How we won the 35th America’s Cup
As is ever the case with the America’s Cup, the fastest boat won. ETNZ designed and developed a cat that was faster on all points and in all conditions than any other. How was this achieved on a budget a fraction of those enjoyed by its rivals? Words & Graphics Christian Stimson, Images Supplied
The answer lies in an open-minded approach to design, identifying the best-performing option and working to remove the obstacles to it rather than compromising. That, and having great faith in your design and simulation tools, since the budget, geography and timeline did not allow for full-size testing alongside the other challengers.
Much has been said of the ‘brutal debrief’ the team went through following the defeat in San Francisco. Grant Dalton recounted that 20 or so points came out of the meeting as the highest priorities for next time. Top of that list was technology and capable people, encouraging the ‘new generation’ of sailors coming through the ranks, and giving individuals responsibility without constraints.
Glen Ashby explained the direction given to the entire team was to ‘throw the ball as far as possible, and then run after it hard’. That meant envisaging the best possible design, then overcoming obstacles along the way to make it real.
And that’s exactly what they did – combining exceptionally efficient yet sensitive foils demanding a complex control system, with plentiful hydraulic power to allow precise and frequent adjustments to keep them flying.
We’ve heard how cycle power was discounted early on by other teams because ‘the clips would get tangled in the netting’ or ‘the mount/dismount would be too slow’. With each of these arguments the other teams inched themselves away from the Cup-winning configuration.
Credit has to go to the entire ETNZ design team. Without doubt there were small victories along the way that made attaining the big goal possible. That said, the design process was overseen by Dan Bernasconi, who no doubt brought a rigour to the team’s processes from his experience with Formula 1, and it is assumed he was responsible for the delivery of – if not the actual ‘go/no go’ decision – the design path ultimately adopted.
Guillaume Verdier, responsible for designing the foils, developed a concept and geometry that drew upon the lessons learned in San Fran and took it to the next level. ETNZ’s foils reportedly generated the same lift from 20 percent less area (and hence less drag) than Oracle USA’s.
They could be configured to generate more lift for the same drag, or anywhere in between. Chris Draper, wing trimmer for Artemis racing, acknowledged the ‘aggression in the design choices’ ETNZ had made.
What made ETNZ faster?
A combination of many things, each reinforcing the effectiveness of the others in a virtuous, upward spiral rather than a vicious, downward one. The adoption of cyclors is the most obvious, but they are really the solution to another problem – how to provide sufficient power to control the foils.
They had the additional spin-off of less windage/drag than traditional arm grinders while affording the helmsman a clear view forward, and also allowed cyclor Blaire Tuke to have his hands free and focus on the fine control of the foils.
This division of labour – allowing Peter Burling to focus on steering and observing the race track, Glen Ashby to trim the wing and jib and Blair Tuke to control the ride height – ultimately turned out to be more effective than the traditional helm-tac-nav triumvirate which loads the helmsman with both ride height control and wing sheet during a turn.
The dagger-foil geometry is the next major difference.
ETNZ’s foils (in blue) had a slightly inboard curved ‘shaft’ (the part inside the boat and that just reached the water) coming in to the upper foil at a noticeable knuckle. The upper foil curves outboard then a sharp transition into the horizontal flat and kinked tip.
Many comments greeted this kink, but to my mind it’s the horizontal element that holds the secret to their success – more on that later. Oracle USA (in red), on the other hand, exhibited foils with a straighter shaft, a much less pronounced knuckle between shaft and upper foil, a tighter radius into the lower foil and either a gentle curve or no deflection at all in the lower foil to tip.
Less easily detected was that ETNZ’s foil tips had a much higher aspect ratio – they were longer in span for approximately the same area. Broadly speaking, foil area determines how much lift you can generate; the span determines how much drag is attendant with that lift.
The greater span means that ETNZ’s foils generate less induced drag, or less tip loss. If you can generate less drag in total, the thrust from the wing will make the yacht faster in the same windspeed. If the foils go faster through the water, they generate more lift, increasing VMG, so you could reduce the area, which will reduce the drag further…
The down side is this is at full extension a reduced proportion of the total force generated (magenta) is vertically upward (red), and so reduces the amount of lift opposing gravity. As a result more area is required, which comes with its attendant drag penalty. (see pix below)
Back to the significance of horizontal element. Oracle’s foils looked like a more refined version of what they had in San Francisco. The foil is a slightly curved shaft, so that its lower part oil changes its angle to the horizontal as board extension is increased.
ETNZ adopted a horizontal tip with all the force acting vertically – it needed less area to lift the same weight of yacht. As well as adding area, the kinked tip increased span to reduce induced drag and possibly, quite cleverly, drawn with twist to reduce the local angle of attack and thereby move the centre of pressure (ie; lift) outboard which would increase righting moment (I concede this is pure supposition).
It’s reported that the basis of ETNZ’s control system was Luna Rossa’s ‘auto-pilot’ from their AC72 programme, which was offered once Luna Rossa had withdrawn from AC35. We don’t know which team submitted Interpretation no.70, but it looks a lot like what Glen Ashby and Blair Tuke were using.
The balance of transverse forces on an AC50 is such that the heel force from the rig is opposed by the lift force from the leeward dagger-foil and the windward rudder T-foil. The AC rule allows the rudders to be raked differently on each side, and adjusted in each tack or gybe, through a maximum range of 3o. This allows the windward T-foil to have a negative angle of attack, pulling the windward corner down and increasing righting moment, while the leeward one is neutral or close to it.
To increase the force generated upwind, when boat speeds are slower, the yacht is sailed in an exaggerated bow-down trim, pitching the stern up and inclining the T-foil further to increase its angle of attack. But there is risk associated with this as it brings the windward T-foil closer to the water’s surface and increases the danger of it breaking the surface.
