Chapter 16 The Sails
One hears of extraordinarily expensive sails being “designed by computer” and constructed from the latest “high tech” materials. Our look at the flow of air over model sails shows that, at the time of writing this book, the accurate prediction of the flow by computation must be impossible on cost grounds alone. This is not unusual in engineering and engineering designers do the best they can in the prevailing circumstances and let the design proceed by creating the best possible theory and then making, trying and improving. It has proved to be very successful . For sails this involves a two-fold attack. The first requirement is the evolution of a good shape and the second the evolution of a method of manufacture that makes the desired shape possible. Of course the use of a computer will help enormously in handling the geometry of the sail just as it can be used to produce shadows for a hull once a shape has been chosen. It can also be used to facilitate the lay of monofilaments radiating from the corners to carry the loads through the sails. The makers of sails for models could emulate the makers of full sized sails but this is a road that no one has yet followed. Instead we work with materials that are manufactured for other purposes and do the best we can.
The design of sails for model yachts
The preceding chapters have all been necessary to bring us to the point where we can consider how we might design our sails. The first thing to recognise is that it is no use having good ideas on the shape of sails if these shapes cannot be realised in practice. It follows that we must be clear about our constraints. A sail is simply a piece of fabric or film that is attached to a sailing rig at three points, the head, the tack and the clew and, usually along one edge, the luff. Normally the sail is stretched between the head and tack by the uphaul and downhaul. The foot of the sail is allowed to bow to form a curve the shape of which we measure as the camber. No matter how much we tighten the sail between the head and the clew the sail will always twist under load. We have seen that the standing rigging is not ideal but it is all that is available to use and we must accept it.
We need now to gather the other factors that affect the design and then look at the outcome. One might start by asking what we can actually design. We can find out by looking at jigs for building sails. Picture 16-1 shows a partly-built jig for building a fore sail for an A boat. The jig is made on a board by setting up sides to the shapes of the luff, the leech and the foot and fitting shaped profiles between them to support a ply lining. Pictures 16-2 and 16-3 show a completed jig for making the main sail for a Metre boat. Clearly no attempt is made to incorporate a twist.
Before jigs like this can be made we must know the plan of the sail, including the luff, and be able to draw the profiles of the sail. The jig for the Metre boat is typical. In the pictures the black lines mark the seams between the four panels which will be used. The lower panels are 6², 5², and 5² wide. The other lines show the profile of the single panel that forms the upper part of the sail. The camber at the foot is 8%, at 6² it is 10%, at 11² it is 11% and from 16² is 12%. Throughout, the maximum offset in the profile is at about 40% of the chord. Others may not agree with this design but that is not the point of showing the picture. The important piece of information from this jig is that, at the joint between the first two panels, where the change is greatest, the curvature needed to match the edge of one piece of the ply lining to the straight edge of the next is only about 1/16². This is very small and tells us that attempts at making concave sails will require careful work and the outcome may not be all we hope for when the sail stretches under load. Furthermore, if the pieces are joined with the popular double-sided sticky tape then any prolonged load on a seam will cause the tape to slip and, with such a tiny curvature in the joint, we really cannot afford any slip at all. If the joint is reinforced with stitching it will slip until the stitching is tight and still go out of shape. If the sail is rolled for storage it might as well have been flat to start with because it will end up flat. Indeed if no way can be found to solve the storage problem then the sail need not be shaped.
This all means that an important constraint is the storage of the sails that, if it cannot be met, means that we may as well make one-piece sails. One is forced to think of leaving the rig under tension (not the sails) and, for instance, suspending the mast horizontally to keep the sail in shape. Fortunately aluminium and carbon do not take a permanent set just because they have been under load for a long time.
It is also important to check the constraints imposed by the class. For instance, the class rules for Metre boats precludes any sophisticated sail design by limiting the number of joints and preventing these joints from being placed where they might affect the “belly” of the sail.
All this tells us that we only have control over the sail plan, the profiles and the shape of the luff for main sails. Most of the other refinements reflect our eternal optimism.
The sail plan.
