Peter Lynn Newsletter March 2011 RE-INVENTING THE WHEEL


The ‘something more dramatic’ that was to have been my subject this month is delayed on account of unsuitable weather- that is, it’s been kite flying as usual.

But my ear to the ground tells me that no-one is breathing down our necks or looking over our shoulders on this one (would this be called a double mixed metaphor?).

So, no time pressure (for a change), and the unaccustomed luxury (matching the accommodation here in Kuwait) of being able to wait until conditions do suit.

And, I ran out of page space last month and have been rather chaffing to tell the second part of that story, so read on for some more boring “why kites don’t fly stuff” stuff instead.

Every person who has ever made a kite, probably everyone who’s ever flown a kite, maybe even everyone who’s ever watched kites fly, is familiar with hunting and looping instability. This is when a kite starts to hunt from side to side with ever increasing amplitude as wind speed increases, eventually crashing out in a series of violent loops.

This sequence is set off if a kite, when correcting after getting a bit out of line, goes too far the opposite way, then in correcting from this over-correction, over-corrects even further the other way, and so on.

Whether corrections build or fade can depend on unimaginably small effects- which is why kite making is so rewarding and annoying:

*Rewarding when they fly well: Irrefutable proof of our cleverness.

*Annoying when they misbehave: Persecution by malign external influences.

*But in reality: We’re usually just being fooled by randomness.

Very occasionally it’s possible to identify some small difference that has a big effect on kite behaviour.

An early prototype of the 10sq.m 4 bridle Pilot design I described last month was stable in all winds while the next one was irretrievably unstable. Between these two kites, I had made a new template for the outer ribs (profiles) as the original was a bit wobbly- but when smoothing out the bumps, had inadvertently moved the widest point forward by just 80mm (each rib is 3.25m long). The actual change in the rib depth at any point is not more than 10mm or so, an amount that gets lost in fabric stretch when a cut rib is laid back on it’s template- but the first kite flew and the second didn’t- and with this change reversed, the next one flew just as well as the original.

Having lots of lateral area (keels, flares, vertical surfaces etc) is effective in preventing hunting/looping, but a 5,000 or 10,000 year older solution (from the first kite I expect) is to add a tail.

All kites are pendulums; their weight (acting at the kite’s centre of gravity) hanging beneath wind generated lift forces, points them upwards. When they tilt over to one side or the other because of unsteady winds, this pendulum effects corrects the lean and gets them back to pointing straight upwards again.

Tails enhance this pendulum effect because their weight , acting at the tail attachment point, has much more leverage. The tail’s aerodynamic drag also helps keep everything pointing upwind and works to damp out any hunting that threatens to take hold.

Tails have a ‘dynamic’ effect, in the sense that their corrective effect increases rapidly when a kite starts to behave badly- which is pretty much the opposite of how the “Department of Corrections” (prisons) function in New Zealand-, but I digress.

And there are other dynamic stabilising mechanisms that kites use.

Rokakus have one that I admire a lot. When a Rokaku begins to turn to one side or the other, aerodynamic forces and the kite’s structure combine to cause the camber of it’s centre section on the faster side to increase and that of the slower side to flatten- which resists the turning tendency and shuts down any hunting/looping build-up before it can establish.

Deltas (and Genki’s, and fighters when their bows aren’t too stiff) have another:

Their wingtip angle of attack decreases as wind speed increases: reducing lift (which is the driving force behind instability) and increasing drag (which has a stabilising effect). This is very clever; not only do they automatically de-power to maintain stability as wind speed increases (because lift decreases and drag increases as the wingtips flatten off in response to more wind pressure), but if one wing should happen to fly faster than the other, then it’s extra flattening-off response acts to slow that side, restore the kite to straight flying and damp out hunting before it can establish it’s baleful grip.

And there’s a dynamic stabilising effects that some ram air inflated kites use also.

The 4 bridle Pilot’s automatic de-power system described in last month’s newsletter (which works by lifting the kite’s trailing edge as wind speed increases to decrease angle of attack, reduce lift and increase drag) also has a dynamic stabilising effect. This is because, analogous to the delta kite dynamic stabilising system, when one side finds itself flying faster than the other, that side de-powers preferentially, and this immediately damps out any hunting behaviour that might have had a mind to get in on the act. Embarrassingly, I’ve been using this system for years without realising that I have been; my standard pilot designs have always (or at least from when I first ditched the central flare) used profiles that are cambered in their rear half (induced by the flares so that if the kite overflies it reverts to reflexive form to resist luffing) and this has been providing, I now see, at least some auto-depower and dynamic stability.

