SSSL Development from, 4 April'14 (Chongqing) to 4 March '17

SSSL Development from, 4 April'14 (Chongqing) to 4 March '17 .

Reprise:

SSSL 3 was reasonably promising but none of the attempts from then until SSSL13 were very hopeful (some did fly OK in smooth mid range wind).  SSSL4 to SSSL12 were all pure sleds of one form or another (no central keel).

Some observations so far:

1. Using a lower leading edge seem to enable lower A of A without LE collapse- but SSSL19 green?

2. It appears that pure sleds become too wide at rear, tapering in LE then causes collapse.

3. Or, is the unsupported LE the problem- consider Robert vW's.

4. Supporting the LE in centre span (tried on some of the sleds) causes luffing at higher apparent winds.

5. For SSSL13 to SSSL19, left/ right tuning is very responsive to "c" sides.

6. Not having taper-in at the LE (in plan view), seems to be critical to prevent tip collapse.

7. LE collapse always occurs at tips first- so far- even for SSSL's with extra centre panel camber near LE.

8. Later SSSL series are least stally when central keel bridles are at low A of A- but the keel then flaps.

9. Reducing the keel A of A causes the rear part to narrow- offsetting the A of A effect.

10. SSSL16 (shallow central keel) was clearly less stable than any of the series with deep keels.

11. There doesn't seem to be any stability disadvantage in the central keel being shallower towards the LE.

12. SSSL19's lack of a lower LE (it has the arch section pieces) doesn't appear to have reduced LE stability.

13.Has SSSL 19's centre panel forward camber helped extend the lower end range?- appears to have so far.

14.Tails don't seem to help light wind stability at all- but maybe would in conjunction with reduced A of A.

15.Up to SSSL 19, recovery from dives is very slow- a function of Cof P being very close to C of G?

16. What causes light wind /low line angle tip-overs?- is it fundamentally C of P falling below the C of G?

17. 4th April Chongqing, SSSL19 (purple) eventually flew reasonably in light conditions; down after 9sq.m Pilot. 

18. Low pull and line angle though;- would SSSL 18 (different keel camber) have been better?   

After Chonquing

20 was cloned from 19, no significant changes (was it light weight fabric?), given to Orlando in Uiseong Korea (April'14?) to see what he can make of them.

 21(green) had minor changes from 19, but no lower leading edge. 

Quite a bit of testing, but has not yet (October '14) flown as well as 19, which does have a lower leading edge- for now this has to be assumed to be significant.

At Uchinada  (May '14) 19 flew well as a pilot above a Ray +- was really excellent, steady, better angle than any of the pilots and much more pull relative to its size- pretty much the same as a 9sq.m Airbanners.

This is hopefully a harbinger of what may be possible from the SSSL approach but at Uchinada there was a soft steady sea breeze- just enough to keep above 19's stall point, not so much as to cause LE collapse on that bridle setting- (above about 30km/hr).  When the wind dropped, the 9sq.m Airbanners and 19 came down at about the same time- though I strongly suspect that the 9sq.m has a generally lower light wind threshold than 19- because the SSSl's weak point currently is their inability to recover from a stall.

After this sequence of events, tried a 3 cell with higher aspect ratio- 22 (red),

22 also went to 3 bridles per cell- 12 total  (earlier series are all 4/cell, 12 total) tested at Wakanui  in Sept, initially with chronic outside LE collapse.  Sewing in a pleat to choke A side flares by 50mm was a substantial cure for this.

Taiwan (Sept '14). 22: Appeared to be much more stable off the launch than lower AR SSSLs- goes straight every time, not as loopy at high angles of attack as the lower AR versions, but also had an edging type behaviour- would not stay at the appex for long, would always eventually fly off to one side or the other and crash- maybe a bit less inclined to do this in stronger wind than light.  22 also appears to have a creasing problem on the outer cells especially at B, from point bridle loads- a consequence of  having just 3/side- but maybe also because B flare is too small (B outers are out about 75mm further than A and C).  Just before Taiwan, sewed out pleats beside the outer flares to reduce overall camber.  Unfortunately I accidentally made one of these adjacent to the centre cell instead, so consequently 22 did not get a good test in Taiwan. After unpicking both these pleats it didn't seem to fly as well- suggesting that less camber is indeed the right direction.  It did still fly in mid range winds, still inclined to edging though.  Took 21 (no lower leading edge) to Taiwan also- wouldn't fly very well, I expect not nearly as well as 19 would have there, even after some bridle changing.

19 was in Mongolia with Craig and Gavin as a pilot while I was in Taiwan - did they use it?

After Taiwan, re-cut the outer panels on 22 to make the tip choke more accurate- and sewed pleats in the outer cells adjacent to the centre flares to generally reduce camber.  Wakanui test looked promising- definitely holds its form well now, quite resistant to tip collapse, but still edges.   

Shenzhen October '14, very light wind

19 (purple) wouldn't stay up- stalled and collapsed, neither would  22 (red) for any length of time- would apex then eventually dive off to one side or the other with a serious crease at B outer.  At a school demo, we could tow launch and run 19 around- definitely more inclined to ground loops than 22 is on launching, but does stay in the middle while flying

Weifang , Nov 2 '14

1. Is 22's edging behaviour an aspect ratio effect (super stability) or is it just point bridle load at B outer?- make B outer flares larger and re-test at Xiamen.

2. Choking the tip panels by 50mm over 250mm (approx 1m chord) definitely reduces wingtip collapse.

3. Tapering off the lower leading edge to the flares at the tips and in the centre seems to have worked well- doesn't seem to reduce the effectiveness, provides access for sewing on the bridle without a lot of stuffing around.

4. It's tempting to try an Arc style SS- flat choked centre panel, 2 long flares per side (so that  A of A can be changed, but I'm anticipating that if such a kite stalled it would not easily re-inflate- would hand clap like ram air arcs only maybe worse because of having no structure to initialise the opening.

5. Flat multi cell fully bridled SS's after the style of Michel's 4 line SS traction kites seem to be able to fly at lower A of A's without stalling or collapsing than the Skin style.  Maybe the significant measure of A of A is a line from the stagnation point at the LE through to the trailing edge- and using bent down TE's as the Skin and current SSSL's both do may not allow a lower A of A - but just moves the overall C of P rearward, putting it closer to the C of G and maybe exacerbating instability approaching the stall. 

6. The current inability of SSSL's to recover from stalls may be just that when they fall backwards, one side rear collapses preferentially causing the kite to tip to that side even to 90degrees or more- from which re-inflation and recovery is impossible in the very light conditions that there must have been to cause the stall in the first place. Could also be just that the cambered down TE tips them over backwards.

7. But is there any solution to this except just reducing the stall speed?  What if the outer rear flares were made with reverse choke?- though this would also cause them to push in when flying normally- but maybe there's enough pressure back there to resist this, at least so as this wouldn't cause collapse. Or what if the SSSL could be caused to drift back with reducing A of A (that is, glide back) rather than at increasing A of A as now (which is a caused by a combination of long bridles and TE camber?)

8.  Try a flat, hour glass shaped, AR 0.75  (19's AR seems to be too low for best stability and recovery, but 22's is too high?), multi cell 5 bridle/cell SSSL with no TE pull down and a 10% (+/-) lower leading edge.

Say 7 cells (8 sets of bridles spanwise), Flares 60degrees included angle or even a bit more tapered?

9.Flare choke could be applied by adjustable cords between the rear edges of the outer A flares- and this might also have a useful feedback function in stability- choking the side with the least bridle load on it (bridle tension decreases on the LE bridle as that tip begins to collapse)  his could also work for the rear flares reverse choke- would this cause the rear to narrow or collapse?.

10. Another approach might be to make in the form of a laterally symmetrical trapezium,  narrow LE, with a row of flares down each side, all angled so as to pull outwards- maybe the front flares could have more choke than the rearward.  This shortens the leading edge a lot and maybe would work with bridles only to the side flares- try 4/side to start with- and overlapping.

11. Another approach; try making from a single sheet with folded over LE like the centre section of a NASA, bridled to each tie, 4 (?) side flares all bridled and many centre bridles directly to the skin- sew on cording for attachment points.

Weifang, November 3 '14;

Made; 23 green: 7 cell, parallel sides, with matching convex LE, concave TE to maintain equal chord,  1.2m span (nominal), 1.3m chord, 4 bridles/cell, 32 bridle total, about 5% lower leading edge with 6mm x 2 of pleat per cell to match bridle camber.  All cells spanwise are to have equal bridle lengths in 3 cascading sets.  Concern is whether LE tips will hold form without choke- but the outer cell sweepback is around 45 degrees, so this will be a test of how much sweepback contributes to tip stability in the absence of choke.

It has no camber at all as cut- completely flat skin LE to TE.

First pull up test at workshop door, appeared to fly OK on first (guessed) bridle dimensions, which are very forward. No noticeable tip or LE collapsing.  

Weifang, Nov 4 '14:

Made 24 yellow: 7 cells spanwise, semi circular shape (curved leading edge) 24 cascaded bridles (spanwise rows of 4, 4, 6, 8), approx 5% lower LE, but this tapers to no leading edge in each outer cell (LE in this cell is very swept back.  approx 1.3 m span, 1.3m chord, perhaps AR 1.2.   There is around 5 degrees choke in every cell. As for 23, no camber at all as cut- but 2x6mm pleats in the centre 5 cells to accommodate arch.

The thinking behind this design is that if sweepback does prove beneficial, this is the type of shape that will be used.  Their difficulty is that the rib to panel connections at the high sweepback sections are difficult to shape right so as not to wrinkle- and difficult to sew( the more the sweepback the trickier this is).  Maybe making the LE of individual spanwise strips like for the 19 would make this easier- maybe only necessary for the outer more swept back cells.

24 was also bridled very far forward for first pull up test outside factory door.  Appears to fly, no LE or tip collapse apparent, but has major wrinkling the 3rd rib LE each side caused at least in part by having paired 3A and 3B bridles; should have used only spanwise pairing..

Xiamen 2014-11-07

Strongish wind off the sea- 30km/hr.

23 green;:leading edge smooth and completely solid (seems fairly forward bridled too).

Trailing edge (actually the rear 25% of the kite)  loose and flapping-and pulling in the 4th row even to stalling didn't stop this- why?- not enough (or even negative) chordwise camber which will need to be corrected by letting B's and C's out, not by shortening D's?

 Flying off the bridles, seemed to be a bit volatile.

Quite a bit of pull- but I'd say nothing near as much as 19 would have had in this wind- but very much doubt that  19 would have flown- or that its LE would have stayed solid.- check this.

24 yellow: LE solid but notchy, and sides need an arch pleat- very loose.

TE is solid too- doesn't flap at all like 23-  the skin does seem more cambered  than 23.

Flying off the bridles was towards superstability- or is it ROI (roll over instability)- and noticeably right wing (bridlings on 23 and 24 are still the fairly random first guesses from Weifang.     

Directions; Need to test now in very light winds- check if stall collapses are a problem. With these designs also.

Generally 23 and 24 are a breakthrough.  Still to be determined: Best- or at least functional-combination of :  lower LE, LE being deeper, having ribs closer together, and less camber?  There doesn't now seem to be much doubt that SSSLs will be successful kites.

The next 'pilot" version could be some combination of 23 and 24 maybe- 23 with higher AR or 24 with slightly lower AR and the LE panels shaped better. 24 has the virtue of only 24 bridles (32 for 23). Could try releasing the D bridles on 23- and if it flies OK (will need to shorten A's a bit to compensate for C of P shift) then try cutting the rear ¼ off.  This increases the AR, and reduces  bridles to 24.

And make a new test kite in the shape of a Ray's wing; could standardise on shallow ribs , making bias cut strips, slight scalloping, bridles at about half the current spacing and say 4 or so different depths, which could then be used for the different chords required on show kite wings.  The primary bridle set could be in chordwise sets , all the same length (perhaps the LE bridles should be singles).  Strips like this would probably have a better appearance than the current triangular flares. There might be some possibility of just using close spaced bridles, no ribs at all- like NASA 'spines'.

Xiamen 7 Nov '14.

Very light to no wind.

23 does seem like it will fly without the rear quarter- all D flares. With D's released rear was still not very solid. LE and shoulders are very solid though- better than 24 which has excellent centre but creased and collapsing shoulders.  Did play around with 24 extensively- flattening off to see how far forward it can go.  It's a lot, especially when D's are released until just short of getting a crease towards the front of the D flares. If the LE is pulled down too far it dives off to one side or the other irretrievably sometimes- presumably because one shoulder or the other is collapsing more (has this appearance).  With the flattest profile it will accept without creases (perhaps 50mm to 100mm camber chordwise and at the most forward bridling it will fly without too much narrowing of the LE spanwise or indenting. It's just volatile unstable in the lightest wind it will fly in but steadies up if there's 10% more.  With a V tail, was quite reliable.  It would fly in wind as light as the latest 9sq.m Airbanners but only when continually pumped.

Tried taking a duct tape 2 x 15mm arching pleat from the shoulders- did improve appearance but didn't by itself reduce preferential shoulder creasing. It recovers well from lulls- drops back, sometimes turns quite a bit sideways, but then zooms off to the side and  comes back up in a long curve.  I did briefly try 19 in the same conditions- 24 is noticeably easier to launch, is more forward bridled for the same or better LE stability, and recovers much better from lulls.  24 does still exhibit the same volatile behaviour which steady's down as it gets to a higher angle- and in more wind- but it has it to a lesser extent.

Tested 21 with the larger B outer flares- still wouldn't stay central- always dived off to one side soon after getting to apex.  More and more likely that this is ROI from too much camber towards the rear.  Didn't try different bridles though.

Next prototype: a 22  with higher aspect ratio and maybe only 3 flares/cell (they need to be slightly larger than for 23 anyway to reduce its volatility.  Should the TE be concave or convex?

Xiamen 9th Nov '14:

Light onshore wind building to  35km/hr by 5pm.

Tested 23 in light wind mode; did narrow LE a lot and was inclined to edge terminally because of shoulder collapse, but seemed quite close to being flyable.

Ashburton from 13 Nov '14

Tried a 4 bridle per cell 7 cell developed from 23, did fly quite acceptably in a gusty northwest, but also seems to have a hang-up mode when it sits at lowish angle stalled.  In light winds it tends to edge quite badly.

24(orange, 3 x 8), developed from this was not improved in the edging- what's causing this?, but does seem to be able to accept quite forward bridling without LE or shoulder collapse.  it's interesting that even without chordwise choke, the LE shoulders do stay well inflated

25 (purple 3 x 6) had a lot of flying and development but is plagued by this edging problem- when it gets perturbed by more than about 20 degrees it inexorably dives off to one side or the other and crashes.  Tried many things, but nothing seemed to make much difference.

26,(yellow, 3 x 6) same basic patterns but small changes to the shoulders and some camber in the ribs.  A lot of testing with this in mainly gusty winds, no useful improvement.

27 (red, 3 x 4) reduced spanwise to 3 cells (12 bridles total).  Did fly for periods in gusty easterly, especially after adding another bridle set between A and B. Very sensitive to tuning and with the dreaded non-recovery when flying sideways

28  Orange, developed version of 27, with 4 bridles/cell (16 total) small changes to bridle points and with a larger lower leading edge-  not significantly better - is quite volatile so;

29, purple, same as 28 but with a 300mm wide cemtre panel (28 is 225mm)  many hours of testing with this in gusty easterlies.  Initially it wasn't nearly as reliable as 28 but reducing the LE return by 30mm- 25mm helped somewhat, as did cutting away the central ribs between bridles to reduce pocketing.  A theory now is that the edging is caused in part by transverse flow.  When the kite is lying on it's side towards the edge of the wind, it functions as an airfoil with the rear outside upper edge generating more lift than the front outside upper edge- which causes the rear to tip up and the kite to slide/dive off to that side.  If the dive gets steep enough, the kite will loop and recover, but on a short line and in the tree wind shadow it more usually crashes to one side or the other.  Shortening C on one side is an excellent and sensitive tuning system- but this  kite is extremely sensitive to tune- if it is set to not deviate more than a degree or 2 either way it will stay up for periods, but more usually hangs off a bit then gets into a death dive.

Tried cutting 2/75mm dia vents in each rib near the LE for cross venting- did this reduce edging?- didn't make it worse anyway, but top skin now has more bumps.  

Tried cutting away the rear outside bridle triangles to reduce area at the rear relative to the front and in this form 29 does still fly about as well as before this- but is suffering from poorly supported rear area- develops creases, makes flapping noises and hangs low- or crashes.

Summary: it now seems likely that cutting away the front flares from 25 on, which  did improve the shoulders a lot (in the absence of choke) but had the initially unrecognised consequence of shifting the lateral area rearward which seems to be causing the edging/tipping. A solution may be to return to using larger area front flares in the centre ribs only (and go back to 3 bridles per cell in the centre ?) while continuing with using the cutaway front flare at the shoulders and 4 bridles/per.  It may be that once this edging is solved, a return to 7 cells might be possible (and helpful if volatility is a problem for the lower aspect ratio versions). 

25 Nov '14 Ashburton; made 30 green 4 x 4:

Reduced lower leading edge to what 29 had been modified to above, but didn't add cross venting.

Added flares back to the centre A's and reduced centre B, C and D flares to 100mm shorter than for 29 and earlier prototypes - to move the lateral area C of P forward as much as possible, and to reduce the flopping of the centre ribs which may have been contributing to edging. 

Added flares to outer B's and reduced outer C's and D's by 25mm also to move lateral C of P forward.  Kept outer A's cut away so as to not have shoulder collapse.

First test (gusty easterly behind the trees) looks promising- definitely recovers now, much less tendency to edge, maybe a little volatile- so make bigger or higher AR?   Recovers well from stalling- probably the best yet.

Quite a tidy smooth kite and appears to be able to retain inflation at low A of A.  Need to test and tune in smooth wind and try very light wind.

Up to 2 December '14

Extensive testing of 30 in easterly at 105 and at beach- can't get it to stay up reliably, even with a lot of bridle adjustment.  It's OK for a while but as soon as it tips to one side more than about 60 degrees it dives over to the ground, doesn't recover.  Is good in light winds, doesn't have the tendency to tip over whenever it drifts backwards and tip over that 19 has.

Despair!

Took  19, 24, 25, 26, 30 to the beach and tuned/tested them- 20+ km/hr north easterly (not from the sea).

The only one of these that looks promising is 24 (yellow, 24 bridle plus 2 extra) made in Weifang, Quite high aspect ratio, every cell is choked a bit very curved leading edge.  This kite did look promising at Xiamen  but was very distorted at the LE of the outer ribs. 29 Nov, after the beach, added a fabric wedge to eliminate the distortion and added an extra bridle on the LE between the outer rib nose and the rear outer flares.  After some more bridle adjusting, this kite has become an excellent flier- was as good as any of the pilots and lifters at the 30 Nov Ashburton Domain flying day - when the wind moved from south to east (every minute), was quite strong (difficult to hand hold a 12sq.m Lifter) and VERY gusty.  In current mode, 24 is very flat in the profiles- almost reflexive.  It is exceptionally stable, not quick recovering but doesn't tip over so doesn't ever need to recover.  Pulling in B on an outer rib (there's A B and C on this rib) works for tuning - 5 -10mm has a big effect in strong winds.  In the strongest conditions the trailing edge tramps- but it continues to fly.  In even stronger this will probably cause super stability- was just starting to at about 40 or 45 km/hr. 

See discussion and photos in December Newsletter.

On 2 Dec, made a development of 24; single panel with scalloped ribs (4, 4, 3), slightly lower (10%?)  AR, slightly shorter lower LE and it's made from a separate curved piece, not an extension of every panel as 24 was. A bridles come straight from the LE lower edge, do not overlap 20mm as for 24. 

In initial testing it was fairly cambered and zoomed around a lot- would loop, but more inclined to dive to the left and crash.  After letting off TE about 25mm, pulling in B's and C's about 20mm and letting A's out 20mm, started to look and fly somewhat like 24- though isn't yet as flat in the TE  and appears proportionally narrower at the TE.  This kite will be a bit wider than 24 on account of not having chordwise panels with seams, so maybe AR is closer to 24 than I intended.

