus style head with a single tail - Serpents.

An immediate problem was that they became weaving unstable from about 20km/hr. I then spent years trying to lift their wind range, trying variations in bridling, but also different tails- MANY different tails. None of which worked, because I was mistaken in thinking that the early onset weaving instability was because of insufficient tail drag. It wasn't. The problem was with the way that the tail was attached to the head. Serpent style tails are attached full width and taper from there to the tip. This doesn't allow the head to correct from a lean except by pulling the entire tail around with it, causing slow correction which manifests as weaving instability in stronger winds. Many attempts at rigging Serpent tails to heads in ways to allow some independent movement failed. And yes, some used bungies and I tried leaving a gap. I also tried slitting the tails in the section nearest the head to mimic Octopus tentacle attachment, but this caused SR instability because they generated too much drag at the slit ends, unless the slits were extended full length- when they tangled of course. Many of these attempts weren't graphically acceptable anyway.

Eventually I made two 2m wide Serpent heads and compared them with tails attached in different ways. This quickly showed that a Serpent head will fly quite well without any tail at all- though they're inclined to lean over because their C of G is a bit close to their C of L, which adding a tail mitigates. The tail's job in this case is not to provide drag, but to move the kite's C of G rearward.

But by then, the upper wind limit for the 30m Serpents had been lifted to around 35km/hr by fitting aeolian bridles, which was good enough for some event flying.

These were developed for the 1Skins, on which they were only marginally useful- but for Serpents and other SSSLs with tails, they extend the upper wind limit by a lot:

Aeolian bridles are named after the Aeolian or wind harp. They are a set of braids that are used in place of a kite's rear bridles. As wind speed increases, they vibrate with increasing amplitude, pulling down on the rear of the kite and increasing its angle of attack (A of A). That they do this at a rate greater than other aerodynamic forces increase with (which is with the square of the wind speed) is a key characteristic. Another is that, for a given wind speed, they generate the most pull-down tension when the kite is at its lowest angle of attack. This is because when the kite is at its highest A of A (during launching) the rear bridles are close to being parallel with the wind, while when at its highest A of A they are almost at right angle to the wind. This is the desired relationship for reducing leading edge collapse in SSSLs.

However, while all SSSL's require their bridled A of A to be increased progressively in stronger winds to prevent leading edge collapse, doing this by pulling the trailing edge down tends to cause SR type instability for tailless SSSLs like 1Skins. This is because a cause of SR instability is for the kite's C of G to be too close to its C of L - as has been explained earlier. Pulling the trailing edge down does not usually cause this problem for SSSLs like Octopuses and Serpents that have substantial tails (around 75% of total kite weight), which shift their centres of gravity rearward.

For some reason, before making the 2m wide Serpent heads I had decided to make a double sized SSSL Serpent (4 times the area, so 4 times the pull). This wasn't a good idea at the time because there was every reason to expect that its upper wind limit would be the same as for smaller Serpents, around 30km/hr, which is marginal for many kite events. This is because, single skin kites scale- that is, their stability responses remain relatively constant when they are made in larger or smaller sizes, as explained in the Scaling section.

To my very great surprise, this scaled up Serpent is stable in any wind speed I've yet had the courage to fly it in- which has been to around 60km/hr so far. Only at this maximum have I seen even the slightest beginnings of weaving instability. I don't really understand why 6om Serpents are so stable, as they have the same weight/area ratio as 30m Serpents and none of the minor exceptions to the scaling rule seem sufficient explanation. What could be causing them to recover from an angular or lateral displacement without getting into the escalating series of overcorrections that plague smaller versions? I had thought it was the way the tail had been cut to take tension down its centre rather than along the edges, but this was thoroughly de-bunked by the next 60m built. And entrained air mass (see Scaling section) should have the opposite effect, if any.

On the other hand, I've built enough kites that I expected to be stable that then weren't, to gratefully accept this one- swings and roundabouts (or rather, no swings and roundabouts in this case). But also, there's something to understand here that will be another useful insight, if it can just be figured out.