I didn’t understand why or how this worked until a few months ago, just knew that this camber somehow made these kites fly better.

And here’s where credit is really due: Flowforms have had this dynamic stabilising feature for at least 25 years, and even cleverer, they enhance it by having a giant vee in their centre trailing edge so that one side can twist freely relative to the other. My hat goes off to whoever first figured this- it’s an idea that’s right up there with clever dick Deltas and Rokakus.

Aagh!: Seems like I just invented the wheel, then find out I’m years too late and that my version isn’t even properly round!

Peter Lynn, Kuwait City, Feb 28, 2011

PS But the 4 bridle Pilots did get lucky in two other respects.

Moving their centre front bridle to the top skin- which was done to improve initial inflation and reduce collapsing- has provided two additional useful features:

This centre top bridle point can be used to attach another kite; these Pilots are completely stackable (and much more reliable when stacked than the centre cell slot system that is available with some 8 bridle models).

And, this top bridle point provides an excellent de-powering system. Pulling on this bridle (when it can be reached) de-powers the kite and makes even the 22sq.m versions easily handle-able in strong winds. I expect that a second line rigged from this bridle to the ground will become an effective safety system when using Pilot kites for lifting things (banners, trains etc) and will save a lot of heavy work when pulling stuff down if the wind doesn’t co-operate by dropping off at the day’s end (everywhere except Italy).

Tail Stabiilising
Tail Stabiilising

News Letter Feb 2011

Pilots and Lifters (skirt lifters that is)
(A passing wind to get your attention, that last bit).

The first America’s cup in 1851 was an unequal contest between the schooner ‘America’ and a fleet of the best of British- and to tilt the playing field even further, ‘America’ had to sail across the Atlantic to Cowes while the English fleet were all specialist inshore racing yachts.
But America still won.
During the extended after match functions, the Marquis of Anglesey (who’d lost a leg at Waterloo 34 years earlier), went for a sail on ‘America’, and climbed out on it’s bowsprit for 20 minutes to study the bow wave.
English hull design theory of this time mandated that yachts should be wide at the front, long and tapered towards the back (sternly eschewing nautical terminology here). This derived from the false belief that raindrops ‘naturally’ assumed an optimal shape for minimising drag.
By contrast, ‘America’ was the opposite; long and thin at the front and with a rather curvaceous back end.

Back on deck, (after nearly falling overboard while checking for a rumoured secret propeller) the 83 year old Marquis remarked; “– I’ve been sailing my yacht stern foremost for the last 20 years.”

I now find myself in somewhat the same situation with respect to single line kites (and maybe even traction kites) -by which I mean I’ve been building kites upside down (rather than also being an old fart sometimes to be found hanging legless from the rigging).
In hindsight, I obviously should have taken more notice of how often my kites seem to prefer to fly belly up.

How embarrassing!

In defence, blame really belongs to Ray Merry and Andrew Jones, who, back in the ’70’s, invented the Flexifoil power kite with it’s revelatory upside down airfoil sections.
Conventional “cambered” profiles have more lift and are more efficient than other single element airfoils but aren’t angle-of-attack stable. They have a distressing tendency to suddenly rotate forward and dive down nose first out of control (called a luff). Aeroplanes control this tendency by having a tailplane out the back (or sometimes a small canard wing out in front). Flying wing style aeroplanes with only a main wing, have a distressing tendency to luff (aeroplane people call this pitch instability) with often fatal results- hence they aren’t very popular.
Reflexive ‘Flexifoil’ style profiles are angle-of-attack stable. They strongly resist luffing, hence their popularity with kite designers. But, unfortunately, they can’t generate as much lift for their size and aren’t as efficient (their drag being higher proportional to lift) as cambered ‘foils- which is why bridled ‘foils (which are conventionally cambered) dominate buggy racing even though they’re trickier to fly, and why flying wings can’t use enough reflexivity in their profiles to make them completely pitch stable without becoming uncompetitive against aircraft with conventional wing and tailplane form.

Thinking of single line kites as basically tethered flying wings, almost every soft single line kite I’ve developed, from the Octopus, Ray and Trilobite through to Pilots, has therefore used reflexive profiles for stability and luff resistance.