 27 Dec '14: Why does 32 (green, 4 Dec '14) collapse while 24 (yellow, 4 Nov) doesn't?

1. 24 has chordwise choke, 32 doesn't have.

2. 32 is larger than 24 (10%?)- fabric stiffness effect

3. 24 has compound triangular sewn on flares,  32's are one piece scalloped.

4. 24 has marginally greater leading edge tuck.

5. 24 is flatter in profile- test by de-cambering 32

6. Is it to do with progressive collapse?- 24 collapses in such a way as to remain flying, 32 folds to one side.

7.  Bridle overall length?  No, checked this by trying much longer and much shorter on 32.

8.  Fabric type?- 24 is lighter, softer- so how could this be?

Wakanui, 27th Dec '14 Strongish NW: 24 stayed inflated, flew marginally, LE crumpled back a bit but not destructively, TE panel tramped.  32 not much behind- but LE would eventually crumple enough on one side to cause a dive -off.  No TE panel tramping (bridling?) and much more noticeable creases at 2nd and 3rd bridles.  Became noticeably compressed chordwise compared to 24. Tried with cambered ribs and without, same bridling, no significant difference- maybe a little less inclined to collapse chordwise with the rib camber removed.

For 33 , try same as 32, same chord and ribs and bridles but 10% less span; by theory that there is a spanwise as well as chordwise max bridle spacing.

Built 33 (Orange) on 28th, flew 29th at Wakanui and 105, strong gusty easterly. 

Definitely the best SSSL yet- flew well immediately with initial bridling (all the same length), not even checked for symmetry. Much less lateral wrinkling than 32 or 24 .  LE stays inflated better too.

There is no other explanation for this improvement than that spanwise bridle spacing matter- narrower cells , closer spaced ribs hold better shape.  There is likely to be a formula for this- perhaps some product of spanwise and chordwise spacing is approx. constant - though deeper flares do appear to help as well.

30th Dec '14, built SSSL 34 (Red), quite strongly right wing, didn't fly anything like as reliably as 33- but had let out centre B's by 15mm to eliminate crease at B- which it did. 

1 Jan '15, re-sewed one side's ribs (except one nearest centreline) to get lateral symmetry, restored bridles to original- all same length.  Now flies as well as 32, and slightly left if anything.  Does compress chordwise and LE bends under a bit in strong winds, but recovers exceptionally well and remained flying when 12 sq.m Airbanners Blue was crashing every 30 minutes or so - and 34 was on half the line length- strong easterly.

1Jan '15, began templates for larger version; scaled x 1.5 chordwise, x1.4 spanwise- all flare depths scaled 1.5x

2nd Jan '15.  This one, SSSL 35(yellow) is the best yet- flying reliably (came down twice in 3 hours, each time because an unsettled bridle had stretched out 20mm- (used soft Dyneema not hard finish)  on a short line in a strongish easterly/southerly- trees are thrashing a bit.  Has a LOT of pull- not really possible for me to hang on to it on 2mm line.   There is a serious crease at C bridles centre, none at A or B or D- but trailing edge flaps loudly in strong wind, compression collapse.  This seems to be having effect of moving the C of P of the functioning (forward) part of the kite rearward which ensures that the LE stays inflated.  Definitely an effect worth keeping if it's doing what I think it's doing- working like a reverse spring bridle. 

SSSL 35 has 26/ 3m long bridles, weighs 360gm without line (45gm acrylic coated fabric).

Built SSSL 36 (green), same as 35 but centre ribs shorter by 150mm, 2nd ribs by 120mm, outer ribs by 80mm- 2.5sq.m effective lifting area.

SSSL 36 flies OK- quite right wing, more creases on right rear, does tramp trailing edge but not as much as 35- does it fly as well in stronger wind?

SSSL 36 compared to 9sq.m Airbanners (latest model) Pilot:

Light gusty intermittent easterly at 105 AF Rd, 4th Jan '15

36 has more pull on average, holds a higher angle, doesn't wander off to the sides as much (quicker recovery from perturbations), seems as reliable or more reliable for same height, but presumably won't have the upper end.  Light end seems equivalent but this needs testing in smooth unobstructed gradually dying wind.

7pm same day, More north easterly now, 30km/hr?:- Trunks on main pines are moving noticeably.  35 came fairly straight as bridles settled in- some tendency to SS in strongest gusts, either way.  Angle same or only marginally better than the 9 Pilot in this wind, pull around the same for both-  20kg to 25kg.  Pilot is much less stable (but 70 m out as compared to 80m for 36), collapses, inverts and usually recovers.

Lots of TE tramping on 36, but the LE holds form very well in the strongest gusts- strongly suspect that this is because of C of P shift deriving from TE panels disengaging from lift? Much less centre panel creasing than for 35.

All the bridles the same length exactly still seems to be better across the wind range than any other variant- even just 15mm +/- on a few bridles.

6th Jan '15.  Built 37 (Purple)  shortened as for 36 green but with more even spacing of bridles, especially on outer ribs (35 has big gap from B to C)  Flies exceptionally well with reduced TE tramping in strong winds (still a lot though) after 8 or 10 hours heavy wind flying, quite noticeable wear marks in TE panels.

Tuning: was initially left, then caught in a big tree, damaged pulling out, After repair, strongly right.  Tried very many things over 5 hours+ continually.  Pulling in left side 3/B and 2/B,C  and letting out right side mirrors did almost correct the lean but in stronger winds it re-appeared.  Max 30kg pull (gusty easterly).

Tried releasing left TE bridles - but didn't work at all.

Then tried letting out left LE on shoulder and 3A- works immediately and precisely, no other bridle asymmetries required.

TE Tramping:

Is it a necessary A of A modifier that allows the kite to hold LE inflation at higher wind speeds, as per above, or is it just inadequate TE support.

If the latter, fixes could be:

1. Shorten TE a little to create some hook- works for sports kites

2. Shorten TE of ribs so as to do the same.

3. Move rear bridles rearward by 50mm or so.

9 Jan '15 Tried 2. By sewing 2 x 12mm pleats in rib and tip TE's of 37 purple.  Quite strong easterly, 20kg + pull, does seem to be flapping a bit less - and does seem to now fly at a higher angle in this wind (or is the reason that it's now flying straight?)

38 red; as for 37 but with rear bridles (ribs and tips) moved rearward by 85mm (other bridles spaced appropriately). Flying 38, strongish N E; initially strongly left wing, after making bridles symmetrical is straight- but responds very well to tuning ONLY with shoulder bridles- +20mm on left shoulder was enough to cause strong left, when corrected, kite is central.

TE tramping a LOT less severe than earlier versions, probably better than modified 37 above.  Because 37 modification did reduce tramping and doesn't appear to increase instability, suggests could do this to 38 in addition to moving rear bridles rearward- which didn't seem to cause any additional creasing anywhere either.

10 Jan '15 Built SSSL 39 (gold 30gm). - Very minor template tidying from 38.

Flies straight off but strong right (lightish easterly) Still to the right after shoulder bridles made the same- and still a little right with LH shoulder 15mm longer.

Doesn't appear to be any less inclined to hold LE shape than heavier fabrics, and no more (perhaps less?) creases in the main skin.  This bridle line is very stretchy- even after each bridle has been individually tensioned to 2 or 3 kg, there is still a lot of settling to come.  Weight is 250gm complete.

In stronger winds, terminally right wing- had to lengthen LH shoulder and 3A 20mm, shorten 3B 20mm, lengthen 2B and 3B 20mm to get almost straight- is this stretchier lighter fabric inherently more likely to distort asymmetrically?

There are now 5 of the larger 26 bridle SSSLs in good flyable condition, each one significantly different in some respect- maybe best to get some hours at events on these before too much more development- need to check light end for example.  Is tempting to try hooking the TE on 38 or 39 though, either by way of TE tightening or rib pleats.

Is also now possible to establish some formulas for bridle spacings and lengths?

11 Jan '15 SSSL40 orange: Increased spanwise length of lower leading edge by 45mm per side- previous did look to be too tight causing unnecessary looseness in upper LE.  Also reduced cutaway at LE of tips- now back to 90 degrees.  39 shows quite a notch at the shoulder bridles, but their spacing seems appropriate- have increased the width of the LE at this point by 20mm. Also tried cutting LE pieces on the straight not bias.

Cutting out- 30minutes to 45minutes?  All sewing took just less than 2 hours.  Bridling, total, including pre-stretching; 45minutes.

Light Norwest, slight drizzle, flies really well straight off- LE does seem better and shoulders aren't as notchy, but there is a bit of a flare out in LE of tip flare. Quite right wing in stronger gusts- but haven't checked bridle symmetry yet.

Seems to be flying at a bit higher A of A than previous- quite a bit stiffer and a little slower to respond to line pull in the light - is this because the leading edge is now 90mm longer, increasing forward area, so moving C of P rearward?  Tried shortening all LE bridles by 15mm and it does now feel more like 35-39.

No noticeable tendency to nose collapse yet- nor to instability- and this will have increased camber so decreased wrinkling in forward area, which was noticeable for 39. Need to check this in stronger and lighter winds.  Does now (15mm shortening of all A's) behave in no wind more like 34-39, that is, lower leading edge does fold in a bit during acceleration. Cut away LE of tip flares- a lot less flare out now, changed template to this.

15th Jan '15 SSSL41 pink; Made all new rib templates with difference between rear and front flare depths all ribs, now 50mm (was 40mm)- to shorten front bridles matching SSSL40.

Light gusty easterly; does fly OK but appears to be significantly SS caused by quite noticeable creases at B and C (less) during gusts: noticeable chordwise compression.

Let out all B's by 20mm, chordwise compression much less now flies much better, reasonably determined right wing though- as is 39 and 40. 

Tried shortening all bridles by 720mm to increase camber (after restoring all B's back to original length) Almost as good as with full length bridles, B's lengthened, but crease at B again- so let out B's by 20mm again and it's probably now flying best- does seem more volatile, less inclined to stall in gusts.

After some more testing of this version- need to try it in strong wind, 40 also- and if the increased camber hasn't caused instability of some type, should try pleating ribs to match camber again, see if this can now help. Damaged in a tree, repaired; restored bridles to 3.1m, not worse, maybe better.

16th January, 42 green: had to lengthen centre A's by 15mm to reduce LE collapse, but then flew well at Nelson - a bit right - and earlier kites, back to 34, noticeably flew as well or better in the quite strong gusty onshore winds there - with some exceptions.  39 gold 30gm soft fabric was a stand-out- very straight and seemed to get better as the wind became stronger. There seems to be a limit to the efficacy of tuning by lengthening/shortening shoulder bridles- possibly the resultant Aof A changes overlap with LE inflation in their light wind/strong wind effects- whatever, neither shoulder bridle tuning, nor outer rib B (camber changing) are always sufficient.  Shortening the rear outer rib bridle on the right to pull the kite right does always work though- unlike for parafoils.  I presume this is because the SSSL's are always at high A of A.  This shortening is just like steering a 4 line kite- but does seem to have slightly opposite effect in light winds sometimes- why could this be?- doesn't fit any theory I can think of.     

19 January '15; 42 orange (After Nelson); 3.0m bridles; cambered centre and 2nd ribs 40mm by a vee between B's and C's, also added 15mm to A's centre to reduce nose collapse.  Light to strongish gusty easterly at 105.  Spanwise creasing is noticeably less than for earlier prototypes.  This one's pretty good. "Natural" camber for 3m bridles is 60mm for 1.2m between A and D so  40mm shouldn't create any more camber than the bridles are requiring, but should decrease rib flutter- which it has.  Noticeably less flutter in rear panels, and sometimes flies at a very high angle.  Do they now have a greater tendency to "roll-over" instability (a consequence of progressive attached flow over the cambered top surface as A of A decreases)?  Not sure about this yet, but later prototypes do seem to have an apparently increasing problem with sensitivity to tuning- they can dive off unrecoverably to one side or the other in gusts  . 

Shortening D outer on the preferred side by 15mm +/- does seem to be effective tuning in stronger winds.

One possible solution to "roll-over" could be a flap strip on the upper surface at around 1/4 chord.  Another idea worth trying could be to make the rear 25% of the skin in flag type porous fabric; would be heavier to reduce super stability, porous to improve stability in stronger winds, and more resistant to flapping fatigue.  

To January 31 '15; Goa-Pune-Dubai.  35, 39, 42, 43  All performed well.  In very light winds they are definitely a match for standard pilots- and they hold their pull right down to when they are about to fall out of the sky, unlike ram air pilots.  They pump so easily it's possible to keep them up in almost nothing by active flying- also unlike ram air pilots. In very light conditions, 42 and 43 have a noticeable tendency to figure eight automatically (basically an effect that is a function of size and aspect ratio, everything else being unchanged). This is very likely to be improving their light end by generating apparent wind speed.  39 did this a bit less- why?   Less rear camber? When they do get pointed down though (very rarely), they can only be recovered by releasing line in classic Rokaku technique.  This would be desirable if they were meant for fighting with but isn't good for pilot style kites- they do need to fly in upward curves not straight lines. Earlier versions did this; how to get this back without losing anything else?

Currently, to fly in strong winds, tuning needs to be very precise- and they don't seem to stay the same from one day to the next.  The marker instability is to hang increasingly off to one side or the other, and very occasionally this will progress into a slow inexorable dive.  Maybe this isn't classic super stability, but rather an effect by which the LE on one side of the other gets progressively pushed in as the kite leans (caused by the apparent wind not hitting the LE squarely in this case) creating ever more asymmetric drag, which increases the lean. Perhaps try cross venting (again) to equalise LE pressure- this should have an opposite and restorative effect by firming up the downside LE relative to the upside LE.  Maybe taking the shoulder bridles through a pulley would do this too.

To March 10th 2015; Pasir Gudang, Malacca and Satun.

Four flying days at Pasir Gudang: had, as usual, every wind every day- some times every five minutes.  Sudden sweeping gusts, wild wind swings, periodic lulls. In 2015, probably no winds above 60km/hr for more than a few seconds though, but the most consistently turbulent kite festival there is; an excellent test place for the SSSL's at this stage in their development.

Malacca one day only; wind was very smooth mid-range, dropping slowly to nothing by 9pm.

Satun had every wind every day, long periods of thermals, no identifiable direction, more variable even than PG but generally lighter except for a last hour last day thunderstorm with rain and estimated 70km/hr peak.  The SSSL's were exemplary, on average as or more reliable than conventional pilots, enough pull to hold the maxi kite train up even in lulls, rarely requiring re-launches and much less moving around.  50(blue, hard 30gm), 51(yellow, soft 30gm) and 52(red, 45gm) did each require 4 or 5 tuning adjustments before they would fly straight (or a bit to the desired side) in the strongest conditions and 52 did hang and dive strongly right when launched on the last day, but when adjusted 10mm or so, took off terminally left- and the tuning for straightness then had to revert to exactly as it had been the day- a bridle snag?- or was there more wind during this first 15minute period?  I can't be 100% sure that this wasn't a "random leading-edge event" though- in which the LE buckles in asymmetrically on launch, and stays that way until the kite gets into lower apparent wind- which it never does in the case of a single full-power dive to either side.  On the second day at PG, 50(blue) flew more and more to the right as wind increased during the day (not having been tuned for stronger conditions until then) until it hit the wind mill and was chopped up.  With hand sewn repairs and 100% new (overweight) bridles, for days 3 and 4 it seemed no worse for wear and was gradually tuned to just right of centre where it has since flown consistently in all conditions.  51 (yellow), soft 30gm seemed less consistent, much more difficult to tune than the other two, but eventually came good- though Volker, who used it briefly in light conditions at Malacca, said it was a bit right (had been straight in much stronger winds at PG).  At Satun 50 had 6 bridles cut and required some re-tuning after this but was then OK- maybe not quite as reliable in stronger winds as the other 2?  These three kites are all bridled with around 20mm steeper A of A than per patterns (to hopefully keep the LE's out a bit better without too much adverse effect on light end) and all have 3.0m bridles for a bit more camber.  They also all have V line shoulder bridles set at around 40-50mm longer at the skin than at the outer rib LE.  Tip bridles are all 20mm to 30mm longer than other bridles - more than this (say 60mm to 70mm) causes shoulder collapse. In minimal wind, 50, 51 and 52 all seem to be a bit more stally than the early 47 series or than 42 series kites- though two of John and Irene's 42 series were also flying at PG occasionally and weren't better- though these have been made to steeper A of A than my 42' s also.- should have tested against 43 (purple) which is original I think (and a 42 series). In very light winds, the heavier 52(red) held up better than the 50(blue)- 52 weaves, 50 doesn't.  They were good in very light winds when actively flown but none of them (including 51 and a J and I 42) ) stay up hands-off as well as some ram air pilots do- at Pune they did seem to. 

Of the events the 42 and 47 series have now flown at; Nelson, Goa, Pune, Dubai, Pasir Gudang, Malacca and Satun were all complete successes.  At Ashburton, Kuwait, and Kuala Lumpar, according to reports, they failed: Always: "they just dived off and crashed".  Clearly this is not to do with wind,  must be entirely that I wasn't at these events to tune them straight- amazing that no-one else appears to be able to do this yet (John Tan maybe?- but he was at Ashburton and in Kuwait).  So simple, and a big worry; I doubt that SSSL's can ever be made so as to never require tuning- pilot kites can't be and they are inherently much less sensitive (lower pull/weight ratio).

In the Satun thunderstorm, 52 looped once to the left, catching and tearing its leading edge in 2 places on an aerial, then 2 to the right and stayed up- VERY difficult to walk down- definitely more than 40kg pull.  In the extreme wind, centre leading edge was pushed in and also the left shoulder when it looped left, the right shoulder when it looped right.  Final bridling for 50, 51 and 52 was generally about 20mm  pulled back from equal lengths, tips about 30mm let-out from equal, shoulders around 40mm longer at the skin than the outer rib leading edge.  Tuning: the adjusting cords on the 47 series seem ideal- responsive but not too sensitive.  50, 51, 52 all finally required less than 20mm  to centralise.

 Some more reinforcement at each leading edge/rib connection may be sensible- all these kites did tear at some of these points- but only from catching on things in strong winds. 

An observation: In lightish conditions, 50, 52 (and 51 I presume) fly at a +/- 60 degrees without any leading edge collapse.  In stronger winds and gusts, their leading edges push in and they drop back to 45-50degrees- and become subject to diving-off incidents if the leading edge is not holding out symmetrically and if they are not tuned straight.  The only explanation for this I can currently think of is that from about 25 km/hr, flow over the top surface becomes increasingly turbulent and is probably separating from close to the leading edge.  This would also explain the trailing edge flutter in stronger winds -that this is caused by decreasing A of a in stronger winds is clearly not correct- because the A of A is actually higher in stronger winds (lower flying angle).   An at least partial solution to increasing leading edge collapse as wind speed increases (which is the cause of most current problems ) is therefore likely to be by re-shaping the leading edge so as to retain attached flow to a higher wind speed. 

To try therefore; replacing the current semi-circular rib LE  profile with a smoother slower upper surface curve for at the forward 10% to 20% of chord.

Ashburton, March 11 to March 16 '15:

 53 (purple); Tried cutting the rib LE profiles to slope the upper leading edge down (by 25mm at LE top marks).  Flying was almost volatile unstable in light winds, and it did luff once or twice without recovering.  No noticeable improvement in LE compression at all.  Appeared to have a significantly lower light wind threshold, but  dived over worse than 52 et al.  Shoulder bridles were 440/460mm

Then tried 54 (green) with rib fronts back to original height but parabolic rather than semi- circular and using a full 2 piece LE with widish lower surface.  Eventually 52 flew reasonably well in mid-range gusty winds, but the lower leading edge appeared to push in a lot in gusts, distorting the front flares. Shoulder bridles 440/460mm

Then re-built 53 (purple) to the same parabolic leading edge profile as 53 but with scalloped lower leading edge. This appeared to be a substantial improvement- though probably still not back to as good as 50 (blue) in strongish winds (but 52 hadn't been extensively tuned at this point and is also a bit huckery with all its sewn in pieces).