During one testing session, at maximum windspeed, the Red 60m SSSL Serpent dragged my (substantial) van downwind in the gusts. I was then usually able to regain lost ground in the lulls, all hail 4wd, and eventually help arrived (another vehicle) to pull it down with a pulley.

After this I fitted a take-down line which has been through some refinements and can now collapse the head instantly with just a few Kgs of one-handed pull. An incident at Berck sur mer France in 2019, when this one launched unexpectedly and could have resulted in fatalities, was a strong reminder to be wary of large SSSLs, even when fitted with take-down lines.

Lots of pull aside (an advantage when they are used as mega kite pilots), SSSL Serpents have only one significant fault- a tendency to stall off to one side when falling back if the wind drops a lot. Auto-bridling has reduced this tendency, but it is a fundamental characteristic of the tail attachment to the head, so may not be able to be improved much more.

The first SSSLs I tried were sleds, mainly because they are simple and don't require lots of bridles. The early versions used trailing edge hook-down to prevent skin and leading-edge collapse, and had lower leading edges supported by partial ribs.

The Arc Sled (below) is the simplest SSSL I can envisage and could be refined to require just a two- leg bridle. Unfortunately, this one had very slow recovery, showing almost no inclination to fly at all. Later understanding suggests that the amount and disposition of its lateral areas are likely to have contributed to this. Even back then I must have had some suspicion of this, judging by the holes I cut in the keels. The large trailing edge hook would also have contributed to SR (its LE is to the right). Worth another try?

I did then occasionally try other sleds, later ones all using ways to reduce chordwise compression and leading edge collapse other than by trailing edge hook. They have additional leading-edge bridles, so are not strictly sleds, but it may be possible to eliminate these LE bridles again when everything else is optimised, so I've continued to call them sleds.

In early 2020, after becoming aware of the advantages of pocketing (see SSP section), they were further refined. Although they now have adequate wind range, flying angle and even have SSP type collapse resilience (the ability to self- correct from shoulder indentations) they are bad leaners. While staying up reliably, they're just as likely to hang right or left by 20 degrees as to fly centrally. With ARs of around 0.5 this is unlikely to be a C of L position problem so has to be either too much frontal area or caused by having too much lateral area towards the rear; most likely the latter seeing as SSPs have similar leading edges but don't usually lean badly. There is limited scope to do much about this as rear keel area is there to hold the trailing edge open (in the absence of trailing edge hook). Fixable? Or are SSSL sleds another lost cause?

This series started when I finally accepted that 1Skin style SSSLs were fatally flawed and from the realisation that Serpent heads fly very well with ribbon tails- and up to mid-range winds with no tails at all.

Initially I'd made a Serpent head with flares instead of a tail (2 January 2019), but it wouldn't fly at all, seeming to have little idea as to where 'up' was and absolutely no inclination to go there (see Lateral Area section). Fortunately, I'd had problems with lateral area disposed too far rearward previously, so cut the flares off and tried again. It then flew exceptionally well with a small tail and automatic bridle: Wind range from less than 10km/hr to more than 60km/hr, high flying angle, lots of pull, minimal leaning, and no diving over.

But 62 bridles seemed a bit daunting for other than one-off making so I then started on a bridle elimination mission. The first attempt was to graft a Serpent leading edge onto a 1Skin (by the assumption that the key difference was the Serpent head's bridle-supported leading edge).

Its leading edge did not seem to be much better than the rib supported system, and then I belatedly re-remembered that the Serpent head's self-correcting response to leading edge collapse requires the centre head area to be free of ribs and flares (see the description of this in the Octopus section).

I then tried several semi-circular leading-edge open centre SSSLs with reduced bridles. but couldn't get really satisfactory flying - though they were more reliable than many of the 1Skins, so some progress.

To reduce weaving instability, and to reduce the number of bridles required down each side, the size of the flares was then increased. This made a problem of what to do with the extra keel depth around the curved head, so the leading edge was squared off.