But now I know that this has been unnecessary or worse; at least in some cases.

The first glimmers of doubt came from various pilot kites and then the fish kite for UKS (Ultimate Kite Show, now Air Banners).
Contrary to what I’d expected, using camber rather than reflex seemed to make these kites more stable.
Solid evidence was there if I hadn’t been so blinded by the reflexive profile mantra; Jurgen Ebbinghaus’s cambered fish kite flies very well in every wind from nothing to too much,- and Flowforms often have extreme camber in their side profiles to apparently beneficial effect.
So, later last year, fresh from solving the Ray problem that had plagued me (and everyone who flew one or near to one no doubt) for 22 years, I thought to try a new Pilot design.
My bosses (licensees) want high flying angle and lots of lift in light wind, with as little increase in pull as possible when wind speed increases,- and there is no such thing as adequate stability in their view.
What I soon found was that positive camber not only improved the flying angle, improved stability and provided heaps of de-power, but that it didn’t provoke luffing either- which was still something of a surprise.
The better flying angle is obvious, but I now sort of have explanations for the other gains.
For the de-power and stability it’s that with a cambered airfoil, the kite’s angle of attack decreases in response to increasing wind speed- which reduces lift (hence the de-power) and allows a single line kite to maintain stability (instability is driven by lift forces while stabilising forces are a function of drag).
But how come this camber driven reduction of angle-of-attack with increasing wind speed doesn’t continue until the kite luffs?
Good question!
But it doesn’t, so once again the theory has had to follow rather than lead.
It appears that luffing can be avoided when using cambered profiles provided they are free to find their own angle-of-attack independently of line angle. If kites have ‘long’ bridles in the chordwise sense (and 4 line traction kites have the longest- they extend right to the flier) then they require either quick reactions from the flier (4 line kites) or reflexive profiles, so as not to luff when they overfly.
They also require either fairly symmetrical leading edges, or high leading edges (like are usual for Flowforms) so as to prevent their noses dipping under the wind when rearward camber lifts the kite’s trailing edge. Combinations of camber and a hooked down leading edge definitely cause uncontrollable luffing, regardless of bridling.
These new series Pilots have just 4 bridles and are relatively free to optimise their angle-of-attack. In response to gusts and during climbs to altitude, angle-of-attack decreases when required- slowing down apparent wind surges to provides de-power and better stability. When I tried reflexive profiles on a prototype instead, it just zoomed right to the apex (because the angle-of-attack then stayed relatively constant irrespective of speed through the air), pulled like crazy during this surge, and, as a consequence, started to figure-eight when lift forces overwhelmed stabilising forces.
This is a new insight for me- but there are kites out there (some Flowforms for example) which have always used camber rather than reflex, and our standard 8sq.m Pilot series has used some camber (induced by the flare shape) for years also- though I never really understood why this worked until now.
Anyway, though I really dislike having to dredge up theories to explain things after the event, I dislike having kites that don’t fly as well as they might even more. So, theory has again had to follow in the wake, catching up when it can. When will there be a truly predictive theory of kite stability?!

They have some promise I think; can fly in lighter winds, have more pull down low- but less in stronger winds, and are so steady as to be scary sometimes. No gain without pain though- currently when tipped over (which rarely happens, even in wild turbulent conditions), they sometimes go straight down without recovery.
An interesting innovation is that their centre leading edge bridle attaches to the upper rather than the lower leading edge. This is better for launching and re-inflation, and allows another pilot to be stacked above without having a central slot like some previous pilot designs have sported.
All the prototypes were 10sq.m, but I expect that the range for this design will be: 7sq.m, 12sq.m and 22 sq.m- for about the right size gaps and to suit fabric dimensions. These will be additional to, rather than replace current models.
Because the mass of air inside soft kites (which effects how quickly they respond to wind direction changes) scales with the cube of dimension while aerodynamic forces scale with the square, making larger sizes is not just a matter of increasing all dimensions proportionally. So that it responds appropriately, the 22sq.m versions of this series use an aspect ratio (span to length) which is 3.5% less than for the 10sq.m (but this is not a linear relationship so sizes smaller than 10’s can probably be simply proportioned).
22’s are available from PL Kites Ltd and from Airbanners now- and they really are lifters rather than pilots- will just sit up there solid as a rock lifting whatever you want into the sky.
I would like to publish plans so that kitefliers can make their own versions- and there are a few tricky features that even an unpick-and-copy person might not twig to- but my licensees have the view that this would likely lead to commercial scale copying, which could significantly cut into sales. They’re right, we’d probably be shooting ourselves in the foot (feet?)- but with other experimenters trying variations based on their(rather than my) half baked theories perhaps a few other things I’m no doubt getting wrong might get fixed.