55 (pink), a tidied up version of the modified 53, was wildly off (asymmetric LE collapse) when first made.  Letting the shoulders out in steps to an eventual 40mm  (490/510mm now- try this for 53 and 54) helped a lot, but noticeable tension creases at the outer rib leading edges appeared to be initiating LE collapse (mid-range gusty southerly with rain). 

Later on March 16; added 60mm Vs to outer edge of leading edge just outboard of the outer ribs.  In gusts the shoulders then showed flappy "ears" about 200mm out  from the outer ribs (worse than they had beem before the 60mm Vs were added).  Then took  a 2x12mm V out from 250mm further out along the shoulder, which corrected this (tried 2x25mm first, but this was too much).

Tip flares were still bulging out between the rear shoulder bridle and the tip bridle though, so took two V's of 2x12mm out, reducing this a lot. Very much improved LE appearance and flying now: LE  holds out quite well in up to moderately strong winds at least (say 20kg max pull). 

These changes have been made to the 47 series patterns, but 55 needs to be tuned a lot more and tested extensively against 50 and 52 to be sure that they are actual improvements.

Singapore 17 March'15 : Observations and questions

*Letting out D's on all ribs by 25mm improves light wind flying and recovery across the range but doesn't seem to have any noticeable effect on LE form, though it does promote creases at C which can become destructive. 

*Pulling in on B's  by 20mm DOES seem to improve LE form- and without noticeable increase in stalling- but this makes creases at B which can also be destructive if they are big enough to initiate general chordwise compression in strong winds. 

*Using spanwise arch and more shoulder/ tip bridles is the traction kite solution to eliminating spanwise creases without adding more bridles.   But would anhedral effect SL stability?  Sleds CAN be SL stable, but arched down shoulders have been unmanageable in earlier prototypes, so this needs further testing.

*Letting out tip bridles eventually causes shoulder collapse- which is a much more severe cause of diving over than compression of the LE at the ribs.

 *Did 53's first iteration, when it had substantially more "nose down" (and a shallower LE) show reduced separated flow (less flapping) as expected from Satun notes?  Perhaps, probably even; it did seem to fly at a higher angle and did seems to flap less (rear of ribs still did flutter though).

*More "nose down" did improve light wind performance but caused worse LE collapse in stronger winds. But if this had been combined with an appropriate A of A increase?  Why hasn't adding camber in the forward 20% of chord ever been beneficial?  It shouldn't have C of P consequences and should  reduce spanwise creasing at A and B- but has so far always seemed to make LE compression worse. 

*Using even more LE sweepback seems likely to be beneficial for reducing the effects of LE compression- the theory being that the shoulder areas become easier to hold out as they sweepback more because the external pressure is progressively less than stagnation pressure .  Also, when the part of the leading edge that is square to the wind direction is only in the centre, when it does compress it doesn't cause the kite to dive off to left or right as much.  Pointier noses are a suggested direction therefore.

*What effect does trailing edge shape (in plan) have?  Is the "Flowform"  style concave TE useful stability-wise?  At the least, a straight or concave TE should reduce flapping- and this change could be combined with a pointier nose while retaining the current ribs- for 42s and 47s.

*What is the optimal LE depth relative to chord?  Too much and its lower edge just folds in, too little doesn't seem to have as much rigidity. 55's parabolic LE form does appear to be an improvement in that this kite seems to fly at a higher angle, especially in stronger winds.

*Would using LE bridles rather than keels/flares reduce LE compression?  It should do, at the least because there would then be the minimum possible rearward pulling component of bridle tension (is this really true?). Using lines rather than fabric would also eliminate the drag effect when the flare LE's are folded back by LE collapse. 

*Or should the forward rib flares overlap the leading edge a bit so as to support the ribs to nearer their leading edge?  With scalloping, moving the leading edge of the front rib flares further forward wouldn't seem to have any other consequences anyway- just cut the leading-edge pieces back a bit to suit.

*What is the optimal slope angle for the front rib flares?

*Wrinkles and creases in the leading edge, especially at the shoulders, act as an initiator of LE collapse. The outer rib LE area on 55 can be cut better- but current templates should be better already than 55 is now (and how does 55 compare with 50 and 52 in this respect? )

*Using heavier hard finish Dyneema bridling is MUCH better- never tangles and won't get cut as easily as various SSSLs did at PG-Malacca-Satun either.

*Call the 47 series "One Skins"- retain the name "Boomer' for the 42s?

Chengdu, March 19 '15.

John Tan says the 42s they make are now reliable to 35km/hr, only have problems above this- and this is probably true also for 47s- all of them are OK to 35-40, but to get reliable flying above this takes intensive (and not always obvious) tuning.

Essentially, where they are now is that some leading edge collapse in strong winds seems to be an inevitable consequence of having adequate light wind performance. I can't currently get the full range except by intensive and time consuming tuning of every individual kite (and not always then?) - and the light end is much more important for most users than being able to cope with winds above 40km/hr.

But enough improvement in strong wind reliability to 'complete' these designs may now be available from small changes: 

*Wrinkles and creases, especially diagonal tension creases, are almost certainly initiators of leading edge compression.  This is supported by most kites improving after they've been used a few times in strong winds- they get stretched in?- though kites in soft stretchy fabric don't generally seem to be as resistant to leading edge compression.  A 3D program would be very useful at this point- but staking the kites off at bridle length in strongish wind and marking changes required will suffice. As for an earlier change in the 42 leading edge, de- creasing the leading edges may also require small compensatory bridle changes.

*Should the forward edges of the front rib flares align with their bridling angle?  As LE compression progresses, the forward edges of these flares folds back, increasing drag in a vicious circle.  Steepening these flare leading edges should reduce this- even without going to individual full length bridles to each rib LE. Making this change would increase bridle spacing over the rest of each rib by 50mm or more- which will make spanwise creasing in the central skin areas worse-  but this could be offset by reducing TE convexity at each rib by the same amount- which should also reduce flapping.

*Scalloping; too much or too little? It may be that scalloping is only useful currently to diminish the consequences of LE compression- if LE compression in stronger winds can be reduced by other changes, scalloping may best be reduced also.

*The seam between the leading edge piece and the main skin may not currently be ideally placed - move this rearward a little to get more taper in each cell LE?     

* Further refinement of the leading edge profile? Perhaps the apex of the LE at each rib could be sharper still- and maybe raised a bit to keep the upper leading edge camber about as now? 

* The shaping of the leading edge piece around each shoulder bridle can be improved- maybe even small triangular extensions at each bridle (say 50mm or less so as not to risk spanwise collapse from aero forces on them).

* Chordwise camber, currently an arc, is unlikely to be optimal.  Try progressively reducing camber towards the TE, dealing to spanwise creases by reducing TE convexity.

* The tip flares are not well shaped; they bulge out between the shoulder bridles and the tip bridles.  Moving the tip forward to reduce this will just exacerbate TE flutter,  The cause may be that the leading edge of the tip has been progressively cut away in the 47 series to reduce this bulge but could be having the opposite effect.  

* The relationship between the tip shaping and bridle length, and the shoulder bridle lengths and the shaping of the leading edge piece through to the outer ribs needs further exploring/refinement.  

Flying at Chengdu, 20th, 21st, 22nd March '15.

 50(blue), 52(red), 53(purple), 55(pink) and the first sample ex Kaixuan (red) all flew very well in light and shifty winds- as soon as there is just enough to keep them up they have really useful lift- probably more even than the 13sq.m Airbanners Lifter (blue) until about 25km/hr.  Didn't fly 54 (green) which has a full lower leading edge.   55(pink) was noticeably better in the lightest conditions than any of the others- usefully volatile too- but would fall before conventional pilots if untended.  If actively flown it would stay up in lighter winds than any pilots ditto. 

All of them showed various degrees of leading edge compression in stronger conditions (on the 22nd there was an hour or so of perhaps 40km/hr- a Tattoo ray plus 52 required  3 or more persons to take the tension.  55 would fly OK but, was inherently straight, but seemed to hang to left or right when tried on a short line, so I used 52 instead as it's a proven flier (Satun etc).

At Chengdu, users found them to be excellent in light and mid-range winds but were put off by uncertain behaviour and went back to using their trusted ram air pilots in strong winds.  To some extent it was that they didn't let them up high enough to get clear of the wind swings behind the trees- were mistaking bad wind for kite misbehaviour- as I maybe did with 55. 

Clearly, the strong wind end needs to be more reliable- even if it's just to a cut-off point, like "OK to 40km/hr, not suitable for winds of more than this" (Tan's and Lee's suggestion).

Current SS traction kites appear to be able to fly in higher apparent wind speeds and at higher flying angles and lower angles of attack than these SSSL's without collapsing their leading edges - why is this?

1. There doesn't appear to be significant differences in profile- maybe the traction kites have a bit less camber- but the earlier 42 series were also much flatter and they still had LE compression at <40km/hr.  Camber was added to reduce mid chord skin creases which traction kites avoid by spanwise tension. Successful SS traction kites do all seem to have flat or even reflexed trailing edges

2. Upper leading edge camber as tried on 53 (first variant) didn't help- unless bridling angle also needed to be steepened.  But if flow separation is the cause of LE compression in stronger winds, then scaling down to half size should discover this.  

3. SS traction kites are bridled significantly further rearward- as is evident from their higher stall speed, but Michel has said that he's bridled one so far forward that it floats- in preparation for designing a de-power version, and in this mode it didn't have leading edge collapse either.

4. Can it be LE bridling that is the difference?- SSSLs use fabric flares, SS traction kites use direct lines only.  Needs to be tried- just add 4 more bridles to a 47 series.

Ashburton, 25 March to 1 April 2015:

56 (purple); tried front flares with their leading edges exactly aligned with the front bridles. Also, removed camber from rear 30% of chord- around 50mm up at trailing edge. Flew OK in light wind but LE's of front flares pulled back out of line with the front bridles, causing a camber hump at 5% to 15% chord and more LE compression in stronger winds.  Its TE flapped a lot more than 52 et al and various mid chord-wrinkles proved difficult to bridle out.

57 (green); tried front flares with much more sweepback than 47 to 55 added riblets to each cell LE, and added 4 additional bridles directly to the lower leading edge at each rib- and retained de-cambering from 56, but with rear bridles (D's) moved forward 75mm. Could not find any combination of front bridle lengths for which this configuration was any improvement- actually flew best in mid-range with extra front bridles completely slack.  TE still flappy but moving D bridles forward reduced rear creases by a lot.

Then changed 56(purple) by moving rear bridles forward 75mm to match 57, and moved A's rearward by 50mm to reduce forward camber, and cut down nose height from 200/180mm to 145/135mm- and re-cut leading edge piece to approximately parallel from "peaked" at ribs. This showed great promise in light to mid (Wakanui)- the leading edge now seems more resistant to compression compared to all other prototypes so far.  Also, re-configured extra LE bridles on 57 to the riblets at the cell centres (3 only) and this also shows considerable promise- LE does exhibit compression in surges but distortions appear to be held to an acceptable limit by the extra bridles- and perhaps only two extras will be required as compression in the centre cell LE can't cause too many problems.

But 56 and 57 were still much less smooth than 52 (red) and Kaixuan (red) - though they do both appear to be less stally in very light conditions (Wakanui).  56 showed no terminal LE distortions and stayed straight to <40km/hr (tow tested)- ran out of road.

58 (blue); reverted to 55(pink) ribs (have rear camber) but with reduced height LE- semi-circular,150mm/140mm diameter at ribs, and with cell centre riblets.   Showed noticeable early onset LE collapse and a lot of rear chord creases.  As a check, measured the as-bridled profile and angle of attack (at centre ribs) of this one (58) against 56 and they are very similar; not even 10mm difference by overall angle of attack measured LE to TE, and barely 20mm difference anywhere in camber (58 had a little more in rear 30%  compared to 56).  So why is 58 much less robust in its LE? Could be outer rib and shoulder differences- check this.

Moving 58's rear bridles forward by 75mm (as for 56 purple) reduced the rear creases a bit (but will still initiate chordwise collapse in strong winds?)  Extra riblets do not seem to improve its LE.

59 (pink); Same as the modified 56(purple) but with wingtip flare bridles about 75mm further forward to reduce the bulging in the tips LE in stronger winds.

Wakanui, 1 April '15; Testing all of these again in mid range (10kg-15kg max pull), they are all OK, though 52(red) and Kaixuan (red) have many less creases and flap less.  52 and K also figure-eight nicely in threshold winds but seem more stally.  The shallower LE's of 56, 58 and 59 do not seem to have cost light wind end as I thought it might have (an angle of attack effect)- and these all seem to fly at a higher line angle than achieved before.  The visibly greater LE distortion of  52, and Kaixuan does not seem to cause more diving over.

When launched full length (30m) in smooth ground level 20km/hr+ wind , 52, K, 56, and 58 all arrowed off immediately to one side or the other and hung there until 500mm or so of line was pulled in and then suddenly released- when they all then came back to centre.  Neither 57 nor 59 did this- but possibly because they just happened to launch straight up.  This observation strongly indicates that diving-over is caused by gust induced asymmetric collapse of one side or the other which doesn't correct until the kite gets into a lighter wind area or is subjected to some other disturbance. This explains why these kites sometimes head off to one side and hang for a while in strong gusty conditions and then return to centre- or to the opposite side (55 seemed inclined to this at Chengdu). 

It may be that extra leading-edge bridles to the centres of the outer cells will fix this- or at least mitigate it enough. 

If the de-cambered rear sections of 56, 57 and 59 do indeed improve light wind performance- or if this de-cambering has beneficial stability effects, then what, apart from fine tuning of the rear bridle points and in the limit, cutting back their TE by 50mm or so, can be done to reduce the creasing/flapping?

Chongqing 4th to7th April and Valencia 10th to 14th April, '15:

Chongqing wind was very light and intermittent and mostly the kites were invisible in the ground level cloud (event was on Fairy Maiden Mountain) , so not a lot of comparative testing was possible.  Both 56 (purple) and 59 (pink) appeared to fly satisfactorily.  These have flat rather than cambered rear sections- which does cause more flapping but also doesn't seem to exhibit the hanging-off to alternate sides that is a characteristic of 50 (blue) and 52(red) which have been the most reliable SSSLs until now.

Valencia had light to mid sea wind- not that steady though and rising to approx 40km /hr in gusts on the final afternoon (maybe a bit more at altitude).

No testing of 57 (green) which has the extra LE bridles- because I inadvertantly sent this on to Weifang from Chongqing, taking the earlier 54 (green) with the extended lower LE instead.  54 (given to Rolf Zimmerman) actually flew reasonably well at Valencia, even in the stronger conditions, but with quite ugly and noticeable folding in of the LE.

58 (blue), given to Ludowico (Italy) did seem to fly OK but wasn't used as a pilot during this event.  It's extra leading edge part ribs don't seem to offer any useful improvement in leading edge form- but might do so if they are also bridled as for 57.

59 (pink) , which has a shallower leading edge, cut-back about 30mm clear of each front flare, was excellent- used as a pilot above the Tattoo ray both days, it did gradually lean left as the wind increased, but hadn't been tuned for stronger conditions and this leaning didn't appear to be of the diving-off type, but just a minor asymmetry.  Pull was 15 to 20kg when I eventually pulled it down. The tuning adjustments now fitted are very effective; not too sensitive nor too insensitive.  5mm to 10mm makes a noticeable difference.

Above the ST Ray at Valencia, I used 50 (blue) then tried a 12sq.m Airbanners 3 bridle Lifter- which seemed to have about the same pull.  Later on the last day, when the wind was strongest, the AB Lifter misbehaved (why?- I have previously flown this kite in stupidly strong gusty winds) and I went back to using 50, which was satisfactory but did seem to hang alternatively left then right?  59 does now seem to be better than 50 and 52 in stronger winds.  59 is good enough for production now.

Light wind: 47's are a match for any kites when actively flown but come down 1 to 2 km/hr before the best conventional pilots when left tethered.  Theory suggests (and observation tends to support) that deeper leading edges help light end performance.  This would be because a deeper leading edge effectively reduces the angle of attack of the kite's lower skin.  But deeper leading edges do seem to cause problems in stronger wind, by folding in and creating sometimes asymmetric drag, especially when they push in on the front flares.  It should be possible to make a leading edge that folds back evenly and without contacting the flares as the kite's angle of attack decreases though.

Bernhard Dingwerth had a 42 series copy in white Icarex at Valencia that was clearly superior in light winds.  He says, and I could see, that it was not just better than any conventional pilots but equivalent or even better than Genkis for example.  I couldn't see what differences there were by just looking at it, but there must be some; need to measure it.  He says it's no good in stronger wind though- so I wonder if it's just bridled to slightly lower angle of attack. At least I now know that superior light wind performance is possible from this style of kite.

To try at Weifang: 

1. A deep leading edge for better light end that folds back in stronger conditions without pushing in on the flares.

2. The usefulness of extra leading edge bridles in stronger winds  by testing 57(green), and maybe by adding another pair to 59.

3. The effect of cutting back the TE so as to reduce flapping (100mm to 150mm?)

Further reflections after Valencia (sitting in a hotel with nothing else useful to do) :
Another cause of hanging off to one side or the other is the tip flares having loose leading edges - one or the other billows out and holds the kite to one side. 50 and 52 have this problem to some extent I think.

Towards some useful principles for SSSLs (not in any particular order of importance as yet, and with comments added during VungTau later in the month):

1. That a flat or slightly reflexed rather than cambered rear section does not prevent leading edge 'inflation' and provides for better stability, (especially recovery) and less hanging to either side; by pushing the centre of pressure forward relative to the centre of gravity. This should be true, except that the 59 type (flat rear section) are more inclined to hang off to one side than the 52 types, which have slightly cambered rear sections- and flap a lot less.

2. That shallow leading edges (relative to chord) make it easier to keep the leading edge from indenting in strong winds (and makes the effects of any indenting less likely to cause "diving-over") - but deeper leading edges provide better low wind performance (by decreasing the lower surface's effective angle of attack).

3. That it's much easier to avoid compression caused creases spanwise than chordwise. Chordwise camber  should be equal or greater than provided by the bridle radius- and even then some spanwise creasing will be difficult to avoid .  Any more than the minimum necessary camber towards the rear causes stability problems by pulling the centre of pressure rearward (good theory but see 1 above).

4. That close spaced cells (spanwise) makes design of the leading edge much easier (can be just a straight panel), but that wider spacing is achievable with careful panel shaping- and is desirable by requiring less total bridles.

5. That there is an angle of attack, which when bridled for, enables every SSSL to fly first try. For tailless 1Skins and Boomers this is almost exactly having all the bridles the same length as measured from the skin- it's only the leading-edge depth that biases the kite's angle of attack sufficiently towards the LE to enable flying.

6. That as a rule of thumb, lines drawn from each bridling point at an angle of 30degrees (chordwise) to the bridle line should overlap at the skin on adjacent bridles to prevent humps and minimise creases.

7. That bias cut ribs with cording sewn only on a line extending the bridle direction, and only part way to the skin is an excellent way to spread bridle loads.

8. That anything more than a semi-circular leading edge reduces captured pressure in the leading edge- that is, extended lower leading edges are undesirable with respect to LE form. Advantages/disadvantages of scalloping?

9. That, for the same camber, longer bridles reduce chordwise creasing.  Bridles lengths just greater than chord x's 1.5 seem adequate. For convenience, flares and the shoulder flares are best dimensioned so that all bridles are the same length.

10. That pressure is maximum in the centre cell, reduces proportionally in outer cells and shoulders- hence cross-venting is not sensible, unless also valved (and probably not even then because flow rates could not be high enough to make a difference). That this is so is shown by the keels all bulging outwards when the kite is flying.  (But the rib sections closest to the leading edge show this effect least- almost not at all actully- so there's still a question mark on cross venting).

11. That LE form (pressure behind) can more easily be improved  by shortening "B" bridles than by hooking the TE- effectively this is usually best done by lengthening "A's" by a little - but this also increases the overall angle of attack, so should be matched by letting the TE out a bit.  Spanwise creasing- from chordwise compression- does limit the use of this unless more bridles (chordwise) are used or all the bridles are also lengthened.

12. LE Sweepback is useful in that this pushes LE indentations away from the tips and towards the centre where they are much less likely to cause "diving over".

