The Design and Development of a 4.9m Hydrofoil Catamaran.

(As presented to the 2nd Australian Sailing Science Conference, Hobart, January 1999)

E.J.C.Chapman and G.C.Chapman.

South Brent, Devon, UK

Abstract

Calliope is a round bilge wide beam single handed 4.9m catamaran designed to investigate the versatility of a hybrid displacement/hydrofoil sailing boat. Horizontal lifting foils under the dagger boards are controlled by surface sensors via a unique clutch mechanism, permitting three operating modes – displacement, one hull flight and two hull flight.

The background to the design was greatly influenced by the authors’ experiences and observations at Weymouth (UK) "Speed Weeks" held from 1972 onwards. However, this design is intended for recreational sailing operating from the confines of a particular boatyard and slipway, rather than speed record attempts. Recognising the need for good low speed displacement performance, incidence-controlled fully submerged lifting surfaces were chosen over fixed incidence surfacing piercing or ladder foil options. Trailing height sensor arms are used in preference to forward reaching ‘Hook’ sensors.

Dagger board or strut profile was initially based on some inappropriate received knowledge for surface piercing foils, causing severe ventilation when sailing close hauled. The solution to both this problem and one of heave instability were aided by modelling in a low speed flume.

Because she is not a race boat, sailing performance is measured by on-board logging of apparent wind speed, direction and boat speed. Data is presented in both time series and true (relative) wind polar form for comparative purposes and to verify VPP modelling. The latter, using data from wind tunnel and towing tests, has been used in the design of a larger craft currently under construction.

Nomenclature

A foil area

a angle of attack

b apparent wind angle to course through water

Cl lift coefficient

Cla lift slope, dCl/da

Cd drag coefficient

L equivalent trailing height sensor length

l leeway angle

TWA true wind angle

Va apparent wind speed

Vs boat speed

Vt true wind speed

VPP velocity prediction program

Introduction

The authors, father and son, have been designing, building and sailing unconventional small craft since 1971. We took part in the first Weymouth (UK) Speed sailing competition in 1972 and from 1974 competed in the 10 sq. m. sail area class until sailboards dominated the event from 1979. For us, the quest for speed was replaced with one for stability and improved all round performance. In 1991 we decided to apply the lessons we had learned from the early speed sailing competitions to the design of a recreational hydrofoil catamaran for single handed operation from our local boatyard. The aim was to produce a boat that was stable, fast and fun in sufficient wind for foiling, but would also have good performance to windward and at low speed in light winds.

Background

While many competitors’ craft at Weymouth displayed short bursts of speed, few managed to traverse the 500m course without a "crash" of some kind. Of the small craft, the best results were from those who adopted the classic aeroplane surface piercing hydrofoil system shown in Figure 1. Attributable to Philip Hansford in 1971 (Alexander et al, 1972) the use of surface piercing main foils supporting most of the weight of the craft plus a fully submerged inverted "T" rudder for pitch stabilisation became known as the Mayfly configuration after the small catamaran to which it was fitted. Copied by others, most notably James Grogono and the Icarus syndicate, this arrangement yielded a string of world records for both boats throughout the 1970’s (Grogono, 1987). Unfortunately, this system suffers from the close coupling between lift and side force on the main foils as a consequence of their dihedral. G. Chapman (1985) showed that any change in the division of side force between the rudder and main foils, as occurs during an alteration of course, shift in sail centre of effort or wave action, alters the lift on the main foils. Often exacerbated by the helmsman attempting to compensate, the ensuing motion of skying and diving may lead to a "crash" – a dramatic reduction in velocity as one or more hulls hit the water – and a reduction in average speed. To physically isolate the lift and side force producing functions of the appendages necessitates geometrically orthogonal hydrofoil surfaces. With the absence of the inherent passive height control of the surface piercer, either a ladder foil arrangement or an actively controlled incidence fully submerged lifting surface is required. Active control held the promise of greater potential, particularly in waves.