And there are two other consequences of the camber rather than reflex approach to stability.

For traction kites, there is the potential to increase the lift coefficient (pull for size) and lift-to-drag ratio (efficiency) – provided the flying lines can converge close to the kite.
For single line kites there are really exciting (for kite trolls like myself) possibilities for new dynamic stability mechanisms. This will be the subject of a newsletter later in the year.

Next month I expect there’ll be something rather more dramatic to report.

Peter Lynn,
Ashburton, New Zealand (briefly) February 1 ’11

PS Media reports out of India last month that I featured in were bogus. I never had any Tigers cut- didn’t happen- and as for being intimidated by “slum boy kite fliers” – yeah right! The Indian press certainly doesn’t let facts stand in the way of a good story. An excellent set of festivals though!

Peters Kite gets cut out of the Air at India

Check this out

AHMEDABAD: This was Peter Lynn’s 600th international kite festival. Known among the world s kite manufacturers, Lynn, from New Zealand, was flying two inflated tiger-shaped kites, each worth 400 US dollars, on the bank of the Sabarmati river, when a puny paper kite came from across the river and slashed his tigers, on Monday.

The loud music on the west bank of the river where Gujarat was hosting the 22nd international kite festival, drowned the slum dwellers screams on the opposite bank. They realised what they had done only when the police came looking for them.

Lynn shouted for help and with the festival volunteers, tried to find the boys who had cut his kite, but they disappeared. The New Zealander who has also judged one of the kite contests in this festival, told TOI, I refuse to be part of a festival where my kites are attacked by others flying with strings which has glass pieces. I have kites which are of thousands of dollar and I can’t lose them to boys on other side of the river.

The Uttarayan festival in Gujarat is all about cutting each other s kites, but in this case, it was far more serious. Most foreigners brought their kites down after the incident. At least 200 kite-flyers from 36 countries and 10 Indian states are participating in the festival, flying kites shaped as dragons, angels, scavenger birds and so on.

Lynn said, this is crazy and happens no where. An Indian fighter kite attacked my tigers and before I could realise what was happening their strings entwined in my kites . One of his kites fell on the festival pavillion and another fell far away and was found only after an hour. The police went to the opposite bank and scolded the slum boys.

We stopped flying for 30 minutes and started again after the fighter kite was pulled down and the police began guarding the other bank, said Rami Al Khal, a participant from Lebanon.

Read more: Slum boys cut Kiwi star’s kite in Ahmedabad – The Times of India

Slow Thinking About Rays

An annoying problem finally identified and cured

Ever since I’ve been making large ray shaped soft kites (from 1987) they’ve been afflicted by a tendency to fall off to one side or the other.  And it’s not just my ram air inflated ray type kites that have this disease, Volker Hoberg’s SFT and Andreas Fischbacher’s Manta Ray sometimes show the same symptoms – they probably caught it by getting too close when they were hanging out here.

In 25 years of trying to find a cure for this affliction- I’ve gone through more bright and shiny new theories than there are trees falling in the forest with no-one to hear them.  In darkest moments I’d begun to think that the entire lifetime edifice of kite stability theories I’ve constructed was of no more value than completely random guesses and was feeling uncomfortably like as weather forecasters and people who write horoscopes must regard themselves in moments of self honesty- if they ever have any.

When watched closely, what seems to happen is that the leading edge on one wing or the other loses inflation; gradually crumpling up and pulling the kite to that side.  Especially when the wind is very light this will be terminal, but in stronger winds, the kite will sometimes then recover and repeat the sequence but to the opposite side.  Ray style kites with lots of tail drag are less susceptible than the later smooth tail styles- but they all do it.  With a mega version about to be started, this most annoying behaviour had really been giving me sleepless nights. 