13. Some SS traction kites don't suffer from LE push-in even to 100km/hr and more apparent wind, so holding LE form across the required wind range is also possible for single line single skin kites- but is this compatible with the profiles and very low stall speed that SSSL's require in order for them to be useable?

14. Highly cambered sections- like the 4Skin and NPWs can have very low stall speeds (less than 5km/hr) but the rearward camber that these kites have probably makes across-the-range single line stability unachievable- because their centre of pressure is not far enough in front of the centre of gravity to enable good recovery in the single line sense.

15.  Adding stiffening rods (usually weed whacker cord)- or using stiffer fabric or stiffening patches can push the onset of LE compression further up the wind range, but fundamentally, using stiffening to do this will always just put off the problem- better, at this stage, to concentrate on getting more pressure in behind the LE- this is true also for SS traction kites.

16. Lighter fabric does not necessarily help light end flying- for the 42's, heavier fabric causes useful figure- eighting in very light winds which builds up their apparent wind enabling them to fly in lighter winds than kites made of lighter fabric do- everything else being equal.

Weifang 15 to 22 April '15;

Made K2 (yellow) , identical to 59(pink),  K3(gold) as for 59, but TE cut back by 100mm at tips, center and outer ribs, K4(grey) 59 patterns but with "adaptive" 90 degree LE, K5 (blue), 59 pattern with 16 bridles, reduced chord , approximately  straight TE, and K6 (grey), 59 type.

At the Weifang festival: 59(pink) and K2(yellow)  flew well in lighter conditions but dived off to both sides uncontrollably in strong winds- almost seemed to be volatile instability, as their leading edges, while indenting noticeably at their lower edges, do not seem to compress in at the upper leading edge seam to any significant extent.  K4 (grey), has this behaviour but even more so. 

K4's leading edge is designed to project downwards at 90degrees or less , but only when the kite is flying at high angles of attack (low line angle).  This is to decrease the effective angle of attack of the kite's lower surface in stall conditions so as to improve light wind performance, while bending back to more conventional form progressively as apparent wind speed increases.  This does appear to work, though needs a careful comparative test with say 59 or K2 to be sure (test flying in strong smooth wind at Vung Tau on the 23rd nd 24th  April appeared to indicate that K4's adaptive leading edge cures the hanging-off problem that all the other 59 types have by comparison to type 52s- which is really strange, the effect should only be lower stall speed, because in stronger winds their leading edges take the same shape)  This  "adaptive" LE must (can here be any other reason?) somehow be the reason for K4's greater volatility, (it's the only significant difference between K4 and 59, K2).  Could it be that it causes K4 to come out of stalls (which did often enough occur momentarily because downwind of the Weifang Stadium roof was very turbulent) earlier and with greater speed.  Sounds like a tenuous theory, but the only one currently on offer.

K3 (gold) and K5 (blue) were both seemed more stally than the "59" series- as I had expected because their TE's had been cut back 100mm and 200+mm respectively.  (But shouldn't this make them less stally because their lower leading edges are proportionally deeper relative to chord, thereby decreasing the effective angle of attack of their lower surfaces). But K3 and K5 seem to hold their leading edge shape better in stronger winds than type 59's and even type 52s.  Is there something else happening here or are they actually not more stally?  Need to make a more careful test of this.

K5 with 3 bridles instead of 4 to each centre rib did not appear to be more creased than versions with 2 more bridles- suggesting that there is the possibility of reducing to 16 total eventually. 

On the 18th (frst day) at Weifang, of the kites tried, (K2, K3, K4, K5, 52 and 59)  the only 1Skin that was really reliable when the wind was at its worst was 52red (but a lot of conventional pilots were not managing conditions very well either).

With 50 (blue) 52 was also the most reliable at PG and Satun- better than earlier styles and, it now seems, better than all subsequent prototypes ( a step backwards then- which is typical enough of this stage of development  and in some ways a hopeful sign in that it suggests the 1Skin design may be nearing an optimum.).  But the "52" series do have a LOT of leading-edge compression in stronger conditions- and I theorise that their stability in stronger winds is deriving from the resulting extra leading edge drag.  This would explain why they are so difficult to tune for stronger winds - if one side pushes in even marginally more than the other they crab off to that side, and when any asymmetric compression does occur (might be initiated by an uneven gust), it doesn't seem to pop out until the apparent wind on that side drops substantially.  The behaviour of the 42 series fits this theory quite well also.

Unfortunately,  before Weifang day 2,  I sold 59, K2 was stolen, we couldn't find 56, and 52 had all its bridles cut by locals using Kevlar lines (had a maxi train cut away also), so I had to revert to conventional pilots  (not very satisfactorily either, most fliers gave up and went back to the hotel- the conditions were atrocious ).  However, K4 did fly very well in the lulls but looped uncontrollably in gusts- clearly volatile unstable- WOW!

I had not thought to ever get the SSSLs to volatile instability, expected to be condemned forever to battling superstability (which is much more difficult to cure). The reason for this is that holding leading edge shape at low enough angles of attack for satisfactory light wind flying seemed sure to cause destructive leading edge compression in stronger winds (and leading edge drag is a prime cause of superstability- as is having too much lateral area in front of the kites centre of pressure- and having too much tail). 

Volatile instability is easy to fix- just add appropriately placed lateral area (deeper keels) or a tail.

20th , afternoon, at Kaixuan: Test flew K4 again in the factory yard (strong gusty wind and rain)- and did get it up briefly, but was lucky to get it down again without losing it in wires or on a roof.  Then added a 10m x 150mm wide heavy fabric tail. It was now remarkably gust tolerant and stable- and this is a place where flying conventional pilots is not practical, even in good conditions.   Need to test this further at Vung Tau of course, but strong grounds for hope (if a tail makes a kite more stable then it must be volatile unstable, because tails always make superstability worse).

21st at Kaixuan: Made and test flew (in the same impossible place) a 1.6m x 9m single skin single line Octopus kite: O1. This has the K4 style adaptive leading edge (constrained from billowing too far forward by bridles rather than by the fabric gussets used on K4) and no keels or ribs- just 64 (?) bridles on a 200mm(spanwise) x  225mm (chordwise) grid.  The leading edge projects downwards 170mm . Without the tentacles it appeared (not stable of course) to be comfortable (not at all stally) with all bridles exactly the same length (about 3.5m).  After the tentacles were added the rear bridles were lengthened progressively towards the rear, 20mm per, 160mm total, to achieve the same flying feel. Only tested in the Kaixuan yard (which has to be the worst imaginable kite testing place anywhere) but it does fly- see it now on Facebook!

How well does it fly?  More testing at VungTau this weekend should answer this. 

Vung Tau (Vietnam), 22 April to 24 April '15 

K6 grey (type 59) definitely hangs off to either side in stronger winds to a greater extent than 52 red, but K4 grey looks to at least equal 52s reliability in stronger wind. The light short tails I bought don't seem to make any noticeable difference to any of the SSSLs I have here- either because they're too small (K4) or because the kites are towards superstable (56,  K5, K6- didn't try one on 52).

I gave 56 and K5 away to other kitefliers and unfortunately lost K4 when it rolled away from my bag in gusty winds and seems to have immediately found  new owner (we were flying in an enclosed area).  No more useful testing of the 1Skins will be possible here therefore, but first job in NZ will be a new K4.

O1, the first SS Octopus shows remarkable ability to hold its form- including the leading edge shape. "Inflation" is excellent in the entire range bridling range from all set to the same length right through to the rear bridles being 160mm longer than the front (with intermediate bridle lengths progressively longer.).

Unfortunately, it's also strongly volatile unstable.  I should have anticipated this; it's because the head has the typical SS characteristic of 3xs the pull of a parafoil. while the single strip tails have nothing like as much weight or drag as tubular tails.   In the strongish winds (taking down a Tattoo Ray later in the day was a minimum 3 person job- actually, Andreas and myself plus 4 or 5 hangers on), the longest flights I could get (less than a minute) were by setting all the bridles to exactly the same length.  Strangely, letting off the rear bridles completely (thereby making the "kite" smaller and the "tail" effectively larger seemed to make it less stable.  Needs longer tails- 3 x's?- but holding the head shape seems to be the least of the challenges with this design now.   

Kelantan 26th May to 1June;

1Skins; No new ones from Kaixuan yet- patterns were held up in Hongkong for 8 or 9 days, but there were still always more Boomers and 1Skins being used at Geting Beach than conventional pilots- 8/2 at times, but a lot of them had strong leans- people either don't know how to tune them (probably), or think it's too difficult, or else the kite is up already and they don't want to pull them down.  Ron Spaulding took a lot of persuading before he would even try- and yet it's such a simple adjustment-he wanted to give back one I'd given him at Vung Tau and wait for a new one. Must get printed tab with instructions onto every kite. 62pink was initially strongly right, couldn't work out why, lengthened the  bridles back to 3m from 2.5- but didn't fix it.  Ron then noticed a pulled thread on one longitudinal seam- with this corrected , was straight again. Otherwise seemed perfect, wind never strong enough to cause LE collapse. Does require more wind than the 20m octopus SS below, but 1Skins pump so well and have so much pull for size in light winds that people are really taking to them.

20m yellow SS Octopus is outstanding in light unsteady winds; In the offshore winds typical at Geting beach in the mornings it flew reliably on 150m line, often in the company of only Helmut Georgi's Genki.  It is not quite as good as Helmut's Genki however- but only because it doesn't respond to pumping, either flies or doesn't.  Much better than any pilot kites ; the various 4 cell pilots and lifters, 1Skins, Boomers and the exceptionally good light wind PG 8sq.m.  It really is exceptional- I can't recall any other new single line kite design that I've done which has performed so exceptionally so early in the development process (the first prototype was built after Weifang 2015).  Letting out the rearmost edge bridles 100mm appeared to improve both light and strong wind flying.  And it didn't significantly fall-off much even while being pulled up in zero wind.  In stronger winds (25km/hr?) it was volatile unstable (would loop), and linking all the tails didn't appear to improve this at all. Unexpectedly, after shortening A and B sets (at the tertiary bridles) to the same length as C's, D's and E's it did then fly satisfactorily in at least middle winds on the 30th - A's had been +60mm, B's +25mm.  On the 27th, did try lengthening A's and B's until it would fly OK in probably 30km/hr; A +150mm, B +60mm.  Had ferocious pull on this setting and did 'stagger' a bit but was then stable. With A's and B's shortened to the same length as C's etc , the leading edge pushes in a lot in gusts and when overflying, but this doesn't cause terminal instability up to at least 25km/hr.  Max pull is perhaps 30kg in this wind/mode.  On the 29th at Sudara, tried the half sized pink Ist prototype in winds of around 20km/hr- and it was terminally volatile unstable on 10m line length- also unexpected because in Vietnam it would fly OK in much stronger conditions than this- what had I changed except for linking the tails, which during Wakanui Beach tests seemed to be a stabiliser?  Three possible explanations:  First; was the wind at Geting actually stronger than the wind at Wakanui because it was warm?  This might be true, but the Vietnam wind was much stronger than either.  Second; could there be a mid-range when this kite is unstable- maybe it becomes stable again in even higher winds- and this would also explain the yellow 20m  Octo at Geting on the 27th?  This is not impossible- have I ever seen this before?- can't remember a specific occasion, but think there may have been, easy to check for as soon as I get it into really strong wind.  Third; I had let  F,G and H bridles out progressively since Vietnam- and hand flying the 20m yellow at Sudara, pulling in on these back bridles does seem to reduce volatile instability in strongish winds- an ugly solution though, and needs further explaining because letting them off just makes the head smaller relative to the rest, which would improve VI i would have thought.  Anyway, it does seem that the 20m style (overlapping tentacles) and current bridling are better than O1pink.  Will be worthwhile trying an extended leading edge- current ratio of held/ free is around 3/1- try 1/1.  And try this only in the front leading edge area- with the sides maybe cut back to having no overlapping edge? Also, maybe, probably, tentacles will tangle in strong winds in sustained flying- Si fabric? - or will tentacle-end ties of  some particular length cure this?- Try.  Until a bridle system that works for every wind is developed (if any such is possible), a drogue should lift the upper wind range to whatever is required, and using a pilot definitely does- flies VERY steadily under a pilot (52red).

1 June '15; O2yellow sent back to Kaixuan with Tan Xinbo as a production pattern- will make 10, 20 and 40m versions.

Mongolia, 15th to 20th July 2015;  Three days flying; one period of an hour or two of turbulent mid-range winds, but mainly no wind except for occasional brief strong gusts.  Generally thermals only, no consistent direction.  Hagar's mega kite (red diamond, 400(?)sq.m) never flew during this event (many attempts).

1Skins: Flew 62 pink as a pilot during the only mid-range wind period. Seemed to be more volatile than usual and than conventional ram air pilots, with occasional strong movements off to the left, but generally OK.  In the frequent lulls, it came down more often than other pilots, even on 90metres, but had much more lift than a 9sq.m Airbanners (which inverted fairly often as did other pilots there, including largish flow-forms), so was more useful.  Actively flown it was a match for everything else when there was no wind- though one 8sq.m (approx.) inflatable lady bird kite was generally superior to everything else (except eagles), including the O2 yellow 20m Octopus .  Responsive kites (which this lady bug was) can be kept up in almost zero wind by active flying using the standard eagle kite type reels- which are now often used in China for even quite large kites in light winds.  Always Kevlar though- one of my 3.5mm Dyneema pilot lines now has 9 knots, and in this environment,  re-splicing is a constant job with the 300kg Dyneema I use with the 1Skins.

The 1Skins are now fairly settled; no major changes required.  The tuning system works easily and precisely- but not many other fliers seems to be able to use it.  Some are clearly unaware of its existence as they still try to use bridle adjustments for tuning.   The most useful improvement now will be extending the wind range- up and down.  Extending the depth of the adaptive leading edge would seem to be the best opportunity for lighter wind flying and this wouldn't effect anything else, so worth trying.  Seeing how both smaller and larger 1Skins behave in stronger winds should indicate how best to reliably extend he upper range.

Weifang, 21st to 23rd  July '15:

Test flew 1Skins; K1 1.5sq.m, K1 6sq.m, K25 3 sq.m. (all with adaptive leading edges). 

Main civic square in Weifang; almost no wind, then strong wind and heavy rain, then rain and no wind.

K25 3sq.m yellow and K1 1.5sq.m yellow appear to be satisfactory- didn't try during the brief windy period, but their responses were as they should be.  This 1.5sq.m (the first one) was very volatile until it had some line pull- then straightened up and was stable, as expected for a smaller kite.  It will be interesting to see what upper range this kite has- if it will fly reliably in stronger winds than the 3sq.m's, as I suspect, this implies that fabric stiffness is probably a significant factor at the upper end.

K1 6sq.m orange (the first 6sq.m), was flown straight-off without checking on a long line during the squalls to lift an ST Ray. In the strongest gusts it would eventually dive inexorable to either side (shoulder collapse)- and didn't seem that keen to stay central except in very light winds.  Is this a scaling effect (fabric stiffness as above?), or is there some scaling error or other influence.  A bridle check showed the shoulder bridles to be all 60mm shorter than they should be proportional to the current 3sq.m's- and this would definitely cause diving-off.  Lengthen these and re-test.  It's also possible that its behavior was just very bad inner city wind- the square is surrounded on all sides by sky-scrapers- though none closer than about  200m .  

SS show kites: The first sample 20m SS Octopus (blue) from Kaixuan looked generally OK but was terminally inclined to 'diving off'.  Visually the bridling appeared to be significantly different to O2 yellow, but I didn't have time to find the actual differences and correct them.  They have also used bridling and cording line that is far too heavy - the bridles alone probably added a kg (to a kite that should weigh less than 3kg total).  Easily fixed; and perhaps in future the bridles should be all individual rather than cascaded like for O2, S1 and this sample.  This would be quicker to do and have many less possibilities for errors.  Suggest cutting all bridles to the longest length, then marking in bands back to the shortest.  This would allow the person bridling to quickly tie every line without further measurement- just by counting back the bands for each position.  Left-over ends could then be cut off (the longest 'tail' will be less than 500mm). Should some capacity to change the relative overall lengths, leading edge to trailing edge, be retained by keeping, say four secondaries at the line attachment point?  Maybe there will eventually be 'strong wind' and 'light wind' settings, but I don't have useable ways to do this yet for the Octo or the Serpent because far too many changes are required.  The goal for now is; one bridle for all conditions- though at present, without changing anything for different winds, they're a long way from this with less than the best achievable light-wind performance (still amazing though) and leading edge collapse from about 15km up for the Octopus, maybe 25 for the Serpent (at probably some cost to its light end).

Both O2 yellow 20m Octopus and the S1 green 35m Serpent flew very well when conditions were suitable- their ease of launching is phenomenal - quite unexpected- just let wind into any part of the leading edge and they 'inflate' and launch themselves- provided they're pointing approximately upwards.  In light winds, pulling or running doesn't help much, unless there is more wind at altitude- they either fly or they don't.  The Octopus was more susceptible to leading edge collapse than the Serpent but seemed to require (very) slightly less wind- need to check this though, might have been because it was flying higher where the wind was stronger. The Octopus was also slightly inclined to diving over when the wind was zero- would slide off inexorably to one side or the other as it came down in the lulls.  The Serpent didn't seem to do this to the same extent- a function of its slightly greater spanwise camber?  Developing the Octopus and Serpent bridling completely independently has been an excellent approach- it is clearly identifying the effect of differences. 

Changes for the Octopus: Individual bridles as per above, taper tentacle ends and don't have a cross hem (to reduce tangling), try increasing the active proportion of the leading edge for less leading-edge collapse in stronger winds, and try one with Serpent bridling.

The Serpent had less leading-edge collapse in stronger winds but did appear to be on the verge of volatile instability at about when its leading edge begun to push in.  The lengthwise slots in the tail (to allow some airflow through so that the rear upper side of the tail does not become a big low pressure zone) have very successfully stopped the tail from folding over- and without the visuals being too detrimentally effected; a triumph for theory!   Appearance is generally good- plenty of fliers wanting to buy- but sharper colour contrasts will be better.  The eyes are too wide apart- one eye is partially obscured when looking from the side.

Changes for the Serpent: Individual bridles as for the octopus above, move the eyes closer together, try increasing the 'active' proportion of the leading edge for better light and strong wind flying.

Generally both these new designs are a very good beginning- will find market niches I expect, but the challenge is to increase wind range without adding adjustment complexities that fliers will simply not accept.

Using a 1Skin pilot above the Octopus did help its upper range a bit (was excellent in Kelantan), but in the briefly much stronger and gustier Mongolian conditions it was not sufficient- its leading edge collapsed so extensively that the Octopus would wrap itself around the pilot line when recovering.   I didn't try a pilot above the Serpent at this event because it was generally OK- only came down twice in strong winds .  Need to try it with a pilot, and also with a drogue.  Best will be if a combination of active leading edge changes and bridling can push the wind range up without, at the least, costing lower end. They're pretty good for such radical new designs already though and will get better.   

For the next style of SS show kite, to consider is that the Boomers, 1Skins, Octopus's and Serpents all have approximately semi-circular leading edges- is this contributing to these successes by comparison to the SS Ray and SS Flag which don't (yet?) fly satisfactorily? Also, currently the SS show kites (Octopus and Serpent) are both upper range wind limited (though I haven't yet tried a pilot with the Serpent or a drogue with either). The Boomers and 1Skins have better top ends but are not as good as the Octopus or Serpent at the light end (except if they're actively flown).  How to get more wind range? It seems that having lateral area stability (Boomers and 1Skins) helps the upper end (both the Octopus and Serpent appear to be volatile unstable in stronger winds- if their leading edges don't collapse first).  And possibly, tails help the lower end by preventing overflying and LE collapse when more forward bridling is used- though it's more likely that the Boomers and 1Skins would fly in much less wind with more forward bridling but would lose their upper end (which is barely good enough now). The Octopus and Serpent's lack of upper range could be corrected by bridling back (which would prevent leading edge collapse), but uless taken to an extreme, they may then become volatile unstable without either longer tails or a drogue. Except that I did try a rearward bridling for the 20m yellow Octopus  at Kelantan,  and this did enable it to fly in much stronger winds (lost the light end though).  And, the 10m pink O1 did fly in VERY strong winds in Vietnam, so perhaps it is just that the current bridling arrangements do not have wide enough ranges rather than there being any fundamental upper range limiting factor with these designs.  Too much pull is also a consideration.  More rearward bridling will increase pull- and they all have more than enough now.