A general arrangement of an aeroplane submerged foil system is shown in Figure 2 , while Figure 3 shows three different methods of controlling the angle of attack, or in the case of a fixed foil with a trailing edge flap, the flap angle, of the submerged lifting surface. The first of these was devised by Christopher Hook as the "Hydrofin" for powered craft. Applied to sailing craft including briefly Hansford’s Mayfly in 1974, Figure 4, structural deficiencies put back its development and it was not until 1976 that the system was successfully adopted for use on the sailing trimaran Force8 (Pattison and Wynne, 1980). It comprised of a forward reaching arm with a small planing shoe at its front end working a fully moving foil via crank arms and push-pull wires. A change in boat height above the water caused the arm to rise or fall relative to the boat, which in turn changed the incidence of the foil such that the lift it developed opposed the initial disturbance from equilibrium. As such it is a simple proportional system, but Pattison also incorporated a manual input for roll control, similar in function to the ailerons on an aircraft. To us, any forward reaching arm of this type appears unsatisfactory, for example if the planing shoe digs into the back of a wave or patch of floating weed. Although this was not reported to be a problem, the spray thrown up by these surface sensors was. Force8’s main shortcoming was with the foils sticking, having used acetyl journal bearings rather than balls or rollers, resulting in occasional broaching and crashing. Force8 was the inspiration for Greg Ketterman’s "trifoiler" Longshot (Ketterman, 1994), which persisted with forward reaching sensors but used a flexible structure rather than crank arms and bearings to break the dominance of sailboards in speed sailing in the 1990s

The second control arrangement in Figure 3b is that proposed by Phillips and Shaughnessy (1976) and consisted of an almost vertical "wand" pivoted about a point on the hull. Connected via a crank arm and push rod to a trailing edge flap, the variation in drag on the wand with depth of immersion provided the feedback needed for level flight. Initially believing that a bob-weight mounted in such a way as to cause the flap to rise with upward acceleration of the boat would dampen out the rigid body motions of the boat in waves, their analysis showed this to be incorrect. They suggested that a more practical arrangement would be to have a planing device on the end of the wand to follow the water surface. The steep wand angle and aft-of-foil location is a feature of Dr Sam Bradfield’s series of incidence controlled hydrofoil sailing boats, including the Windrider Rave (Guillaumin,1998)

The third arrangement is that due to Simmonds (G.C.Chapman, 1977), with a trailing, planing height sensor arm working a flap via linkages. In the case of Mark Simmonds’ A-Class catamaran Rampage, only one such foil was fitted to the Port hull, and it was not until Hansford adopted the system on both sides of his trimaran Dot (later Philfly) in 1985 that truly stable, level flight on a variety of headings was seen.

The authors’ own Bandersnatch, a 14’ ultra-light catamaran (G.C.Chapman, 1991) served as a test platform for a variety of configurations, finally converging on the trailing, planing wand/trailing edge flap system, Figure 5 and Figure 6. As such, this craft demonstrated excellent stability foilborne. However, with low displacement hulls of almost zero rocker and large surface piercing additional lifting surfaces, low speed manoeuvrability and displacement windward performance was poor.

Design Aim.

The requirements for a new boat were:

(i) Retain the excellent flying characteristics of Bandersnatch

(ii) Improve windward performance, displacement and/or flying

(iii) Reduce time to deploy and stow lifting foils

(iv) Be operable single handed from our favourite boatyard in Plymouth, UK.

Design Realisation

To meet these objectives we chose to retain a catamaran configuration but switched to vertical struts sliding through the hulls, as on a conventional cat, and adopted fuller, semi-circular underwater hull lines with modest rocker, Figure 7. The boat was to be called Calliope after an RN ship of that name whose Captain lived at the authors’ address in the 1800s.

The beam of the craft was limited by the width of the boatyard gate, and the length by the size of our garage! Although a wider foil base was desirable it would have probably been necessary to choose a trimaran platform, which in turn would need to be dismantled to get to and from the water. The ability to adjust trim through moving crew weight fore-and-aft and athwartships was considered desirable, and this could not be done if restrained in a central cockpit.