At which time they also acquired a name: SSPs. - Single Skin Pilots.

By January 2021, SSP's were flying in quite strong winds with no diving over, but reliable strong wind flying while also retaining good light wind performance remains a challenge.

SSP 5 has slightly higher aspect ratio - to reduce weaving- and 22 bridles only. It also has the flares curving round the trailing edge slightly which has reduced leading edge collapsing markedly. The added drag from these lateral hooks helps keep the leading edge from collapsing, reduces chordwise buckling, and does so without the downside of shifting the C of L rearward, as was a problem with pre-Boomer TE hook. The drag from these corners is equivalent to the drag from tails, but as they have very little weight, they don't shift the C of G rearward as tails do. But for all that, their addition has improved tailless flying.

For some reason I don't yet understand, these SSPs require less bridle adjustment than Serpent head SSSLs- just one for most conditions, another for very light winds and a third for strong winds (2m Serpent heads require 3 plus an auto bridle). For this reason, SSP's haven't yet had auto bridles fitted. The intention is to develop the catenary auto bridle system described under Automatic Bridles for these kites.

They also have the self-correcting response to leading edge collapse first developed in Octopus SSSLs that makes SSSL kites much more reliable in gusty winds. Keel leading edges are critical- if they flare outwards, the kite tends to hang to left or right, rather than in the middle, but if they 'toe in' too much, the shoulders collapse in stronger winds. Currently I'm not sure there's any fabric cut that will not be subject to one or the other.

Will they get to be good enough for general use as pilot kites? Too soon to tell.

Another development has been the use of "pocketing"- or rather, the recognition of its function and its deliberate exploitation.

For a long time I regarded bumps and hollows in SSSLs as imperfections and strived to minimise them as far as possible by adding more bridles and using careful fabric orienation where appropriate to take advantage of bias stretch (where practical, bridles should be connected to cords sewn diagonally to fabric so as to spread loads, avoid stress points and reduce creasing). I then had ocassion to make two identical Serpents, and was puzzled when one of them required an extra row of bridles to prevent chordwise compression. The difference was in the fabric- one was made of soft-finish fabric, the other hard. Looking at these kites it could be seen that the softer fabric was pocketing noticeably more between the bridles- and this was on the kite that needed less bridles. This kite also flew in stronger wind before weaving instability onset.

About then I started to use the grid pattern of reinforcing and bridle attachment cording to increase this pocketing.- by setting the thread tension higher while sewing on cords. Extra drag from these deeper pockets replaces drag that would otherwise have to be added elsewhere - like from tails, tapering lateral area or higher bridled A of A.

But it wasn't until the SSP series that I realised that this pocketing was also improving overall compression resistance (spamwise and chordwise). For SSSLs like the 1Skins and some Sleds, which are quite smooth, compression creases become more evident as wind increases. These creases rarely develop symmetrically and are a cause of high wind instabilty by causing one side or the other to eventualy compress, diving the kite to that side. For Octopuses and Serpents with discrete pockets defined by bridle spacing, very little chordwise or spanwise ompression is visible in any wind speed I've flown them in yet. For SSPs, some distortion is visble near the trailing edges in strong winds but this doesn't seem to be destructive of stability.

I can see why the central part of each pocket resists spanwise and chordwise compression forces better than flat surfaces. I don't clearly understand why they also resist compression better at the 'fold points' on their boundaries- but they do seem to.

This is the account so far. The dynamic relationships underlying instabilities won't be resolved until single line kite stability is numerically defined and having to achieve stability without using internal pressurised spaces or any rigid elements adds enormously to the complexity. Seeing as the dynamic stability of something as simple as a bicycle is not yet properly understood, this won't happen in my lifetime- and possibly not in yours.

For now then, I'll keep on making and testing promising new variants along the general directions I've established and continue the approach of waking up each morning questioning everything I believed the day before.

Until I run out of fabric or wind.

Peter Lynn, Ashburton New Zealand, February 2021