For the next design, is there a show kite style that has some lateral area as well as some tail drag?

A caterpillar maybe (legs could provide lateral area)? 

Weifang,  21st to 23rd July '15;

Test of  K SS 20m Octopus number 1 (blue). after changing to individual rather than cascading bridling- and to July 21 bridle chart.  Main civic square in Weifang, almost no wind (1Skins wouldn't stay up).  Flew very well, maybe a little more inclined to LE collapse than O1 20m yellow   (but bridles are a bit different because of the change to individual lengths. Then tried with a mid-sized fish drogue attached to the centre two tentacles.  This clearly stopped leading edge collapse and extended the wind range a lot, but made it more susceptible to falling off to one side or the other in lulls. 

The Serpent and Octopus bridling are quite different- up to 300mm in places- but both kites fly quite well. This suggests there is further improvement to be had for both styles as they aren't fundamentally different kites (same heads). The Octopuses are slightly more susceptible to falling off than the Serpents and the probable reason for this is less central body spanwise camber (I still don't really understand why this should be so). In particular for the Octopus, bridles in the centre top head area are in general 100mm shorter; lengthen these and re-test with and without a drogue.

At Kaixuan, made a 5m wide x 11m long seven legs per side SS Caterpillar (3m wide semi-circular head that's very similar to the Octopus and Serpent heads but with 1 less set of bridles spanwise).  Not tested yet (no wind and still raining in Weifang) but looks promising because the legs, as well as providing lateral area (as do the mandibles), can be angled outwards (by re-sewing) to whatever extent is necessary to add drag for stability.  Weighs 2.75kg with 45gm fabric and heavy bridles.   

After Mongolia; Guangdong, Weifang, Shenzhen. 20 Oct to 1 Nov '15.

Octopus and Serpent: The Serpent currently has much more curve-down on each side; more sled like than the Octopus (Blue ex Kaixuan) .  The Octopus also has less camber in the mid-top body area.  Flown alone, the Octopus has more inclination to falling-off to one side in light conditions- making launching in light winds difficult.  At none of these events, did it show any tail tangling tendency- and quite strong winds at Shenzhen, almost more than I could pull down- 3sq.m 1Skin Pilot. 

At Weifang, added a crescent shaped ram-air inflated leading-edge element to the Octopus (360mm wide x 200mm deep, independent of the top skin) and the Serpent (320mm wide, 200mm deep using the leading edge and top skins as 2 of the 4 sides). 

The Octopus flew well straight away, had none of the leading-edge collapse in stronger winds it had previously had.  In a Weifang test - strong wind very wind-shadowed area- it flies by itself very well, at least as reliably as the 1Skins, more reliable than Craig's 8sq.m pilot.  It also flew well for very many hours with a 1Skin pilot at Shenzhen (light lifting to strong on shore).  Craig thought it was volatile unstable, but I doubt this, did see some sashaying when the wind was light but this had no tendency to increase as the wind built- could be either a light wind (falling-off) effect or a pilot effect.   It did seem to have less pull than before the ram air element was added.

The Serpent would not initially inflate- the extended leading edge caused flow across the inflation gauzes rather than blowing into it.  This may have been largely because the centre leading edge where the gauzes are is quite set back on the Serpent relative to the Octopus.  After taping off the leading-edge extension it did inflate OK- though still rather depressed in the centre and too far forward in the shoulders.  It then flew all day reliably- no volatility- under a 3sq.m 1Skin.  Didn't try it by itself.  Need to cut-off or sew down the extended leading edge (will this cost some light end?) and set the centre LE forward relative to the shoulders and test again. Perhaps the tail panels could overlap for the same functional effect as the current longitudinal gaps- but for better visuals. The eyes need to be closer together.

For the Octopus, clearly the Serpent style leading edge cavity is lighter and simpler- but will it work as well?  320mm width is sufficient. Need to adjust the Octopus bridling to a bit more like the Serpent to stop the light wind falling off, and test how far forward the bridling can be pulled before leading edge collapse again becomes a problem.

1Skins;

Both the latest 3sqm (yellow) and the 6 sq.m (orange) ex Kaixuan were excellent- did good service at Guangdong and Shenzhen (though not used when the wind was very strong and gusty offshore on the second day at Shenzhen).  Ron Spaulding and Andreas Agren both said the 6sq.m dived-off, twice, tangling them in the tight formation flying during the first day at Shenzhen.  I didn't see this, but it's possible- or could have been interference from another kite or just a patch of turbulence and a slow recovery. At Weifang on the 29th- strong turbulent wind, both these kites were much more useful and reliable than Craig's red 8 sq.m pilot- which was collapsing badly.

While at Kaixuan on the 29th, had a prototype 3sq.m  (blue, no number) with tubular ram-air inflated leading edge built.  At Shenzhen this showed no sign of leading-edge indentation (the presumed cause of diving-off) at all, and adds less than 5% by weight and cost.  This kite did not have tuning lines and was progressively left wing so wasn't tested in the strongest winds but is looking promising. Add tuning, test again- and try pulling the bridling forward to see if this kite will now accept a lower angle of attack.

An alternative way of doing this might be to just make the leading-edge piece of double thickness and let some ram-air in between the layers.  This is worth trying; and may have efficacy for other kites also- like to get earlier leading-edge stability on the ST Ray.

And what if the indentation that presumably causes diving-off is asymmetric because once the kite starts diving off to one side the flow centre/stagnation point is displaced from centre towards that side?

To test for this, try 2 /1Skins, both with ram air leading edges as above, but one with inflation central, the other with inflation at 1st rib each side and with no connection to the opposite side.

Could this also be a problem/solution for the ST Ray?

Wuhan 7th 8th November '15

Octopus and Serpent; The Serpent bridling at the leading edge was mis-rigged in various ways from hasty re-attachments after the active part of the leading edge was sewn down at Weifang after Shenzhen.   After correcting this to some extent, and reducing the amount of spanwise arch- which seemed excessive- the bridling was too low- would barely launch, even with a 1Skin pilot in the light winds. Looked good when it was up though.  Leading edge still does not inflate well- better than at Shenzhen, but not now nearly as good as the Octopus (which has an inflation cavity independent of the top skin). Maybe shortening the centre front bridles will fix the Serpent's inflation and stalling; need to fly it beside the Octopus at Wakanui.  Maybe there will be a useful advantage from adding an inflated central keel as well- for the Octopus also- or maybe with further refinement the double skin idea from the 1Skins above rather than a full cavity will suffice.  The Octopus looked and flew exceptionally well and at a high angle using a 6sq.m 1Skin- wraps only happened when the top kite fell back and the Octopus overflew the top kite line.  The tentacles never tangled if they started fairly clear. Sashaying does appear to be caused by having a top kite at a lower flying angle.  There are definite hopes for these kites in stronger winds now.

For bridling; all the secondaries should be identical length, variations only in the primaries- which should all be cut to the same length and tied centrally- and 500mm x 2  longer to allow for length variation requirements.   

1Skins:   Had K3/14 red, 6/11/15 with double skin leading edge, inflation at each 1st rib (no connection side to side) and K3/13 pink 5/11/15 ditto with central inflation built at Kaixuan.  Thes both look very promising- their leading edges showed no sign of indenting in the strongest puffs at Wuhan (not much).  Need to test in strong winds against K3/25 (sic) yellow 18/6/15 and blue no number with full leading-edge cavity.

The double skin leading edge does not appear to inflate to a tube ever, the ram air just fills the indentations between the rear and front layers.   Key will be whether strong wind diving- off is cured and if the angle of attack can now be significantly reduced for better light wind flying.

These kites are currently excellent in their window, but that window still isn't wide enough at either end of the wind range for them to supplant conventional pilots.  Their advantages are higher flying angle, huge pull and much easier launching and active flying.  In the light to nothing winds at Wuhan, a 6sq.m 1Skin could immediately lift either a max Trilobite or Ray up to flying altitude uninflated -the only maxi kites that did get up enough to fly (briefly) on the second day.

What can be done to improve their light end?  This is worth maximum effort as an even 10% improvement will likely make them the best of all pilot kites when the wind is alternating from nothing to very little-  the usual conditions at China events- and not uncommon elsewhere.

Is there any way to make a reactive bridle that responds to stalling by reducing the kite's bridled A of A for just long enough to get the kite moving up again?  Problem is that this is exactly what will also cause luffing - because the centre of pressure moves rearward in luffs and in stalls.   Such a mechanism is possible if it's active- like a human pilot, but can it be done passively with levers, pulleys, springs and dampers? Theory suggests not because to prevent luffing the rear bridles must be shortened - which requires energy input.  The active leading edge does seem to helping quite a bit to reduce stalling- if it's extended, will it help more without causing other issues?  It is compatible with double skin leading edges. Try this.

From 9 November'15

1Skins: At Wakanui, the Kite factory field and the Nelson event (January '16), the Blue (no number ) 3sq.m with the tubular leading edge, K3/13 pink (flat ram air leading edge with single central inflation) and K3/14 red (flat ram air LE with centre closed and inflation to each side) all flew well and appeared to be a lot less inclined to diving over than single skin LE 1Skins.  The blue ripped its LE at "A" bridles in a strong wind, was retired.  13 and 14 both seemed stally in light winds until the 4 centre "A's" were shortened 20mm, after which they were a little better in the light than previous versions, and quite a bit more volatile.  Both flew fairly faultlessly at Nelson- maybe 2 terminal diving overs during the day, total.  Both were also fairly straight without tuning and their LE's appeared to resist indentation much better.  After the "A" shortenings their LE's were noticeably pushed under along their lower edge, but still held out well. At Pasir Gudang to Feb 28 '16 the wind was very strong for all 3 flying days (holding on tothem was on the edge of not do-able, even with 4mm line.  They both exhibited collapse of the LE at the outer ribs until the "A" bridles at these ribs was restored to original length (the centre "A's" were still Ok at 20mm shorter than standard.  13/pink was used above the Ray, OLTrilobite, and Trilobite stack for one day but then started to lean right increasingly.  When it eventually came down, there were 2 damaged bridles (cut except for a few filaments) After fixing these it was violently left, needs to be re-tuned. 14/red then took over and was handling the winds well- until it dived over catastrophically and was destroyed on a pavilion roof.  Not good enough! But it is fair to note that conventional ram air pilots were not having any better time of it. None of the 15 or more flying survived a day without multiple crashes.  One or 2 Boomers flying at low altitude did continue to do their job- albeit with very frequent re-launches.  

It seems at present that the flat inflated LE's do reduce diving over but not completely- so what other cause (apart from really bad wind) is there?

Could it be a lateral area disposition problem- recalling similar behaviour from prototype ram air pilots 10 or so years ago which was eventually pinned on too much lateral area to the rear, which when the kite is hanging off to one side or the other, prevents the rear of the kite from rotating downward to allow correction. Leong says that his 1Skin (number 42) is not as reliable as Danny's Boomer- but I didn't see either flying.

1 March '16, Pasir Gudang: Re-built the remains of 14/red by removing the LE inflation and reducing the size of all the rear flares (including the outers), adding 6 more bridles to provide support.  Tested in strong winds (70km/hr?) the LE collapsed terminally, even after letting centre "A's" out to original, and even after letting the 4 centre "A's" out a further 20mm and shortening 2 rear centre bridles 20mm had too much leading-edge compression.  Seems a lot more volatile also.  Needs to be tested in moderate winds and tuned so that the LE does not collapse.

Why did the inflated LE's of 13 and 14 allow shortening of "A's"- and why did they seem more stally compared to identically bridled single skin LE versions before this shortening?  Having a more rigid LE explains the first part of this, but NOT the second.

And why does this latest single skin LE 1Skin prototype with more bridles and less rearward lateral area suffer from LE collapse even when bridled to a substantially higher angle of attack?  VERY puzzling!  That it is more volatile unstable is likely a function of less total lateral area.

Malacca, 1, 2 March '16. Strong offshore, surprisingly smooth, except low down.

Tried reducing rear lateral area of 1Skin K25 yellow by adding a bridle to each side flare, taping off the tip (see photo).  In strong gusty wind, this had the same response as 14/red above; leading edge collapse and much more volatility.  Interestingly, the effect seems to be with the shoulders; when they are released a lot (K14/Red, K25 Yellow), not only the shoulders lose inflation, but the centre does also.  This kite did have much quicker recovery- but is this reducing "diving over"?- it seems to head off to either side even more.

Both K13 Pink and K25 Yellow performed much better in original form- K14/Pink best in gusty strong winds close to the ground.

The only improvement from the current standard (K25 Yellow for example) found so far is therefore the double skin leading edge, and it's as yet inconclusive whether a single inflation point or separated sides is best. K13 Red (split sides) was marginally better than K14 Pink (single inflation) in the extreme conditions at PG, but within the range of individual differences and tuning. 

The side flares are currently providing a lot more of the general leading edge "inflation" than would be thought possible from the rib structure, and are also providing a fair amount of drag, which holds the kite back from ascending to a line angle where the angle of attack is too low for leading edge form to be sustained.  Effects of disposition of lateral area are not clear- as yet too mixed up with other effects to establish.

To try next;

Transverse vents at leading edge - has been tried before, but many changes have been made since then; so it will be worthwhile checking this again.

Moving lateral area at the front- which will shift the lateral C of P without reducing total lateral area.

An active bridle?- but optimal strong and light wind bridle settings need to be established first to set the range required.

SS Show kites from 9 Nov '15: 

  After Wuhan 2015 and before Nelson (Jan '16), the 34m Serpent was developed intensively.  First stage; inflation, eventually improved to an acceptable level by moving the gauze inlets downward to roughly split the lower leading edge and adding extra bridles to speed initial inflation.  Second stage was bridling, and the first step in this was to lengthen bridles progressively towards the rear while increasing camber (chordwise and spanwise in the upper head area until the minimum angle of attack that could be sustained was found. The next step was to search for stable flight by changing leading edge and side bridles.  Shortening the side rear bridles has the apparent advantage of creating stabilising keels but also pulled the kite's centre of pressure rearward to an extent that caused inexorable "diving over".  It's very difficult to get the C of P into the upper half of the head let alone close enough to the leading edge to get useful recovery from the sides of the wind window  in any wind.  Shortening the leading-edge bridles in the shoulder, and especially upper shoulder area was the only means found so far that achieves satisfactory flying.  Stability is excellent in mid to strong winds- there doesn't seem to be an upper limit so far, this kite just becomes more stable (in the volatile sense) as wind speed increases without falling into superstability.  Recovery from angular perturbation is excellent; neither too fast nor too slow.  Flying angle is also excellent- at least as high as Boomers and 1Skins, much higher than pilots and other ram air kites.  BUT, this kite still has an annoying inclination to fall off to one side or other in light winds rather than sinking back in the centre. Tried rigging the tail attachment so as to allow the head to pivot more freely but couldn't see any improvement- but this needs to be tried again after everything else has been optimised- if it helps at all in light wind recovery, could be elastic so as to revert to full width attachment in stronger winds.  Minimum wind is only slightly above conventional Rays, pilots etc but its mode of falling as the wind dies is unhelpful.   Nelson flying with this kite was excellent- perhaps only one unscheduled re-launch in 2 days. After Nelson, built a 4.5m x 50m version, one less bridle row spanwise, one more chordwise.  First bridle set up was approximately scaled from the 34, then developed without further reference.  Was found to be quite sensitive to the second bridle row - too short and the head crumpled, kite wouldn't fly. Too long and the collar inflation is poor.  The collar was re-cut to taper from the shoulders for better appearance (but did this make getting the C of P forward more difficult?)  The 50m has a central bridle row (no centre line bridles on the 34m) and single gauze (double on the 34).  Eventual bridling looks quite similar to that on the 34m, perhaps the shoulders a little further forward (which might also reduce the light wind falling off for the 34). Launching is very fast and easy- huge apparent wind on the way up, just a little weaving (was none before LE bridles 4, 5 and 6 were shortened by 50, 100 and 50mm respectively to reduce light wind falling off.  Flew very well in the gusty wind at Pasir Gudang (though neighbours didn't think so as it swept sideways at frequent enough intervals- but conventional pilots weren't handling the wind there any better and it was off the same anchor as a pilot, Ray, Trilobite, Trilobite train.  Malacca (strong winds) flying was also excellent, as was Satun.  Pull is more than a ram air kite of the same head size, but manageable- did break 1500kg(nominal) Dyneema at Malacca and Satun- but evidence of cutting damage both times, and didn't break again after knotting.  The tail on the 34m is in 3 sections with an about 75mm gap longitudinally between each.  The tail on the 50m is overlapped (for a solid visual effect) and works as well, but actually people seemed to like the visible gaps more.  A single piece tail with tensioned outer edges would probably work just as well, is quick to make and would not catch trees, tractors and stuff  as much- safer too. The 50m is a more developed and refined kite now than the 34m- will be the pattern. It's bridling system is easy to fit but does tangle more than the cascade set-up on the 34- split into say 4 or 6 bands chordwise to reduce tangling without adding complexity?  The only remaining problem with this design is their inclination to fall off to one side or the other when the wind is insufficient (currently maybe 10km/hr?) This may be an inherent angle of attack problem which will be difficult to solve without getting leading edge collapse, but even a small forward movement of the C of P position, if this is do-able, will help a lot, and the elastic head pivot idea (above) has theoretical promise at least.  Three sizes; 34m, 50m, 75m?  Before Pasir Gudang, did try taping off the collar inflation- flies nearly as well, slightly more inclination to shoulder collapse (but this was fixed for the octopus below when it was tried as purely single skin by shortening the upper shoulder bridles by another 200 to 250mm).  Without an inflated collar LE appearance is very notchy.

 The 20m Octopus has not responded nearly as easily to inflation and bridling development as the Serpents have; in its original SS adaptive LE form the leading edge tended to fold under in mid to strong winds (top end was around 25km/hr) although the light wind flying was excellent (still failed eventually by falling off to one side though).  Andrew Beattie called it "possibly the most dangerous kite I've ever flown"- the wiss. A procedure that requires care and respect is launching though; in mid and strong winds the speed of launching is phenomenal; far too quick to allow an disentangling from the tentacles- this goes for the Serpent as well. On the other hand, they haven't as yet shown any inclination to looping (because tight turning causes some shoulder or leading edge indentation which either disrupts the loop or causes collapse. When figure eighting, any dip below horizontal can, in light turbulent winds, cause head compression and collapse (Serpents and Octopus).  This maybe be a source of "stability" as well as a cause of crashing (suspect as various incremental improvements are made that looping might become possible). 

The inflated collar added to the blue 20m Octopus before Shenzhen 2015 was independent of the top skin and this flew well enough under a pilot (Shenzhen and Wuhan), was volatile unstable and dived over badly when flown pilotless. After the 34m serpent was eventually developed to fly successfully, a similar collar and bridling was tried on the Octopus, but after extensive development (before and at Nelson 2016), would not fly well alone and had terminal tail tangling. Flew at a noticeably lower angle than the Serpent in this format. After this, mainly at Northpark, linking the tentacles (necessary to prevent tangling, had tried careful tip tapering and Si coating to no avail) and pulling in the upper shoulders a A LOT eventually had it flying , but still with a lot of sashaying- which seemed NOT to increase with windspeed.  It was then tested with the inflation taped off- and flew very much better, to the extent that the inflated collar was completely removed (and then the centre leading edge was cut back about 300mm to reduce leading edge collapse more and  move the C of P forward) , after which it flew the best it ever had- handling even strong turbulent winds on 15m of line beside the kite factory that most (all?) other soft kites would not have been have been able to-  and excellent in the light also. To be considered is whether some of this improvement came from removing the adaptive leading edge?  Mostly though it seems to have been from shortening the upper shoulders and letting out the lower edges (had been through the process of finding the lowest angle of attack it would accept without LE collapse, including letting out rear bridles progressively as far as possible). More and less spanwise camber was tried and some testing of more/less chordwise camber also. But the leading edge notchiness was pretty ugly so for PG, Malacca, Satun, a new flat collar was fitted, after which it went back to sashaying (still fairly independent of wind speed).  The only obvious explanation for this is that the collar improves airflow over the LE, which boosts L/D- hence increases VI.  Backing off the upper shoulder bridles by 50mm or more did reduce the lateral oscillations somewhat, as did shortening all the bridles to half length (more spanwise and chordwise camber).  A possibility is that making the head more free to pivot (recover) by taking the tail load centrally might help- worth trying.  This Octopus flat collar is easier to make and cleaner looking than the Serpent's  system - worth while trying on the Serpents therefore.  If these two changes don't fix the sashaying, easiest next step may be to  replace a good-flying Serpent's tail with Octopus tentacles- to isolate a lot of variables.