By operating the foil control system on the windward side only, it was envisaged that in light winds the windward hull only would fly, overcoming the disadvantage that a wide beam cat has in those conditions.

Structure

The hulls were built from 3mm tortured ply bonded with glass tape and epoxy resin. Simple 3" diameter 16swg aluminium tubes formed the main beams and a 2" 16swg tube the rudder beam. The hulls have three watertight compartments each, the centre one containing the centre-board or strut case. Each bow is a separate structure, originally concealing the linkage from the height sensor to a crank arm at deck level. These foam/GRP bow units are detachable so the reduced hull length fits in the garage.

Foils

Initially, the struts and foils from Bandersnatch were used. These have a lifting section approximately NACA 0012 with the last 25% forming the trailing edge flap, Figure 8a.

The struts remained of bi-ogival section but were stripped of their anti-ventilation fences so that they could be retracted through the centre-board cases.

Two rudder blades are carried and deployed from a single, central rudder stock. A spring loaded ‘hack’ blade is used during launch and recovery – it does not matter if it hits the ground – and an inverted ‘T’ foil that can be locked down once in sufficiently deep water. The horizontal incidence of the rudder foil can be pre-set by rotating the rudder stock around the rudder beam and securing it with a pin. It is set to zero degrees and is not usually adjusted afloat. The lifting and strut sections were originally 0012 and bi-ogive respectively.

The height sensing wands (or ‘feelers’) were pivoted from within the bow pods and shaped to fit snugly under the hulls’ centre lines when not in use. Small triangular planing surfaces at their tips provided enough force through the crank and wire linkage to pull the foil flaps to rise (i.e. flaps down) if the boat was too low. A powerful elastic pulled the flaps to dive (i.e. flaps up) if the boat rose too high, the wand having to overcome this as well as the flap moment when boat height low.

Foil Structure

The initial Bandersnatch struts, foils and rudder were made from a glass fibre/polyester hand lay-up on a softwood core. Carbon reinforcement was included in the high stress areas. The trailing edge flaps were made of carbon in a two part mould, core material being high density filler.

The shells of the later fully moving foils, described below, were made in two part moulds similarly to the flaps. After laying up the shells the tube to accommodate a stainless stub shaft and a continuing wooden spar were firmly glassed in. After reinforcing the edges of the shells and filling each half with polyurethane foam the two halves were glued together. This proved to be inadequate. On two successive outings first the Port inner, and then the Starboard inner foils shed their top shells. In both cases they were the lee foils. Surprisingly the boat went on flying, though at reduced speed. These two had been ‘up-side-down’, with the tube and spar glassed to the lower shell. Replacements, and later models, have been carefully handed to make sure the major strength of the shell-to-tube join is upwards.

Clutch and Lost Motion Mechanism.

A unique feature of Calliope’s foil control system is the clutch at the top of each strut. Controlled by two strings they permit remote engage/disengage of the wand-to-foil connection, locking the foil in neutral and allowing the wand to trail freely for minimum drag when de-clutched.

A fixed incidence surface piercing foiler would take around a minute to change from displacement to foiling mode by stopping and lowering the cumbersome foils, or put up with sluggish performance from dragging a considerable wetted area, at incidence, through the water. With clutch operated foils the switch from displacement to flying takes only a second and can be done while moving. This offers a huge advantage when sailing in light winds marginal for foiling.

A lost-motion linkage at the top of the main foil strut was intended to prevent excess loads developing in the control system, the wand-to-stow or fully up position exceeded the corresponding flap-to-rise.

Sail Rig.

For the rig we chose a fully battened main/rotating mast with a wishboom sheeted to the mid-point of the aft beam. Bandersnatch had suffered from an undesirable bow-down twist of the leeward hull due to the high sheeting load applied via a track to leeward of the centreline, which we wished to avoid. A small jib was retained from the previous boat as an aid to tacking and helm balance, its tack carried on a light bowsprit.

To begin with, a circular section mast was used, later replaced with a D section Z-Spars Z-170.