To mitigate this, ETNZ employed two more techniques. The first was allowing the platform to twist under backstay load. This required clever engineering to ensure the platform only twisted just enough without breaking, and also meant less laminate and thus less weight was expended on the structure. ETNZ’s was reportedly the lightest of all the AC50s measured in Bermuda, and that again helps unload the dagger-foils further.
Second was the adoption of slight windward heel, very much like a windsurfer. This has three principal effects: one, to bring the weight of the entire wingsail over the centreline, increasing righting moment; two, to direct the force from the wingsail above the horizontal which marginally reduces the heeling moment and lifts the leeward hull, both of which unloads the leeward dagger-foil, meaning it produced less drag; and three, to immerse the windward T-foil more and so mitigate the risk of it being flown so close to the surface.
To maintain this tricky balance of windward heel, bow-down trim, hooked in on a twisting platform at 30 to 40knots on foils on the verge of letting go in a straight line is going to be hard enough. Add in tacking and gybing while the other team tries to stop you and it’s clear the reliance on the control systems functioning and the hydraulic oil pressure (to allow micro-trimming of the control surfaces) is critical to the success of the concept.
Furthermore, the smoothness and stability through tacks and gybes was partly due to the horizontal element. As the yacht carves through the turn with both boards down, the ‘old’ board on the outside of the turn is travelling faster than the ‘new’ foil on the inside of the track. This generates more flow over the foil, which produces more lift. In the case of ETNZ that force is vertical and helps keep the windward hull up as the boat settles on the new tack or gybe.
For Oracle that dihedral vee causes the outside foil to generate more force than the new, inside foil, and crucially that force (magenta) is inclined as it’s perpendicular to the vee. The inward component (green), being greater on the windward side, will try to push the yacht to leeward, increasing the local angle of attack of the new foil until the windward foil is retracted.
The net result is more drag and unevenly distributed vertical lift across the boat. All of which will slow you down and make it harder to balance the boat, and explains ETNZ’s faster speed and greater VMG through tacks and gybes.
The control systems – required to move the foils, rudders and wing (known collectively as control surfaces) and maintain the ride height and give ETNZ that elusive 100 percent fly-time – were equally cunning in their design.
We know from the Rule Interpretations (which make interesting reading in themselves – http://noticeboard.acracemgt.com/home/ac-class/ac-class-interpretations ) that one team (we don’t know which) requested clarification under Interpretation no. 70.
This concerned a tablet screen with a second, electrically-isolated, transparent screen, physically air-gapped and over-laid. The lower screen shows the target values calculated by the on-board computers, the upper screen the actual state/position of the various control surfaces.
But ‘auto-pilots’ are prohibited as the AC50 class rule 15.4 specifies that electrical systems for actuation and position sensing of control surfaces be electrically-isolated from each other but also that only manual inputs are allowed in terms of altering the control surfaces.
The crew member moves the control surface by either touching the upper screen at the position indicated as the target value or, as in subsequent Interpretation no.72, by adjusting some other slider, knob or buttons to align the actual position with the calculated, target position.
A large part of maintaining that balance is the wing. Glen Ashby’s ‘X-box controller’ would most likely have worked in a similar way to the board control system with an air-gapped overlaid LCD, and what was noticeable early on was the very different trimming technique he was using.
Most obvious was the absence of a winch, opting for a ram inside the wing base instead. Many wing trimmers have remarked that, unlike a soft sail, you can’t ‘see’ the effects of trim change, so it is best (easiest?) to rely on the target numbers for twist and play the sheet. While no doubt Glen Ashby had target numbers to refer to, he spent a lot more time watching the wing sail than any other wing trimmer.
He also had a different approach, favouring playing the twist on the (smaller) upper elements and minimised sheet adjustments, as opposed to a pre-set twist and playing the sheet on the (larger) lower span.
It caused BAR crewman Freddie Carr to remark that the ETNZ wing was ‘flapping like a humming bird’s wing’. This would have had a huge effect on the heeling moment. Having the longest lever arm it generated more moment for less hydraulic pressure lost and gave less variance in the side force on the dagger-foils than playing the sheet would.
Oracle took full advantage of the five-day break to re-configure their yacht. They reportedly shed almost 100kg, they combined their heavy-airs dagger-foil (lower drag) with light airs tips (high span, high lift), and opted to run the boards at less than full extension so the tips were closer to horizontal (but risking a touch-down).
The rudder T-foils were enlarged (for more righting moment), but the rudders themselves skinnier to compensate for the added drag. They also changed their wing trim style to be closer to Glen Ashby’s, at least as far as they could given the deck layout and hydraulic pressures.
The end result was they increased the yacht’s speed but at the expense of control, both in the sense of sensitivity and actuation, as they lacked the input devices and the hydraulic oil to manage the now skittish yacht.
The one race they did take off ETNZ, in the strongest winds of the match itself, could most likely be attributed to ETNZ nursing their critically-damaged foils – cracked as a result of being taken out of range in a heavy airs race with Artemis, but kept secret so as not to expose the vulnerability.
As the breeze died, so did Oracle’s chances, and ultimately ETNZ righted the wrongs from San Fran. By adopting efficient foils, sophisticated controls, plentiful power generated by cyclors, who, with their hands free, were able to share the workload and with their heads down generated less aero drag, ETNZ had a boat that was intrinsically faster, and that kept getting faster throughout the tournament through constant refinement.
8-1 was ETNZ’s match point score in 2013
8-1 was ETNZ’s race win/loss score in 2017
1-7 was Oracle’s points in 2017
17 is the name of Oracle’s boat
The writing was on the wall.