From the beginning, a puzzling aspect was that it appears to only afflict larger versions- smaller models are immune, even when dimensions are exactly scaled.   A standard maxi bucket tail ray will rarely fly pilotless for long periods without coming down with a bad case, whereas midi versions of the same kite fly on forever, apparently uninfected.

Ah-ha, why could this be?- a clue!  Something to do with fabric stiffness?- which has an appreciable effect on the structure of a small kite but less effect as kites are made bigger.  Reynolds number effects (don’t ask!) and non scalar surface roughness are the only possible other candidates but they just don’t stack up- so it HAS to be fabric stiffness- there’s no other sensible explanation 

Which pointed to some sort of structural irigidity being the primary cause of the problem

I’d made it this far by the late eighties, but then wrongly seized on poor inflation as the perpetrator- leaving the real villain to continue it’s serial offending until a road to Damascus moment during the Xiamen Festival in November.

In defence it’s fair to say that poor inflation does contribute to errant behaviour of this type.  Careful positioning of the inflation point(s) and the use of flap valves to prevent egress during lulls does help.

But even with perfect inflation the underlying behaviour always still lurked- just waiting for the most inconvenient moment to strike.

OK OK, no more suspense:

I now know that the actual cause is fabric shaping that leaves the leading and trailing edges loose while the centre chord area alone resists the spanwise component of internal air pressure.  And it happens because we all took what seemed to be the easiest way to make these kites:-Sewed up two essentially flat lozenge shaped pieces of fabric, one for the lower skin, the other for the upper, then sewed them together around the edges and added a tail eyes and mouth.

And to visualise why this causes there to be looseness at the leading edge, think of a heart shaped helium filled message balloon.  They’re made from two flat pieces of aluminised Mylar film that are then sealed around the perimeter and inflated.  All those little creases that then appear transversely across the edges are exactly what our problem is. 

Or take two A4 sheets of paper, tape them together around the edges and inflate them. Transverse creases will appear across the edges, especially in the centre of each long side- these are the only way that flat sheets can adapt to taking a convex shape.

Instead of using two basically flat skins for the upper and lower surfaces (albeit with shaping to lift the tips), we should have been taking care to make them conform to the three dimensional shape that they take when the kite is inflated- perhaps even adding a bit of extra fabric through the centre chord area to ensure that as much tension as possible is taken where it matters most- along the leading edge.

Simple isn’t it!- and so easy to fix seeing as we have complete freedom to use many and varied shaped fabric pieces when making the upper and lower skins- they don’t have to flat, but can be made to conform to any three dimensional shape we choose.

But does it really work?

When I belatedly realised what the cause of the problem was (while watching a smooth tail TT Ray) during the Xiamen festival, two tie cords rigged along it’s trailing edge to pull each side in by 0.5m or so immediately smoothed out the leading edge- and the kite then flew central all the time: Problem solved.

So why did this take 25 years? 

Good question. 

The only answers that readily spring to mind are, incompetence, a slow brain, and fundamental lack of understanding.

And now the alert reader (if there are any still following this except Volker, Andreas and Robert van Weers*), will be thinking:-Yes this explains why the leading edges looks wrinkly and un-inflated on these kites, but how come their trailing edges are tight?

This is because, by air flow and Bernoulli’s theorem, there is always less pressure difference between the outside air and the internal space at the leading edge of ram air kites than at their trailing edges.  Any looseness at the trailing edge migrates to the leading edge and adds to the woes there.

And why does looseness at the leading edge cause such bad kite behaviour- even when inflation pressure is good?  It’s because looseness in the leading edge doesn’t remain equally disposed to either side of the kite.  Sooner or later a small wind shift or whatever will cause all the looseness to accumulate along one leading edge while the opposite leading edge becomes tight.  With one leading edge smooth and tight, the other loose and draggy, the kite will inexorably fall off to the high drag side.

Peter Lynn,

Ashburton, Jan 1 2010.

*Because Volker and Andreas will be reading on to see if they agree with me or not and whether there’s anything useful for their designs in this, and Robert will be telling me by now that he never fell into this trap in the first place with his diamond shaped ram air kite (which could potentially have the same problem) because he’s not a Luddite and used a 3D program for skin shaping.

Peters 64th Birthday

On the 23rd of December Peter Lynn turned 64.  Elwyn, Peters lovely wife, laid on a great spread for all the gang and we helped Peter celebrate in style.  I managed to take a couple of pictures early on in the evening.

“When I get older losing my hair….many years from now