Instruments.

Initially, Smiths Instruments apparent wind speed, pitot log and a Speedwatch log all ex-Bandersnatch were carried. The wind speed cup anemometer was mounted on a forward reaching strut on the mast and the orifice for the pitot log built into the leading edge of the rudder foil strut. The Speedwatch impeller was mounted on one side of the rudder strut, with its pick-up coil above the water on the stock.

Preliminary Results.

The first few sessions afloat were disappointing. Although the boat sailed well enough displacement and flying one hull, excessive ventilation down the leeward main foil strut as soon as the leeward hull cleared the surface prevented two-hull flight. Increasing the chord of the struts to the maximum within the constraint of the centre-board cases did little to improve matters. Only when the lost motion control linkage ‘locked-up’, jamming the wand in its up position (foil-to rise) did the boat get both hulls off the water, and then only transiently, Figure 10.

Reasoning that the disturbance created on the surface of the water by the wand tip was in some way initiating ventilation, a simple experiment was set up in a low speed flume to investigate. A section of old foil strut was held at an angle to the flow whilst a kitchen whisk agitated the surface just upstream. Ventilation was immediate. On the boat, the leeward wand was secured in its stowed position and the foil flap angle locked to a mid-to-rise position, reproducing the two hull unstable flight experienced ‘by accident’. The in-line under hull wand location was abandoned in favour of a deck mounted, inboard arrangement.

The ability to fly both hulls, although now possible, could only be done when sailing well off the true wind. What the true wind angle actually was was uncertain, so as a parallel exercise an improved instrument system was developed (outlined below).

With inboard wands, flapped foils and bi-ogive struts Calliope was entered in a privately run time trial event at Weymouth in October 1993, recording a best 500m average of 15 knots.

Foil Sections.

Our choice of strut section had been greatly influenced by Alexander et al (1972) and the success of ogival surface piercing foils. Although we were aware that ventilation might be a problem without fences, we were so wedded to the received idea that a sharp leading edge was necessary to ‘cut’ the water surface, and that an ogive or bi-ogival section would have the required un-peaky pressure distribution that our progress was delayed.

Bethwaite (1993) had conducted experiments on 10 surface piercing rudder blades at speeds up to 25 knots, albeit transom hung. He concluded that the best compromise for low and high speed performance, in terms of lift/drag and anti-ventilation properties, was for a section with a fine but highly polished parabolic leading edge. Describing 18’ skiff development, he reported that a 0.5% of chord radius leading edge proved optimum in moderate winds, falling to 0.4% at boat speeds up to 28 knots. This information came to our attention at an appropriate time!

To overcome the deficiencies with Calliope we adopted a NACA 0012-34 section with a 0.4% nose radius for the struts of a new set of foils. We also chose to use a balanced, fully moving foil instead of a fixed foil plus flap, initially of 0015 section but subsequently of NACA 632-015. To ensure the foil would not stick at one incidence, as had plagued Force8, needle roller bearings were used in the hub at the foot of the strut. Figures 8b and 8c.

The lifting surfaces have the appearance of being swept forward In fact they are straight but tapered about their quarter chord point. Theoretically, for a 2-D symmetrical section both the centre of pressure and the aerodynamic centre lie at the quarter chord point. In practice our initial 0015 foils exhibited a tendency to pitch the ’wrong’ way, and we were forced to extend the trailing edge aft, reducing the aspect ratio. Following their structural failure the opportunity was taken to use the NACA 632-015 section whose experimental Cm(c/4) - a curve (Abbot and von Doenhoff, 1959) indicates that the aerodynamic centre is at 27% chord. When pivoted about the 25% chord line, a small restoring moment is created when at incidence, tending to feather the foil. The Cd-a curve shows a modest laminar flow ‘bucket’, and the pressure distribution a less steep adverse pressure gradient towards the trailing edge than the other sections.

The new foil assemblies, with the 0015 lifters, were available for the 1994 season. Although leading edge ventilation had been banished, we encountered a stability problem.