An interesting observation, applying to Serpents and Octopus both, is when any significant lateral displacement occurs, causing the kite to be leaning out to the side, the inner shoulder (nearest the true wind direction) is strongly "inflated", while the outer shoulder shows significant indentation.  Clearly this asymmetry is aiding and may be close to being the entire cause of subsequent recovery.  Simon's brief watching of this elicited, "no this is not happening", but after many many hours of watching, I'm sure it is. In very light winds the lift generated by the apparent wind is not quite sufficient to drive the kite up and out of this lateral displacement back towards the centre and the kite then collapses to the edge. This is almost like a soft kite equivalent of "Rokaku stability" (by which when one side of a Rok travels faster, frame geometry causes that side to camber more, slowing it down and damping any build up in VI (volatile instability).  What will improve this dynamic for the semicircular head SS and collared SS single line kites?

Anything that moves the head's C of P forward should (by giving the lift component more leverage against weight and inertia holding it back from correction).should assist.

 Allowing the head to pivot independently of the tail/tentacles also should- especially if this is made apparent wind sensitive- to operate less as wind speed increases (when they don't have a recovery problem). Adding tail weight or reducing tail drag will probably also help in this also.  Would more tail weight- or moving the kites CG rearward- have deleterious effects in stronger winds though?  There are ways to decrease drag as wind speed decreases - (elastic aperture drogues for eg). 

From another perspective, anything that reduces the initial displacement/lean will also address this problem- causing the head to correct any angular displacement before the kite has displaced laterally to any significant extent for example (this is the theory that superstability is just very slow acting volatile instability- that correction doesn't occur until the kite is to the edge of the window).

The self-correcting leading edge dynamic observable with the Octopus and Serpent SS and hybrid SS's is about flow not being parallel to the kite's centre line in this circumstance (no keels/or flares), so could be adversely influenced by using the outer edges towards the rear of the heads as keels to any significant extent (maybe explaining why this didn't work when tried early in this development).  Disposition of lateral area has also to be something to consider (thinking back to noticing that excessive lateral area to the rear of pilot keels causes them to lean off and collapse to one side or the other in light winds and momentary lulls and similar possible effects now with 1Skins and Boomers).  What consequences do considerations in this section have for smooth tail Rays- which have a not dissimilar "falling off" problem in light winds?

Add section from PLHimself for through to before leaving for Chengdu.

Chengdu, Tongren, Nanjing, Weifang, Uiseong, Wuhan, Hancheng

1Skins: No changes until Hancheng, having used them very little while taking every opportunity to use the 20m Blue Octo (ex-Andrew Beattie) as much as possible, increasingly as a pilot.  The latest double leading edge 3sq.m 1Skin (blue/purple) was stolen from our base on the field at Binhai on the 2nd day of Weifang, before it had been flown.  At Hangchen there was a Chinese made version of a Boomer flying exceptionally well in very light winds.  White with Chinese characters on it (see photos).  I tried the orange 6m 1Skin in the same wind and it was nothing like as usable- even with continual pumping it could barely be kept up while the Chinese one was hanging up there tethered off.  At this time no other kites were flying- various 11sq.m Pilots were being tried but couldn't hold, and even Helmut Georgi's Genki was struggling.  The Chinese Boomer copy appeared to be much more cambered towards the leading edge, and it did appear to collapse its leading edge when the wind increased to 15-20km/hr.  I tried pulling in the front 4 bridles of the orange 6sq.m. 1Skin by 60mm, and it did then fly as well as the Boomer copy- would sustain in almost no wind, and was, if anything, at a slightly higher angle. A bridle adjustment was then added to this kite (similar to those fitted to the 20m SS Octo before Weifang and the 50m Serpent after Uiseong). With the shorter LE settings, I don't yet know what the upper wind speed will be but doing this definitely provides a useful option for very light conditions, thinking ahead to Chongqing/Fairy Maiden Mountain.

The SS Octopus development (see below) is showing that a leading edge without any lower section is satisfactory in light winds, flies at a high angle, has a pretty respectable wind range- 9 to 15 on the forward setting, 12 to 50 on the second and 15 to perhaps 80km/hr on the back setting- and has excellent stability and reliability within the ranges. It's clear from watching the 1Skins flying that indentation of the lower leading edge is a cause of leading-edge collapse and luffing- would they be better without this extra fabric?  This also calls into question the value of the "adaptive" leading edge used on the first 4 SS Octos (and taken off the one that Andrew returned before Chengdu.  The idea of the adaptive leading edge on the 1Skins- and earlier SS Octopus's- is to decrease the kites angle of attack at low flying angles so as to improve their light wind flying- the SS's Achilles heel.  This is an excellent theory, and does seem to work in practice, but the Blue 20m Octo is now flying in lighter winds than it did with the adaptive LE (albeit with alternative bridle settings, but maybe even on just the second position which has a pretty good wind range).  Could it be that the lower leading-edge flap is unnecessarily causing LE collapse when the kite flies at lower angles of attack in stronger winds)? Does this need looking at again?  I have a pretty seductive theory that the best possible single skin soft leading edge will be when the LE is exactly parallel to the wind flow (which is rising slightly in anticipation at this point so is not horizontal). Try scalloping (as tried on 4 Skins) again?

SS Octopus and Serpent.

At Nanjing (very light wind) the 20m Blue Octo did fly for a while, for one session as a pilot above the 50m Serpent,, and could be pulled up relatively easily.  Tried pulling in the front 5 bridles on each side by 60mm and this improved light wind flying a lot- and almost completely eliminated "falling off". At Kaixuan before Weifang this bridle adjustment was changed to three sets of 4 leading edge bridles with 3 separate positions for each 60mm apart. The new yellow 34m Serpent was made the same and the 50m Serpent also fitted with an adjustable leading edge- about 75mm steps. At the Weifang event- sometimes rain and very strong winds, the blue Octopus flew pretty well with the 3 LE sets all to the second knot- about 60mm shorter, but did move left to right a bit when the wind was above about 30km/hr, and the LE would collapse causing it to dive or loop at above about 60km/hr.  On the original full-length setting, does tend to fall off in light patches but seems quite capable of flying in winds above 60km/hr.  Does it fly a bit more stably when wet- seems to, as does the 50m Serpent.  41 of 56 bridles were cut at some stage late in the day, and the tips of all the tentacles were cut off- clearly by a Kevlar line. Fixed and re-bridled on the evening of the second day it flew well on the final morning- also still on the second setting- but did have some LE collapse causing looping in the strongest gusts (above 50km/hr?). A majority of the primary bridles were cut again- by Kevlar.  The replacement bridles were made as primaries and secondaries, to reduce tangling potential, and with lighter line so that they will hopefully break rather than cause injury during any launching mishaps.  At Uiseong, Wuhan and Hancheng this kite was really excellent- used as a pilot, very stable and reliable except when on the third bridle setting after wind increases.  On the second setting it was barely showing any LE distortion in 30km/hr estimate.

The 50m Serpent with the LE bridles forward 75mm handled the strong wind and rain at Weifang very well, only collapsing and looping in a squall late on the first afternoon- when it took more than 10 people to get a 9kg Ray and 11sq.m Pilot down.  The tail were ripped off in 2 places, probably on a sign.  At Uiseong, Wuhan and especially Hangchen this kite flew exceptionally well- almost no falling off even in winds of around 10km/hr- is not far off being as useful in light winds as pilot kites.

.The 34m yellow Serpent, at Weifang, would not fly with an identical bridle to that which works so well on the Octopus- very likely because of less tail drag.  Changed after Weifang to having 25/75mm crescent shaped strips sewn across the tail it did seem to be much better when tried briefly in very light winds, but this needs to be tested extensively.  An alternative way to increase tail drag will be to make the tail with 4 or more zig-zagging pieces, hemmed and sewn together so as to leave many longitudinal slits.

The 50m Serpent with the inflated LE collar is such a good flier- and seems to have a bit less pull proportionally than the Octopus, that this design should be retained even if the pure SS smaller version can be developed successfully.  The tail should revert to the 3 piece with a 75mm gap between each, proportionally as per the earlier 34 m Serpent.

The general form of these kites can be extended to Chinese opera mask heads with straggley beard tails and many other designs.

The Octopus, Serpent and 1Skins are all clear of the bog and onto firm higher ground now- but how much height remains to be climbed before the peaks of these particular hills are attained?  

Automatic bridle adjustment;

The fundamental issue is that the leading-edge bridles need to be substantially shortened for good light wind flying but this causes LE collapse in stronger winds, and in turbulent or luffing situations- and prevents recovery after the kite has collapsed.  This is almost certainly unsolvable except by using some sort of sensing system which triggers shortening of at least some rear bridles when the kite's centre of pressure moves too far forward.  But there may be ways to mitigate some of this problem reactively. Splitting the bridles into 2 sets chordwise with 2 lines to the flier would of course enable active control of this- and might be a way to the lightest wind flying that is possible. A radio controlled servo system that can pull in or let out the LE bridles by the desired amount could also work.

A reactive system that should help would be to use elastic elements on some rear bridles- in stronger wind this would have the effect of reducing the kite's lifting area and moving the bridle position of the remaining forward section rearward.  This doesn't seem as likely to cause unrecoverable luffing as elastic elements to the front bridles would.

Another way would be to fit a slow spring to the LE bridles- so that with sustained load they lengthen by the required amount, with lower average pull they gradually shorten up.

A way to make a slow spring would be an elastomeric grease filled tube, sealed at each end, and rigged between the line attachment point and the main bridles.  A constricting ring to which the LE bridles are attached that is also connected to the main line attachment point by an elastic element would enable damped lengthening and shortening of the LE bridles.  This could be inside a braided line.

Ashburton to May 11'16

 35m SS Serpent (Green): At Kaixuan, crescent 25 shaped bars were sewn to the tail to increase drag, but in a first test in moderate wind (Wakanui), it overflew, causing LE collapse- even with the LE bridles let out 50mm further than the 20m SS Octopus standard.  After doubling the bars, moderate wind, Northpark, it dived off relentlessly. With a mid-fish drogue to the tail end, it dived off even more.

Which suggested that the bar drag was now adequate to prevent overflying but too rearward.

After shortening the tail to half-length and attaching the drogue close to the head, it did fly satisfactorily.

This is one of the clearest examples of "pulling the cat's tail' type instability (drag too far behind the kite, acting in a similar manner to when the end of a tail gets snagged on something, causing the kite to dive off to one side or the other inexorably.

I then made a new tail (purple, no green left) in 6 sections widthwise, with about 45 longitudinal slits, each 1.3m long.  This slit section of the tail is about 18m long, with the remaining tapering section about 15m long and plain, single width.  The joint to the head is cut with around 50mm cut back at each side so that most load goes thru the centre section, allowing the head some freedom to correct from angular displacement without having to haul the entire tail around with it.  Other joints in the tail are also cut convex to put tension in the edges so that the tail holds its form widthwise.  

Test fly; light northerly, Northpark; overflies a little, terminal LE collapse, but much closer to being satisfactory.  The slits bulge too much so ties were sewn at their mid points- now twice as many but only 650mm long.  The side whiskers were also replaced at this stage; 6/650mm slits per side, 8m total each. Tested at Wakanui on the 11th, light northerly, still just showing too much centre LE collapse.  Need to find a way to increase drag unobtrusively and near to the head.  Also seems useful to sew the whiskers into the main body to differentiate this design from the hybrid style.  How to add a bit more drag? - some short crosswise slits seems to be the obvious answer- but how to do this without adding complexity and cost?

Chonqing ( 12 to 17May '16):

The 20m SS Octopus (blue) was an excellent pilot 150m above Ray, Trilobite, Whale train in strongish wind and rain/mist on the first day.  The key characteristic in this place (a fairly flat area roughly a km in diameter and 50 to 100 below the surrounding ridges) was to get high and have enough lift to pull up completely wet maxi kites until they could get into some wind.  The Octopus was on the second bridle position from standard and was just verging on too much LE collapse at times.  When the wind dropped, re-launch was successful twice but then became impossible because of the wet tails- at least another 20kg drag.

The 50m green Hybrid Serpent flew well on a much shorter line that the 20m Octopus on this day- it did come down a few times but less often, proportionally, than I would have expected given that it was not above the ridge turbulence.  I didn't fly this kite on days 2 and 3 because its lines (70m/2000kg, 40m 6000kg Dyneema) were stolen before packing on the afternoon of the first day.  This is an excellent design, needs to be retained even when the "pure" SS Serpents are fully developed, because it has a bit less pull for size and is slightly more resistant to sudden collapse in extreme turbulence.

I wasn't able to launch the 20m SS Octopus (still dripping wet) on the second day (and still raining) because of light wind at ground level and extra tail weight/drag.

But a 6sq.m 1Skin was launchable (even when wringing wet), by running hard and this had enough lift once up (40kg +) to get a wet maxi Ray and Trilobite up, and keep them up, even when they lost inflation (frequently).  No other kites, except Revs occasionally, flew successfully on this day. The 6sq.m 1Skin was on the second bridle setting and was still really stally on launch.  It also sashayed from left to right in what looked to be volatile instability when the wind was in the lighter range during lulls, which it has never done before.  Is this because it was wet? - a response to the higher ( 4 x's or more) pendulum effect and high angle of attack. It's amazing that there were so many spectators given that it was raining, cold and more often than not the lower kites were lost in the mist- the top kite always was.

Fairy Mountain (flying area at 1800m altitude) third day; middle wind, no rain: 

The 20m blue SS Octo at 150m above the Ray held the Ray/Trilobite/Whale train up all day.  No re-launches after the first, which took an uninflated maxi UKS Fish up on the line. The Octo moved a little more laterally than most of the adjacent pilots, but in a narrow range, and it didn't get into any tangles that required attention.  The Octopus's "blink" stability (the way that the underside leading edge collapses slightly causing the kite to re-centralise immediately a dive to either side begins) is unique to my observation, though similar in effect to "Rokaku stability (by which the camber of the faster side of a Rokaku increases because of the way their structure is formatted, slowing that side down and restoring the kite to straight flying) . Most, if not all the conventional pilots flying (7 or 8?) did not survive the 3rd day without re-launches- basically because the mountain winds were quite turbulent, and when luffed or tipped they take a lot of height to recover.  It was also noticeable that the show kites below the Octopus fared much better than those below conventional pilots, tending to loop and collapse more often because their lifters did not have as much pull nor fly at as high an angle.

Generally, therefore, this event showed, again, that the SS's can be better pilots than current ram air designs and have opened a window for flying that has been closed until now- on the second day in the rain, no single line show kites did fly or could have been flown except under the 6sq.m 1Skin.  As Craig said; they have turned our previously best lifting show kites (PL Rays and Andreas's Mantas) into line laundry. 

Using a 4.5m 4Skin for buggying at Chongqing was also very successful- plenty of pull to go uphill and thru swampy bits and good control/stability/anti luffing downhill and when speed had to be scrubbed off. This design does not need much improvement (a reduction in the dreaded reverse flying tendency and a bit less sensitivity to brake tension) to be a useful buggyng kites. For one session, the brakes were tangled thru the front bridles a bit- courtesy of locals who would try to launch and fly every kite left sitting around- and this was enough to completely prevent recovery from reverse flying (I didn't see what was causing this problem until later).  Too sensitive!

Why do most (not NASA's or 4Skins) single skin traction kites, Boomers and 1Skins have a lower leading edge?  The stagnation point (highest available pressure) is close to the centre of the leading edge, so any extension of the lower leading edge will reduce the available underside pressure significantly- which, top skin flow effects notwithstanding, must ALWAYS result in less secure leading edge form unless stiffeners of some type are in use. Is the usefulness of a lower leading edge merely to provide drag at lower angles of attack to prevent overflying/luffing? Would providing a reflexive trailing edge - so that at angles of attack approaching zero, the TE is pushed down- do the same job with less undesirable consequences (LE collapse).  A reflexive TE would be angle of attack sensitive, whereas the current palliatives (lower leading edge for the 1Skin and hybrid Serpent, tail for the SS Octo) are not, which means they will always have more than necessary drag in lighter conditions but not enough above some windspeed. The 4Skins use rapid flier response to stop overflying/luffing, but this is not available for 1- or 2-line kites.

Lower LE's do also seem to exacerbate LE collapse once it initiates- which seems likely to be true because the remaining still formed LE will be very marginal by then- easily collapsed by even the tiniest extra force.  This could also explain why the "adaptive LE" used on the 1Skins and early model SS Octopus's is not as effective as expected- yes it does reduce stalling at low wind speeds and high angle of attack, but it also accelerates LE collapse once this initiates. 

Using extra long bridles to the LE should also help push LE collapse up the wind range- and these bridles should be as small diameter as possible.

What is the "blink" stability mechanism?  Can it be applied to 1Skins?

"Blink" stability appears to be a function of having the side leading edge bridles shortened so that one side collapses (folds in) just before the centre does.  The effect of this is that when the kite begins to fall off to either side when the wind is strongish, the leading edge on the side it's moving towards collapses a little, which slows down or halts this movement (because of increased drag?). At this stage the flow is then asymmetric- because the kite is leaning but no longer traversing sideways.  The opposite side leading edge- that isn't collapsed- then generates more lift than the side that is a bit, causing the kite to straighten up and centralise. 

If this is a correct explanation, then to apply it to 1Skins will require that the side leading edges are bridled forward relative to the centre and that the keels don't prevent the required transverse flow.  I have, until now, thought that the "falling-off" and "diving over" that 1Skins can be plagued by in stiffer winds, has been caused by one side LE collapsing in such a way as to pull the kite inexorably towards that side.  The Octopus "blinking" stability theory above suggests the opposite- but I wonder if this apparent conflict is a function of rotational velocity?  Could it be that the 1Skins, lacking a tail, get into a curving flight path of tight enough radius that the increased lift from the faster side overcomes their restorative mechanism (the weight pendulum).  Nah, can't be true; 1Skins show some form of superstability, not volatile instability (of which the above is a classic description). And, it's their inner LE that stays collapsed during this- not the outer, faster, side

It's worth trying bridling the shoulders of a 1Skin forward to see if a similar "blinking" mechanism can be contrived.

It's also time to again try, (rather more determinedly this time) moving the centre of area of the 1Skin's keels forward, just to check that their "falling off' isn't just a lateral area problem, like it is, incipiently, for the 11sq.m Airbanner Pilots? (I need to fix this too, even though it's only an issue in very light winds).   

To 25 June '16, including Kelantan, 1-6 June '16

1Skins:  Built 1Skin 64 pink 28th May'16 with keel area moved forward to check, again, whether diving over is a function of lateral area distribution- or to what extent it is. Centre of pressure for this kite is on a line between the second bridles on the centre ribs. This kite is exceptionally smooth and its trailing edge tramps noticeably less than earlier 1Skins, probably because the rear bridles are closer to the TE. From the side while flying, it appears that the lateral area centre is significantly FORWARD of the lift point C of P. Have I moved the keel area too far forward now- will it be superstable for this reason (lateral area distribution either too far rearward or too forward can cause this)?  Except, looking at it again, this side view was not accounting for outer keels, which are more rearward now (though smaller).  Still does look to be too forward if anything.  Flying at Wakanui and later, extensively at Kelantan, no sign of diving over, but wind was never much above say 30km/hr.