Dynamic Response and Longitudinal Stability.

On occasions, usually when sailed by E JC (of less weight than GC), at the point where the leeward hull would clear the water when accelerating from rest, the windward hull would start to heave up and down with increasing amplitude at a frequency of about 1 Hz. Increasing crew weight by sailing two-up eliminated this problem, but this was not a useful solution.

A full treatment of the dynamic stability of this craft would require a six degree of freedom analysis and is beyond the scope of this paper.

A ½ scale physical model of one foil limited to 1 degree of freedom (heave) only, initially built as a constructional pattern, had briefly been fitted with an electronic height control system. This had demonstrated a tendency to oscillate if the gain was set too high. It was tested by fixing it to a trolley and pushing along the edge of a swimming pool.

On Calliope the gain was substantially reduced to 1:1 between the rotation of the wand and the foil. This permitted the removal of the lost motion link and simplification of the clutch mechanism. No further stability problems have been encountered, and we were able to record a 500m average of 19.01 at the 1994 Weymouth event, winning the prize for fastest non-sailboard by more than 3 knots from a standard Hurricane 5.9m catamaran.

Instrumentation and Performance Measurement.

To answer the question "what is the true wind angle" an instrument to display this and other quantities was designed and built, (E J C Chapman, 1993). Accepting data from sensors for apparent wind speed and direction and boat speed, a look-up table containing calculated values of true wind speed, velocity made good (sailing upwind) and true wind angle (sailing downwind) is addressed and the result displayed. Leeway is not measured and is assumed to be small. The size of eprom memory dictated a coarse 5 bit resolution for the measured parameters - 0 to 31 knots in 1 knot steps for Va and Vs, 5.625 degree steps for b .

With the addition of a battery backed memory card, the device can log up to 120 minutes of data at a rate of 1 Hz from the Speedwatch sensors. No running averages or damping is applied to the data. The anemometer and wind vane are mounted at the front of the bowsprit, the rotating mast being unsuitable for this purpose. For improved sensitivity at low speed the log impeller pick-up coil was embedded within the rudder strut. The original Smiths masthead anemometer was retained, which together with the bowsprit unit enabled some wind speed gradient information to be collected.

Analysis of sailing performance is carried out off-line. Both time series and polar diagrams can be displayed for any section of a record. The polars shown in Figure 12 present the raw, 1 Hz data plotted as a histogram or frequency density of boat speed against true wind angle. The darker the colour the more time the boat was being sailed at a particular speed and heading.

The turbulent nature of the wind and the unsteady, non-linear behaviour of a sailing boat mean that the generation of ideal, steady state ‘smooth-line’ polars, with any degree of statistical significance, is a challenging goal. Initially, we ran a peak average search routine over the data to find the fastest average, over typically 10s, for each 10o bracket of true wind angle, (GC and EJC Chapman, 1995). More recently, by searching for short pieces of sailing where the variance of the three recorded parameters is within preset limits it was hoped to get a more reliable, reproducible measure of performance. The drawback with this method is the apparent wastage of data - it is rare for all three variables to be sufficiently steady for a 10s average to be valid. Higher sampling rates and resolution may permit shorter sample periods to be considered, but until then it is felt necessary to return to the raw data and present it in form shown here. It must be remembered, however, that the true wind fluctuates in direction as well as speed, a factor which cannot be accounted for without logging compass heading. For this reason, we call these diagrams true-relative-wind, boat speed polars.

The results plotted in Figure 12 show the improved performance in October 1994 over that 12 months earlier. Both sets of data were recorded on Portland harbour during the Weymouth events in similar weather conditions.

As an aide to assessing the probability of success in a 500m timed trial, the record of Vs can be searched for the fastest 500m or any other distance, Figure 13.

Velocity Prediction

As well as analysing performance as above, an early exercise was to take the data from a fast run at Weymouth and attempt to put numbers to the whole set of steady state forces acting on the foils, sail and hulls. This developed into a velocity prediction program (Vpp) for the three modes of operation – displacement, one hull and two hull flying – and proved useful in determining how best to sail the boat as well as a design tool for successor projects.

Data for the Vpp is largely empirical, but includes the results of full size towing and 1:21 scale wind tunnel tests, previously unpublished and presented in Figure 14 and Figure 15.

The towing test was conducted on a flat calm day on Portland harbour. Calliope, with a crew of two, was pulled by a RIB from straight ahead by a line connected via a spring balance to the forward main beam. The mast was up but no sail set. Without the forward pitching moment from the sail, the rudder foil had to be set at a high incidence and the both crew had to sit well forward for level flight when flying. Three runs were conducted: displacement, flying 1 hull and flying 2 hulls. Maximum speed was 15 knots before the difficulty of steering from such a forward position caused a near shipwreck and the tow was slipped.

The model used for the wind tunnel tests is shown in Figure 16. It was tested at two typical sheet settings, and with the hulls alone at three heights. The small model with its brass sails and low wind speed make for incorrect Reynolds scaling, but to the extent that much of the flow over an object of this kind is separated and ‘messy’, it could be classed as a non-aerodynamic structure for which correct scaling is less important. The resulting plots of Cl and Cd conform to those produced elsewhere, so we believe they are not so far off the mark. The wind tunnel was the small recirculating teaching tunnel at Exeter University’s School of Engineering, with a jet 416mm wide x 455mm high.

What is of interest is the wind tunnel data for the hulls alone (plus 1 crew). The increasing lift and drag with height above the ‘sea surface’ suggests either an influence from the velocity gradient in the tunnel and/or a change in flow pattern. The low height state was with the hulls touching the surface rather than immersed in it, so the displacement condition was not reproduced.

By combining this data together with that from standard references for the foil sections used, the Vpp works by iteration to find a balance between the aero and hydrodynamic forces acting on the boat.

The resulting printout for Calliope flying both hulls, Figure 17, differs from usual Vpp output since it shows the envelope of performance for the chosen true wind, rather than just maximum speed. Square symbols show where the windward foil will be to dive, circles where it will be to rise. There is a blank area in the down wind sector due to lack of down wind sail data, but since the boat tacks down wind this is no hardship. The figure suggests that in this case it is not possible to fly both hulls at less than 8 knots. In reality 9 knots is about the minimum flying speed.

Final Performance and Discussion.

By the end of the 1994 season the boat was performing well. She had met if not exceeded our expectations. The polar shown in Figure 11b remains typical of her performance. Figure 18 shows the final general arrangement. In 1995 a new, larger sail was fitted which reduced the true wind speed required for two-hull flight by a knot to 10 knots, but did not increase maximum speed when sailed single handed, Figure 20. With two-up, maximum speed in stronger winds is predicted to be greater since the drag penalty incurred through providing righting moment by additional weight is less than that associated with the windward foil developing downforce. Unfortunately, attempting to investigate this has resulted in a number of breakages.

Inevitably in a project of this kind more than one alteration to the boat may occur between outings, and so it is difficult to assess the affect of any one factor. In particular we do not know whether the improvement on changing from 0015 to 632-015 foils was due to the increased area and aspect ratio, or from laminar flow. Bethwaite (1993) maintains that when highly polished his blades enjoy laminar flow. Parsons (1998) says "The sure proof that the question of how much laminar flow can be obtained at full scale has not been answered to everyone’s satisfaction is the heat and vehemence that divides the international naval architecture community on this issue. There have been no conclusive experiments carried out and made public that convinces everyone of what is actually going on. Based on full scale measurements carried out on the large K boat for the 1988 [America’s Cup] contest, the New Zealand technical community believes that significant laminar flow can be obtained".

By comparison with America’s Cup standards, our performance measurement and velocity prediction accuracy and resolution are deplorable, but we are seeking improvements in the order of 10% in 10 – 20 knots rather than 0.1% . For our purposes the Vpp is useful and it will doubtless be improved. We believe it may be unique in being applied to a dinghy sized craft and tied to afloat recorded measurement.