Could diving over be because when the TE flaps, there is effectively no lift from this area? Correct?  If so, this will be causing the kite's c of p (for the remaining lifting area) to move rearward relative to the bridle point causing superstability (exactly what diving over appears to be)- and also having the useful effect of limiting leading edge collapse by preventing the kite's angle of attack from reducing too much. 

Even though the lateral area disposition theory (above) hasn't yet been fully tested, this is a 'new' (didn't I consider this in earlier iterations- think I did but mainly from the LE stability point of view) theory, is worth checking out. 

Built 1Skin 65 red, 23 June with all rear keel TE's steepened by about 100mm, full length shoulder bridles and outer keel smaller than pre 64 1Skins.  Made completely new and refined templates for this kite also- in preparation for Melbourne workshop in August.  Test flight in light wind, Wakanui 23June, flies very well, smooth, slightly volatile in minimum winds, need to be tested in stronger conditions before sending to Simon in Melbourne as a pattern.

At Kelantan the 6sq.m orange 1Skin flew exceptionally well in very light winds with the 4 centre LE bridles pulled in 50mm.  It's VERY near a minimum wind match for the best light wind ram air pilots now, and as soon as the wind is sufficient to sustain it- <8km/hr? it's superior because it has so much lift.  With the front bridles shortened this much, it wasn't distressed (no destructive LE collapse) up to 30km/hr at least. This kite has had some other bridle changes, so there are photo references in "Kite photos SSSLs" for when the next one is made- unless it's superseded by the 65 series.        

20m Octopus SS

20m Blue flew exceptionally well at Kelantan, only just started to show LE collapse at 30 to 35km/hr with centre front bridle pair 2 knots forward, others 1 knot forward.  Excellent angle, great pull, no stability issues, no leaning issues.  Only one job remaining with this design; check that the production version replicas are as good.

50m Hybrid Serpent

Using the 6sq.m orange 1Skin as a pilot on this kite at Kelantan was an excellent choice because the 1Skin would keep the Serpent up in light winds with the Serpent bridle on the normal rather than forward setting.  Without the pilot, on this setting the Serpent is inclined to fall off to one or the other side when the wind drops, which is annoying for adjacent fliers.  The Serpent on its light wind setting flown alone needs to be pulled down and re-set for winds above about 20km/hr.  Shifting the top kite attachment point forward to above the LE bridles definitely made for more reliable flying -less wrapping over failures. 

30m Serpent green

The purple tail with 78 diagonal slots, all in the first 6m , slots pulling in not out, did not quite have enough drag, but is a lot better than the horizontal bar tail (green) or the longitudinally split purple tail tried earlier.

It did fly for many hours at Kelantan in this mode but was increasingly volatile unstable with increasing wind speed- and this was with bridling fairly similar to the 'standard' 20m Octo bridle set, though with various minor differences (too many to specify in writing).  Did try adding a midi drogue- attached close to the head was best- but this still caused unflyable superstability.  Volatile instability (above) tends to suggest that allowing the head some degree of rotational freedom relative to the tail is unnecessary.  Clearly it needs a little more tail drag- but not too much.  Going to 4 slots width wise (from 6), same number lengthwise, will increase drag by 50%.  A better alternative (easier to make, better visually, more easily tuneable) could be a series of diagonal scoops sewn to the tails underside edges near the head- tapering off by spacing and size by say 8m along - see design diary page 2226 for sketch.

1Skins to July 20 '16

K65 Red flew adequately in very strong southerly wind on 12th July, at kite paddock.  Is it better than earlier versions though?   Rear centre rib could be, and was, let out 25mm without causing leading edge collapse. B and C centre rib bridles were then let out to match, but some horizontal wrinkles still evident. A very long 75mm wide tail did make it somewhat more reliable; didn't seem to increase edging tendencies, which is interesting evidence.  Built K66 Orange/pink from K65 patterns, refined.  On 14th July this flew quite well in extremely strong gusty Norwest- stayed up for some time then took Craig, Gavin and I to pull down, maybe 30kg pull.  Differences to K65 are minor- just flares changed to return all bridles to the same length again.  At Wakanui, later on the 14th, flew very well, strong South long-shore wind, reducing to light after an hour.  Tip flares were let out 25mm without causing any noticeable shoulder LE change- shoulders are still very firm, some centre LE collapse in big gusts but only on the lower leading edge, never progressing up to the top skin.  Some horizontal wrinkles, addressed by letting out centre rib A and B's and outer rib B by around 25mm. I'm still suspicious that the 65/66 design is inclined to going out to either side for a look sometimes for no good reason (that is, not because of wind changes), but they have much better recovery than earlier series, are quite volatile.  In light winds they will figure eight and can jiggle from side to side.  In strong winds they will sometimes loop, which I don't recall earlier 1Skins ever doing regularly.  Maybe extra volatility is because they have a bit less lateral area?

K65 Red and a set of orange cut-outs for another K66 sent to Simon Freidin in Melbourne on the 14th.   Without extensive comparative flying against earlier 1Skins, it's not possible to be sure that 66 is an improvement (wrt diving over)- but so far it does seem so.

What is "diving over"?  It's the strong inclination this style of kite -tailless single skin single line (SSSL) kites- have in stronger winds for heading off to one side or the other until they get into lower apparent wind speed, (turbulence or ground effect) or hit something.  Although this may have different causes; see below, it is fundamentally a form of superstability (SS) in which kites are too slow in recovering from some angular perturbation.  From here in this narrative, 'diving over' will be called superstability (SS) with the understanding that SSSLs may have a unique form of SS. 

Some recent observations:

The 6sq.m Orange 1Skin with the front centre 4 bridles shortened 30mm for light wind flying performed very well at Kelantan in winds up to 40km/hr; huge pull, no SS, some LE collapse, but no apparent adverse effects of this.  At Chongqing this kite (standard bridle setting) successfully lifted a wet maxi Ray and Trilobite- at least 50kg average pull- but was probably SS at times (couldn't see because we were in cloud) in that the line occasionally angled out to the right. 1Skin 64 (lateral area disposed forward) also performed well at Kelantan- no sign of SS beyond requirement to tune straight.   At Wakanui, 14th July, 1Skin 66 was either SS or the wind was erratic in direction (remnants of a Norwester but acting longshore from the south).

Letting the outer rear bridles, centre rib A's and B's and out rib B's  by 25mm did seem to keep it from leaning off as much, and at Kite paddock, 15th July, on these bridles settings and with the front centre 4 shortened 20mm, performed well in mid easterly- slight right tendency, but good recovery. At Wakanui in stronger wind it was noticeable that the ribs all continued to flare out symmetrically even when the kite was leaning- which suggests that flow is not asymmetrical during SS periods.  It's also perhaps true that this kite is straight and doesn't usually show any SS until it gets to high angle, low angle of attack- suggesting attached flow could be a cause.

POSSIBLE CAUSES OF SUPERSTABILITY IN SINGLE SKIN SINGLE LINE TAILLESS KITES.

1. That the upward pointing moment of the kite's weight acting below its centre of lift is insufficient to overcome aerodynamic forces in stronger winds.

2. That the leading-edge indents on one side or the other as apparent wind speed increases, causing the kite to dive off to that side, and does not "pop out' again until substantially lower apparent wind speed conditions are encountered.

3. That flow progressively attaches chordwise towards the trailing edge over the upper surface as angle of attack decreases/apparent wind speed increases, causing the kite's centre of pressure to move too far rearward relative to its centre of gravity to enable the upward pointing moment above to operate effectively.

4. That lateral area is disposed too far forward or too rearward, causing either the kites trailing edge (lateral area too rearward) to lift or its leading edge to dive (lateral area too forward) when the kite develops a lean to either side, making it unable to recover when a lean induced move off to either side initiates.

5. That TE convexity (in plan view) causes the rear area to become progressively inoperative (by flapping) as the angle of attack decreases in stronger winds) which effectively moves the bridle point for the area that is still contributing lift (the forward section) rearward, stalling it into superstability (SS).    

6.  That the included angle between the bridling in a chordwise sense causes compression of the skin (seen as span-wise wrinkles) as wind speed increases, and that when this occurs in the rear part of the kite the effect is as for 5 above.

7. That the cause is bridle stretch; one side of the kite randomly develops more pull, which stretches the bridles on that side by 25mm or so, which causes the kite to lean to the opposite side, which causes the stretched side to generate even more pull, in a positive feedback loop.  

  Comments on these theories:      

1. A 3sq.m 1Skin weighs 300gm and pulls more than 30kg in strong winds.  300gms acting with minimal leverage (the centre of pressure for single skin kites is necessarily further rearward than for ram air inflated kites, much closer to its centre of gravity) is all there is to correct any lean and point the kite back up.  At some higher wind speed, it seems likely therefore that this correcting moment won't be able to act quickly enough to correct a lean initiated move to one side driven by aerodynamic forces that are already 100 times the weight force by just 90km/hr.  It's noticeable that recovery is slower for tailless single skin single line kites than it is for ram air inflated equivalents - and much slower than for framed single line kites that can be structured so as to put their centre of gravity relatively closer to the trailing edge.  A prediction of this theory is that kites that are bridled more to the rear (placing their centre of lift closer to their centre of gravity) will recover more slowly from any angular displacement, and correcting for leading edge indentation effects, this is clearly true for SSSLs. Conversely, "bridling forward" is generally a cure for SS in ram air and framed kites and does seem to reduce SS for SSSLs to the extent it's possible without losing leading edge form. 

That tailless SSSL SS is at least to some extent a function of insufficient correcting moment is also supported by it's not being a problem for SSSLs with substantial tails (the 20m SSSL Octopus for example) except in very light winds when their bridle point is set to "standard" rather than 'forward", which almost certainly has unrelated causes.  Presumably the Octopus tentacles, comprising 75% of the kite's total weight, increase the restorative moment sufficiently to eliminate this type of SS.

What effect will scaling have on this cause of SS for SSSLs?  Will a bigger 1Skin be even slower to recover and therefore suffer worse from diving over?  Experience with the orange 6sq.m (which Volker Hoberg now has) at events this season doesn't suggest this. Ram air kites become more superstable (recovery so slow that they traverse to the edge of the wind before coming back up) as they are scaled bigger because their enclosed air mass increases as the cube of dimension and adds disproportionally to their rotational moment of inertia. Single skin kites are not subject to this (except entrained air?) so maybe it does make sense that their SS tendency will be independent of size?  Need to check this by dimensional analysis.

2. Leading edge indentation definitely causes SS in SSSLs but that it's not the only cause is clear from two observations: 

That 1Skin series with a ram air inflated leading-edge sleeve seemed to be less subject to SS but not completely exempt.

And that 1Skins can be subject to SS even when there is no observable leading-edge indentation.  

 3. This is a good theory- rollover type failure like the Dutch parafoil at Singapore in 1984. It's supported by the observation that 1Skin SS most usually occurs after the kite reaches a high flying angle (low angle of attack). Progressive attachment of flow over the top surface as the kite's angle of attack reduces would seem likely to be able to move the centre of lift back towards the kite's centre of gravity to an extent sufficient to reduce the effectiveness of the C of G's moment in pointing the kite upwards.  However, reducing angle of attack will also move the underside centre of pressure forward, so the net effect may not be as significant.  A test for this will be to fit a 1Skin with a turbulator to prevent full chord attachment over the top surface.

4. There is no doubt that having the lateral centre of pressure significantly forward of the centre of lift will cause SS (what I have called 'nose area instability' in earlier discussions). This causes a form of SS in all winds, including very light.  Having the lateral centre of pressure too rearward will probably also cause SS in SSSLs, but if this manifests as it does for ram air pilot kites, will be a light wind problem only. From 1Skin 64 (lateral area disposed much more forward) it seems that there the current lateral area disposition is not too far forward, and from watching the way that 1Skin ribs bulge to either side, it's not too far rearward either (even when leaning severely, 1Skin 66 shows symmetrical ballooning of the rear sections of the ribs.

5 and 6. Considered together because chordwise compression is probably the main (only?) mechanism by which the rear section of 1Skins cease to contribute lift as angle of attack decreases (when wind speed increases) so causing the more forward section to become "back bridled", which can initiate or at least exacerbate SS. Average bridle length has, unfortunately, two contradictory effects on this: shortening increases camber but increases included angle, and vice versa.  Check if lengthening bridles, while retaining original camber reduces SS.

The difficulty that addressing this presents is that single skin kites must have substantial chordwise camber to prevent chordwise compression , but chordwise camber pulls the kite's centre of lift rearward which is itself a prime cause of SS- see 1 above.   An earlier theory /expectation was that progressive lift reduction of 1Skin rear areas as wind speed increases acted to prevent leading edge collapse (by effectively bridling the kite back as wind speed increases)- which was the main problem for the first 50 or so prototypes.  Maybe this approach is useful but just needs to be backed off a bit.  1Skins 65 and 66 (better supported trailing edges) with much reduced trailing edge flapping and reduced spanwise compression wrinkles are if anything less inclined to SS than earlier convex trailing edge 1Skins.  They also don't appear, as yet, to be subject to any more leading-edge collapse than earlier versions, while retaining as least as good light wind performance. Why their leading edges don't collapse in stronger winds is evidence that theories to date are wrong, at least in part.  1Skin 66 currently has around 25mm longer rib centre bridles (centre rib's B and C, outer rib's B) than a 2.5m bridle radius would define- that is it is currently "over-cambered ' by around

25mm- but has less chordwise wrinkles and appears to be less (or at least no more) subject to SS than earlier flatter versions.

7. Bridle stretch induced superstability; Ram air pilot kites have sometimes exhibited exactly this behaviour, and the conclusive solution has been to fit less stretchy bridling. 

Is it happening to 1Skins?

Because the SSSLs have many more bridles, it is less likely that asymmetric stretch will be causing SS. A test of this is whether 67 (which is the first of this series to have has polyester rather than Dyneema bridling) is more SS than an equivalent kite with Dyneema bridles- which can be checked by using Dyneema for 68.

Test of 67 black, 23 July '16, Kite factory field and Northpark, VERY strong wind

At the factory site, 67 broke nominal 750kg line (in good condition too), which has never happened before (this line has been used on the 20m Octopus at numerous events, including Fairy Maiden Mountain when the wind was strong and gusty).  Did fly for brief periods in this wind, appears to be less SS than earlier versions.   Turbulator then fitted (see photos).  Tested again, probably no improvement, or at least no dramatic improvement- but needs to be tested in wind that is just strong, not completely stupid.

Before West Java (25 Aug to 1 Sept '16 to after Dieppe (8 to 19 Sept '16). 

30m Serpent:

Very worthwhile developing because it has less tail snagging than the Octopus series- which are a nightmare because of the linked tails.

Yellow appeared to be flyable at Northpark, 22 Aug,, wind NW, fading.  Seemed to need more tail drag, so added another 4 (?) central scoops to the 10 diagonal.  Then seemed to have too much tail drag (falling over to side in light), but also more LE collapse than 20m Octo blue.  Why?  - has to be in bridling and shape of head. 

Tried sewing off 200mm crescent from top of head by theory that its pointy shape (compared to Octopus?) was a cause of LE collapse. This reduced bridles by 4 and did appear to reduce LE collapse, so I cut this extra fabric off.  Now quite ugly head shape- 'ears' at shoulders. Sewed off all scoops, tested again at Northpark, seemed much better but was volatile unstable from mid-range up- test further at Dieppe, re-engage by progressively unpicking scoops until just enough.  

At Dieppe, still no useful steady mid-range winds, still inconclusive. Refined LE bridle set a lot, slightly different to 20m Octo now, possibly better (more progressive).  Does fly very well under 1Skin now- and in turbulent winds (predominantly offshore gusty) it was effective to let LE bridles off a further 80mm from 'standard'. Head looks quite good, curvature of top is not ugly like at Northpark last week. Opened one set of tail scoops.

An interesting observation at Dieppe; this kite now had excessive chordwise collapse between each bridle set, really scooped and a lot of chordwise compression.  Why?  Shortening rear bridle set back to same as 'standard' (that is, all the same except for LE) made a dramatic difference.  This implies that the problem was caused by shortening the head, and that the length of the head (relative to its width?) is critical below a certain minimum.  How strange, consequences for 1Skins, traction kites, Octopus?

Volker Hoberg suggested (Ashburton 19th Sept) that a difference between the Octo and Serpent is that the Octo has slits between the tentacles to release pressure just below the head.  This is definitely worth trying.

20m Octopus:

20m blue flew very well at West Java, most reliable fly-alone single line show kite yet (much better than Rays)- pretty close to 11sq.m Pilots, but has a quite narrow range for each LE bridle setting- too short causes collapse and crashing, too long causes falling off to one side or the other. Currently 4 settings in 75mm steps, shortest setting only for winds of 8km/hr to 15km/hr?! On longer settings, wind ranges are sufficient for wind that isn't gusty and turbulent.  Need to develop either an automatic reactive bridle system or some shaping or other features to extend wind range without having to pull down and re-set bridles. 

20m pink, new (4th Sept, Northpark) does fly but doesn't seem as reliable as 20m blue (or just bad wind?).

Why?

Bridling is the same within mms- amazingly similar- and 20m blue has had 3 replacement bridle sets (Weifang damage) while flying as well in each iteration.  

 LE of Pink is perhaps 50mm shorter around curve in total than Blue (5 cells each side of centre), and is more even.

Stretch?  Maybe Blue has stretched into more 'pocketing' in the cells.   3rd cell adjacent centre measures 36, 42, 37, 37 (38 average) for blue, 56.5 average diagonal.  Pink is 39, 38, 36, 43, (38.75 average), 54, 57 on diagonals (55.5 average). Blue difference (measure of pocketing) is 18.5, 16.75 for pink.  Not a lot- significant?

Porosity?

Check weight differences.

Blue's tails are somewhat damaged now, will have more drag- significant?

Taking centre top pair as reference, blue LE bridle differences are: 0, -160, -140, -170, -110, -50, +80.  Pink is same within +/- 20mm.

Pink never arrived at Dieppe- lost bag, but Blue flew reasonably well in a range of conditions.  Appeared to be volatile unstable in stronger winds, and did crash after some hours flying because of both too much and not enough wind on last Sunday, 2nd setting, falling onshore wind.  With 1Skin 71 blue above and on standard setting it was then painted to the sky for the remainder of the afternoon.  Need to fit thru-bridle to Pink?

1Skins

General observations:

Dirk's Boomers flew well in gusty strong conditions, and in the light. In strong winds, 67 black with turbulator (does it do anything useful? - maybe!) was noticeably better than they were when flown with 2 x 40m x 75mm tails. 67 black and 71 blue with their leading edges pulled in 80mm also hung up better than the Boomers in very light wind (only a Michel Gressier kite was still up- the 11sq.m Pilot would not hold).  But the Boomers use a single bridle setting and no tails- clearly the flat aft TE of the Boomers (and earlier 1Skins) is a valid wind range extender, flapping notwithstanding.

The 2 x 40m x 75mm tails pull the TE of 67 up in stronger winds- and towards their end they are even more horizontal. These tails do seem to be effective in moving the kite's C of G rearward, improving recovery, reducing V instability, reducing SS, without deleterious effect except catching on things.

In very strong gusty offshore wind and heavy rain, 67 black became volatile unstable- even with tail.  71 blue (with tail) was stable in the same conditions. 

Also, 67 black was noticeably more volatile than 71 blue during launching.

Why?

71 does have its front rib flares cut to align with fabric filaments (2 piece ribs) which required about 25mm longer bridles for LE stability.

71 also didn't appear to take up as much water- was noticeably lighter- so probably this is just a classic case of pendulum caused volatile instability.

Does 67's turbulator have an effect here?- I'd expect it to be opposite except for the weight difference.

Could 67's use of polyester bridles be a factor? - 71 has Dyneema bridling. 

Though tending towards volatility, at Dieppe, Simon F's 0.75 sq.m. 1Skin flew well in stronger winds- looped easily without destructive LE collapse in winds that was causing the 3sq.m versions to show SS causing LE indentation. 1.5sq.m yellow was in lost bag, as was hybrid pink they would have been a useful comparative tests with this.

Also, 67 black appeared to be more resistant to LE indentation in strong winds than 71 blue (lighter, softer fabric), same bridling, same patterns (except for extra fabric at shoulder bridles on 71 to reduce creasing there).