There is plenty of scope to model the dynamic behaviour of the craft, a subject we have barely touched on. Measuring performance at a higher sampling rate and resolution to capture the dynamics is an aim for the future.

In the last season Calliope was afloat, 1997, we began to explore windward performance with greater care. We had thought she would go to windward better by sailing displacement or flying one hull. The Vpp suggested that either flying both hulls or fully displacement would be better than one hull flight. Moving the sheeting point to windward and sailing on the Vmg meter appears to confirm this.

Calliope is now in refit and a new and larger sister Ceres is under construction. At 5.8m this craft – essentially a scaled up version – is designed for a crew of two, of more robust construction, and intended to see if this design is competitive against similarly sized conventional catamarans.

Conclusion.

The use of incidence-controlled submerged foils has been successfully applied to a 4.9m catamaran that can be operated single handed from a conventional boatyard. Problems of ventilation on the vertical main foil struts were overcome by adopting a section with a fine, highly polished, parabolic leading edge. Balanced lifting foils of moderate laminar flow section have proved satisfactory. A tendency towards heave instability was eliminated by reducing the effective gain between the height sensor and foil.

Logging data from on-board wind and water speed sensors has enabled a record to be kept of sailing performance. Wind tunnel and towing test data enabled a Vpp to be written which has helped us to get the best out of this craft and in designing a successor.

Acknowlegdements.

The authors wish to acknowledge the help and support of the following: Avril Chapman for use of her drawing room as a boat yard, Bob Downhill for resurrecting the Weymouth speed weeks and towing Calliope, the Junk Rig Association (sponsor of EJC at Exeter University) and Frank Bethwaite for his book and encouragement.

(Return to home page).

References

Abbott, I.H., von Doenhoff, A.E. (1959). Theory Of Wing Sections. Dover, New York. pp 524-525.

Alexander, A., Grogono, J., .Nigg, D. (1972). Hydrofoil Sailing. Juanita Kalerghi, London. ISBN 0 903238 004.

Bethwaite, F. (1993). High Performance Sailing. Waterline Books, Shrewsbury, England. ISBN 1 85310 337 3. pp 237-243.

Chapman, E.J.C., (1994). Small Craft Performance Measurement, Multihull International vol .27 no. 320, pp 274-275.

Chapman, G.C. (1977). The John Player 1976 World Speed Sailing Record Attempts. AYRS Publication no.87, pp 3-9.

Chapman, G.C. (1985). Why Aren’t Hydrofoils Faster. Multihull International, vol.18 no. 206, pp 68-70

Chapman, G.C. (1991). Stable Sailing On Submerged Hydrofoils. AYRS Publication no 108, pp 7-13

Chapman, G C., Chapman, E.J.C. (1995). Instrumented Perofrmance Measurement And Assessment. AYRS Publication no. 119, pp 7-28.

Chiu, T.W., Zino, A.J.P. (1996). Nwing 1.0 For Windows. School of Engineering, Unversity of Exeter.

Grogono, J. (1987). Icarus, The Boat That Flys. Adlard Coles Ltd, London. ISBN 0 229 11803 8.

Guillaumin, J.C. (1998). WindRider Rave Test Report. Multihulls World, no 43, pp 64-67.

Ketterman, G. (1994). Hydrofoils - What Makes Them Work So Well?. Multihulls, vol 20 no. 3, pp 59-61.

Parsons, B.L. (1998). Experimental Techniques For Model Yacht Performance Evaluation – A Review. Ocean Engineering International, vol 2 no 1, pp 1-18.

Pattison, D.R., Wynne, J.B. (1980). Surface Piercing Versus Fully Submerged Foils For Sailing Hydrofoils: The Design And Development Of Two Small Sailing Hydrofoils. High-speed Surface Craft, conference papers. Kalerghi Publications, London. pp 350-365.

Phillips, W.H. and Shaughnessy, J.D. (1976). Investigation Of Longitudinal Control System For A Small Hydrofoil Boat. J. Hydronautics, vol 10 no 2, pp 39-45.