Also, 72 grey (very soft stretchy Si coated 40denier ex OL Whale) barely flies at all in any wind- really bad LE distortion.

Suggests strongly that fabric stiffness matters:

To improve LE stiffness (or at least make it hold form better in stronger winds and at lower A of A);

1. Very small extensions of the 4 centre and first shoulder bridles can have a huge effect on SS in stronger winds.  First tuning step it therefore to adjust these bridles until the LE does not indent- or at the least, indents equally. 

2. A way to embed fine monofilament nylon in the leading-edge fabric? - the idea being; something that makes this fabric stiffer while remaining packable and non-creasing. 

3. Use polypropylene fabric (like used for bags and tarpaulins) for the leading-edge piece and the top skin adjacent to the LE.

4. Use a laminate of fabric and chip bag plastic- this stuff is highly resistant to crushing, creasing, as everyone who has tried to tidy away the peanut bags given out on aeroplanes knows.

5. Shape the leading edge to the form it will take in strong winds at low A of A, not the semi-circular shape it takes in light winds.  The strong wind shape is almost pointed at the front edge- so this should be a seam with stiffer or stiffened fabric to both sides.

6. Try bridling the LE (again)- will still need to be shaped to allow upward curvature between bridles.

7. Generally, the more accurately the LE fabric can be shaped to reduce creasing the better- especially around the rib noses.    

Reactive bridle approaches (also for Octo and Serpent);  

1. Pair front 4 LE bridles with rear 4 (or 6?) on a spring from main line attachment point.  2/1 pulley on front set so that they extend 60mm for every 30mm of rear set let-out?

2. Combine the front 4 bridles and extend to a point say 1m below main line attachment, have a sliding ring pulling this line and the line from all the remaining bridles together, spring loaded with a light spring to pull it towards the kite end and a heavier spring further down the kite line to pull this ring down, letting the LE bridles out as wind speed increases.  Too much ring movement required for 80mm LE pull in?  But does have some natural damping.

3. A, say, 3/1 lever with the 4 front bridles to one end, the main line to the other, a spring from the main flying line to the 1/3rd point on the lever to which all the other bridles are also attached.  This is easiest to make and adjust (spring tension and leverage ratio0 but will cause bridle tangles, especially during over flying and pumping.  This is the best approach for testing the reactive bridle principle.

4.  Use a spindle with different diameter grooves taking, to the larger, the front bridle set, to the smaller, all the remaining bridles and with another groove for a stretchy to the main line, which also attaches to the spindle's axis.  This would probably be less inclined to tangle than 3 above, is equally tune-able.

5. A light "sticky' spring in the front bridle set- which extends promptly when they have a lot of tension on them (strong wind) but shorten slowly when the tension on the front bridles decreases- as discussed earlier.  

6.  Use a spring between the main line and the 10 centre rear bridles (with a self-adjusting ring pulleys between inner and outer rib sets to eliminate stepping).  Then take a line from the tip and shoulder bridles through a pulley attached to the spring upper end to the front bridle set.  This has the potential to be least subject to tangling.

All these reactive bridle systems suffer from not being angle of attack sensitive- at high A of A's the leading-edge bridles should be shortened to reduce stalling but these systems, only reacting to total pull, will not do this.  A consequence might be that stalling becomes worse on stronger winds.  What would be better is a system that is A of A controlled - but the only reactive way to do this appears to be by exploiting that B bridle tension will increase relative to D as A of A decreases- or a powered sensing/servo system.  Best would be a system that senses LE collapse and responds by letting out these bridles or pulling in on the rear set. 

Shape changes to reduce 1Skin SS: 

From Northpark and Wakanui before West Java, it was noticeable that asymmetric creases in the skin can also cause SS- even when the LE's hold form.  At a given moment with 67 black hanging right, there was a visible crease in the right side.  Moments later- after a jerk or 2, a crease would appear in the left side instead, and the kite would then hang left.

This suggests strongly that compression of the skin also needs to be eliminated, in the more forward sections especially.  More chordwise camber accomplishes this- but also moves the C of P disadvantageously rearward. Does more spanwise camber also help? - maybe allowing chordwise camber to be reduced?.   This could be done only in the rear section- perhaps by inserting wedges of fabric into each cell (or go to individual cell panels?)  Another answer would be Octopus/Serpent style cells/bridling- but how to do this without losing keel area? Cut repeating curved sections into each rib?

Reflexive TE? - could be supported by cords or by short top surface ribs. This moves C of P rearward though which is not desirable.

"Blink" stability as per Octopus appears to be a substantial solution to SS (but would it be for 1Skins without tentacles to move the C of G rearward/), could probably be contrived for 1Skins by using a very open net (fish net) strip, say 100mm wide , along every rib against skin.

Would making the rear 25% say of the main skins gauze reduce the pernicious effect of rearward camber without causing compressive creases?  Would also act something like a turbulator. 

1Skin variant designs:

Use Octo/Serpent style LE, cells and bridling with side keels only, 4 flares each side.  Aspect ratio between 0.5 and 1.0.  This eliminates the rearward C of P that derives from having to have enough chordwise camber to prevent chordwise compressive creases. Would also be simpler to make.

Another possibility is to use less bridles spanwise except at LE, maybe a triple sled or double.  I did try a simple sled at the start of this program, might now have enough understanding to make it work.  I expect problem will always be the LE and how to shape and support it well enough without it indenting at the shoulders.,

For next prototypes, seems indicated to try a 1Skin with polypropylene leading edge and skin front strip, one with a gauze rear skin section and one with fishnet rib edges. 

LE factors:

More bridles, longer front bridles, different placement of front bridles, better fabric shaping, shorter lower LE, stiffer fabric, shaping for strong wind not light, stiffeners, ram air sleeves, higher A of A, different rib nose radius, top skin camber near LE, some active symmetrifier, plan form LE curvature, chord length relative to cell width (as noticed for Serpent), TE pull down, 'biplane' LE (like Peel prototype).   

Tried preferentially collapsing the centre LE on 67 black- principle being to lengthen the shoulder bridles while shortening the centre LE bridles so as to have a net decrease in bridled A of A.  The centre notched in while the shoulders stayed tight, causing the kite's planform shape to narrow markedly at the front (moving the C of P rearward) while also keeping the shoulders resistant to SS causing indentations.  Results inconclusive- not obviously a clear improvement, but worth following up.

After Dieppe (Sept '16) to February '17.

This development project is becoming something of a saga, hopefully not of the type when the subject remains lost in the swamp forever! 

Single skin pilots: Up to end '16, I concentrated on improving the LE.  Many prototypes: starting with a shaped polystyrene model, stuck on pieces of fabric.  The challenge is to estimate (guess) the spanwise camber of each cell at the leading edge, which is really difficult to do as this is also very sensitive to small changes in chordwise bridling.  Most single skin traction kites deal to this by having close spaced ribs so that this spanwise camber and camber changes can effectively be ignored.  4Skins, and 1Skins have significant cell camber at the leading edge and the only path to clean shaping seems to be make, bridle, fly, modify, --- in repeating iterations.  By 1Skin 79 (grey/blue) India (12 to 19 January) their LE's were very clean- and as part of this process, the lower LE had been progressively reduced, by the theory that indenting of the lower LE in stronger winds (lower angle of attack), was causing the upper LE to also indent, promoting superstability.  79 and successors show no indentation at all up to perhaps 100km/hr, their LEs hold shape.  But they are still subject to superstability- and after going back and re-testing earlier models right back to the 30 series during January '17- it became clear that the onset windspeed that this occurs at has not changed substantially  (around 40km/hr), even with LE improvements.  Maybe later kites are more reliable and generally have a higher wind capability- the earlier models re-tested were all selected as being the "best of" for their time-  but without doubt some early versions can fly reliably in winds as strong as any of the latest, up to 79.  This suggests that there is a cause of superstability other than LE indentation or chordwise skin compression.  For now, the most likely candidate is insufficient pendulum effect from the kite's weight (300gm) acting below its centre of lift, to overcome the large aerodynamic forces (30kg +) generated in strong winds. The obvious cure (apart from adding a tail, which shifts the centre of gravity rearward) is to move the kite's centre of pressure forward without also shifting its centre of gravity, but this is prevented by the necessity for having significant chordwise camber over the entire chord so as to prevent spanwise compressive creases in the skin (which also cause superstability).

Light wind flying:  The 79 style 3sq.m 1Skins will accept shortening of the front 4 bridles by up to 80mm in 3 stages for light wind flying- the ranges being say <10 to 15, 12 to 20 and 15 to >40km/hr.  In India it was noticeable that an earlier (40 series) 6sq.m 1Skin (ex Kaixuan) with 120mm shortening was not only able to fly in lighter winds than any conventional ram air pilots there (and pretty similar to Volker Hoberg's very lightweight plastic ram air pilot) but even by 10km/hr had substantial lift- enough to keep maxi show kites up when nothing else could.  This kite was also marginally better in the very lightest wind than 1Skin 79 on the maximum shortening it would accept. This is very likely to be a consequence of the proportionally wider LE of the 40 series- which effectively decreases the kite's angle of attack by a few extra degrees (but promotes more LE collapse in stronger winds). The 6sq.m 1Skin also collapses the centre section of its leading edge before any shoulder distortion if left on more forward bridle settings as wind speed increases, flying quite reliably in this mode.   When production at Kaixuan is upgraded to the latest series, it will therefore be sensible to continue production of the 6sq.m 40 series until a better light wind design comes along, if ever!

Before India, I had tried various reactive bridle systems on the 79 series in an attempt to eliminate the requirement to make bridle adjustments for different winds. I've often discussed the possibilities of this before but know from long experience that this part of the swamp is particularly impenetrable, so I've been very reluctant go there, fearing I'll be lost forever.  But surprise! A lever bridle system fitted to 1Skin76 yellow (pretty similar to 79 with a reduced lower LE, just not quite as refined) worked immediately the best- noticeably superior to a pulley system with very similar geometry fitted to 1Skin78 (pink/purple). Flying side by side in India, wind rising from not enough to say 30km/hr max, 76 stayed up more consistently and moved around less.

Before India, had also tried a tried a version 1Skin 77 pink/purple with a porous rear section.  The idea is to allow some pressure bleed in stronger winds so as to reduce maximum aero forces while not significantly effecting light wind behaviour.  First attempt at this with fine mesh gauze was unsuccessful- required substantial wind to fly at all.  Second attempt was also unsuccessful as the uncoated fabric used distorted during sewing causing just about the most left flying 1Skin I have ever built. Third try, stabilising the fabric with hair spray before cutting and sewing was successful.  Light wind flying is not noticeably different, and it does seem a little less violent in strong conditions.

But none of these tries has lifted the wind range by enough for reliable upper end kite event flying so far (though the lever bridle is promising.)

Next attempt was to start a lower aspect ratio pilot kite series, called Singers (but in the same number sequence), basically by removing one centre rib from the 1Skin pattern- plus some consequent changes (every template is different). Singer 80 red is 2.25sq.m and has an aspect ratio of 0.85 (1Skin 79 is 1.08).  Although this doesn't change the relative effect of the pendulum restoring moment at all (it's a function only of the distance between the centre of pressure and centre of gravity) it does improve recovery from any angular deviation because tip drag is closer to the kite's centre line. Singer 80 is therefore quite volatile, inclined to looping, but does not show signs of superstability in any conditions- not even in winds that are blowing branches off trees.  To control this volatility, various tails were tried.  The principle of tails, if they are to control volatility without exacerbating superstability is that they should be self-supporting (generating enough lift to support their weight), have as little drag as possible within this constraint, and be attached at a single point on the kite's centre line.  In effect they provide damping and shift the kite's centre of gravity rearward while moving its centre of lift rearward as little as possible.  For Singers, a fabric strip tail of around 30m length and weighing about as much as the kite itself works well.  With such a tail, Singers 80, 81 (porous rear section) and 82 will fly reliably in VERY strong winds- in 1 case breaking the main line with something above 40kg pull I think.  To reduce volatility a bit, 83 and 84 were then built, still with just 4 cells and 16 bridles, but a little wider in each cell for area 2.5sq.m and aspect ratio 0.93. (Aspect ratio,- span squared divided by area- is the strongest determinant of the volatility/superstability boundary for single line kites, higher aspect ratio kites are less volatile , more superstable).    These are very pleasant kites, not requiring tails in normal conditions.  With a lever bridle, 83 had some of the most reliable flying in mid-range gusty condition that I have yet achieved at the kite factory field with single skin pilots, very steady, with quick recovery from wind induced tilting, no signs of looping (volatility) or diving off (superstability). But when changed to a pulley bridle, it flew less reliably- though maybe this was a wind effect. Singer 84 (yellow/orange) is identical to 83 so as to enable side by side testing of different reactive bridle arrangements.  

From 2nd Feb '17: In light variable northerly wind, (Northpark) even when flying Singers 83 and 84 side by side it's difficult to distinguish kite differences from wind changes. Indications so far are that differences between the pulley and lever bridle systems are probably in the details (especially bridle self-adjustment limits).

Limiting the amount by which the front bridles can be let out relative to other bridles seems to have a profound effect on stability.  Allowing them to lengthen by say 25mm relative to the rear set, and having this change occurring at lower wind speeds appeared to promote quite wild swings to either side in light winds. The most likely explanation for this is that it's a centre of pressure migration effect- the front letting out moves the centre of pressure too close to the kite's centre of gravity for rapid response to wind gust induced perturbations. But by limiting their extension to zero (best for light wind flying), will strong wind flying be compromised?

If c of p migration from small changes in 'as bridled' angle of attack does have such significant effect on stability as this indicates- and this seems likely seeing as these are probably changing the relative length of the restorative pendulum by significant percentages (25% +/-) then total bridle length can also be expected to have significant effects- shorter for more responsiveness, longer for slower response.  This accords with other kite stability experiences- for example, moving the line attachment point further from the kite is a known way of reducing volatility for ram air and sticked kites.  Unfortunately, for single skin kites, shortening (to improve responsiveness) is not a realistic option because their skins cannot resist compressive loads. In stronger gusty north-west at Northpark, and after further detail refinements, Singer 83 with pulley bridle is very little different to 84 (lever) in crash frequency (short line) but seemed to use more sky to recover.  Could this be from other minor differences?  1Skin76 yellow (with lever bridle) was if anything more reliable in the same wind except that it tended to hang to left or right- superstable- and eventually broke the lever.  1Skin 79 grey/blue was pretty good also- scarcely less fly-able.  Best by far though, in these conditions was Singer 82 purple with one tail set.

Next test should be Singer 83 with fixed bridle against Singer 84 with lever bridle- it is easy and quick to immobilise the pulley.  And then try with a tail- 30m x 200mm single strip rather than current double 80mm wide (as sewn) strips.  

 Octopus and Serpent: The latest pink 20m Octopus, made and bridled the same as the older blue one (which is such a reliable kite across the range) had never flown satisfactorily, after more than a year, many attempts and comparisons.  It went to Dieppe in September, where I expected to get it sorted, but was in a bag lost by Southern China/Air France with the 55m Serpent, which I didn't get back until on the way home again.  Before India I finally noticed 4 bridles in the rear body area that were approx. 50mm too short.  This fixed it- which surprises me a lot because I have- now had- the view that these bridles are relatively unimportant.
35 m green Serpent: After very many reconstructions of the head leading edge and the tail- so many that the fabric was almost falling apart from needle holes in places- allowing some airflow to escape through the tail at the base of the head - as suggested by Volker Hoberg- and making these lateral slots into 4 scoops did eventually get it flying.  Earlier attempts at providing some tuneable drag at the tail/head connection had all failed.  At one stage I added orange peel shaped scoops down the edges of the tail- more and more until eventually it became obvious, even to me, that adding ever more drag was not the answer. Although in some specific winds it would occasionally fly briefly, the tendency to dive off over to one side or the other rather than apex was basically unstoppable.  For a long time I thought that the problem was just that I couldn't get the drag to act close enough to the head and that what I was seeing was typical superstability caused by having a tail that's too long.  This effect can be seen whenever a long tail catches on something- the kite then slowly falls off to one side or the other.  However, when I did get to look at this now-flying-OK Serpent from the side (in India) the scoop effect at the head/tail joint was just too unsightly- a big ugly protuberance where there shouldn't be one.   Next try was to re-sew, again, with 4 longitudinal slits in the first few metres of tail.  This does look OK now and is close to flying satisfactorily. It could still do with a bit more drag in this area, but its's possible that enough improvement can come from further bridle adjustment.  What I have found during the many days I've spent trying different bridle arrangements on this kite is that it does respond quite well to the sides of the head being pulled forward to create a sled effect.  I've also found that if the leading-edge bridles are not at least 150mm shorter than the face bridles (which are all made the same length as a start point), the head compresses longitudinally into a series of ridges. 

Still a work in progress!

On the more positive side, both the Serpent and the Octopus can now be trained satisfactorily- which is very useful in the limited sky space at most events.  The top kite needs to always fly at a higher angle than lower kites.  To this end, the 79 series 1Skins are best- or the 40 series 6sq.m in light winds.  But if the top kite ever falls back or drops to one side, the lower kites will almost always wrap irretrievably around the top kite's line. Setting the lower kites for wind speed higher than they are actually experiencing also helps prevent them overflying.  Ram air pilots do not fly at high enough angles to be useable above single skin show kites.  I've now stacked up to 3 single skin show kites below a 1Skin pilot satisfactorily, and expect more will be possible.  

March 2018; Castles in the air:

On account of being physically inoperative at present (Shingles), I still haven't actually tried using aeolian vibrations in rear bridles to automatically increase A of A in stronger winds- potentially leading to the "one setting all winds single skin kite" holy grail. 

But it may work even better than I have understood so far because such rear bridles will not only be wind speed sensitive (increasing their drag far more rapidly than by the square of AW speed in the desired range).

And have the multiplier effect of pulling down on the rear of the kite - I know from experience that 20mm shortening on the rear bridles of a 3sq.m 1Skin is enough to offset 40mm shortening of the LE bridles- which 40mm is pretty much enough to push the light wind threshold down to a very competitive (with conventional ram air pilots) level.

But these Aeolian bridles will also be flying angle sensitive, especially those sections of them that are furthest from the kite and unaffected by kite-caused flow effects.  At low flying angle, these bridles will be almost aligned with wind flow- so will not be inclined to vibrate- but at high flying angles, flow direction across them becomes increasingly square-on- which will promote vibration.

And the best thing of all about this approach to the problem is that it should work equally well for tailed SS's (like the Octo and Serpent and Singer) as for tailless versions like the Boomer and 1Skin.  The best automatic wind speed adjusting system I've found so far is the lever bridle, but this is subject to a reverse effect in very light winds, because tails hang straight down until there is sufficient wide speed for them to stream out, (which requires more wind than the kite itself requires to start flying on account of their low aspect ratio).  This is only costing a km/hr or 2 at the bottom end, and the lever bridle system does still improve the light end for kites with tails, but at some stage this 1 or 2 km/hr becomes critical to performance- because minimum flying speed is perhaps the single most important criteria for kite usefulness.

On the other hand, I think there is now have data suggesting that making the rear portions of the kites porous to improve top end stability is unlikely to be successful.

 The idea is that at low wind speeds, flow is very little changed because the leakage flow is calibrated to be pressure sensitive by hole size (uncoated ripstop works well for this in the range that's of interest). In stronger winds, the leakage causes the kite's C of P to move forward , while also causing the kite's "effective bridling angle" to move rearward - reducing overflying and the lower angles of attack that promote LE collapse.

Good theory.

But as far as I can see now, just lengthening the kite's rear bridles in stronger winds (could be done elastically) should have pretty much the same effect, but when this happens on the 6sq.m 1Skins because of cumulative fabric stretch the kite becomes quite unflyable- terminally superstable as soon as there's enough wind for useful flying. This misbehaviour is dramatically and completely cured by shortening the rear bridles by 20mm or so, which rather suggests that the original theory is a crock.  Maybe there is still wiggle room here quantitatively- maybe some smaller change that will have a positive effect without bringing on the dreaded superstability, but exploring this would require a huge time commitment which doesn't seem justifiable by comparison to other available directions.