The first decision to be taken is how to share the area available between the two sails. We have no choice for Metre boats but there is a limited choice for Marbleheads and A boats and a free choice for 36 inch boats. We know that it is the fore sail that is mainly responsible for the drive when the yacht is beating to windward and so it makes sense to have as large a fore sail as is practicable. There are constraints. For a Marblehead the distance from the deck to the upper mast attachment is limited to 80% of the total height. The distance from the mast to the stem is usually about 22² but, for a normal rig, room is needed for the counterweight and this sets a practical limit to the length of the foot of the fore sail. There is ample room for a swing rig because the area of the fore sail is limited by the need to ensure that the total force exerted on the sails by the wind acts behind the mast. This gives a lower value for the area of the fore sail of a swing rig than one might wish to have. Similar restrictions apply to A boats. The sails also have to fit into the spaces in the rig and allow room for bowsies etcetera to function to set the sail. These requirements lead to shaping of the sail plan at the head and foot. Where rules permit sails may be extended beyond the basic triangle by the addition of a roach supported by battens.
We have seen that the mast may be bent by adjustment of the back stay and the fore stay and that this gives us a mechanism by which the profile over the middle part of the sail can be altered. If we are to use this facility the luff of the sail must be cut to a suitable curve and this can be built into the jig. In all probability this is the single most important variable in our design.
We have seen in the Hele-Shaw pictures that sails must have camber and that camber needs to suit an angle of attack of about 30°. This is not just a matter of letting the luff angle be 30°. We need to look at the Hele-Shaw pictures as well. These suggest that the air can make rapid changes in direction to get under and over the mast and on to the sail so that the luff angle may need to be several degrees less than the angle of attack. Many skippers claim that the shape should have its maximum offset at a point nearer to the mast than would be the case for a circular arc. Not having prolonged access to a lake which has unobstructed flow of wind over it I cannot check this but it may well be correct and it would fit in with the flow patterns from the Hele-Shaw rig. There are other claims that the camber should be reduced in high winds and increased in calm conditions. This may also be correct but I cannot check it or find any basic explanation to support it. Generally starting with a sail camber of about 10% looks to be about right but if a profile with its maximum offset at, say 40%, is to be used the camber might need to be reduced.
We have also seen from open field tests that the flow of air over the concave side is more or less horizontal over the top two thirds of the sail but flows down obliquely near the foot to get under the sail. We might like to attempt to reduce this by letting the camber be 12% at say the one third point and then reduce the camber progressively towards the foot. Then the sail curves from head to foot as well as luff to leech. The sail will be hollow.
We still have to recognise that the very small curvatures needed between the panels of a sail to give it “belly” show us that only a tiny stretch in the completed sail can have a considerable effect on its shape. It may well be impossible to retain the designed shape under load. In truth, as we cannot with certainty make any special shape, we may as well make a jig with profiles other than arcs of circle, provide in the jig for a progressive change of camber, use mast bending for the final adjustment and accept that there will be an element of chance in the outcome. Of course experience in sail making will reduce the effect of chance.
The main sail
Once sail plans have been chosen the sails can be considered independently. Let us start with the main sail.
Before we can design this sail we have to decide just what it is that we are trying to do. We have seen that, at best, the wind we sail in will veer quite quickly through perhaps 10° in either direction and back again. These changes are too rapid for a control response and so, for preference the sails should cope with this without a loss of drive. We also know that the flow over the lower part of the sail is obliquely downwards and the pressure over the middle part of the sails would increase if this downward flow could be suppressed. We know that, for the main sail, the camber over the middle section can be increased as the sail swings out by suitably shaping the mast and luff but we do not know with any certainty that this is an advantage.
In my view the first thing to settle is the connection between the mast and the luff of the sail. We have two ways of dealing with the rapid wind shifts inherent in the air flow. They are the twist in the sail and the shape of the leading edge of the sail. The twist presents no special problem and is probably unavoidable with the arrangements that we must work with. The leading edge is another matter. All the experience of aeroplanes tells us that sharp leading edges are not desirable. Several people died to show us that wings with sharp leading edges suffer sudden and large changes in flow pattern (catastrophic switches in fact) at certain angles of attack. Aerofoils with round leading edges do not behave in this way and a great deal of attention has been paid to the profile of the first 10-20% of aerofoil sections. As always a compromise emerges between having a section with a large nose radius with a somewhat lower efficiency but very forgiving characteristics and sections with a small nose radius with higher efficiency and less forgiving behaviour. With a model sail we start off with an unavoidable and large radius, the mast. We must see how to use it for the best.
We could start by deciding what would be desirable. We know that sails can have attached flow over the convex side of the sail and we want to do nothing to disturb it Given the fact that we must end up with a single membrane it is a matter of how to blend the mast and the sail. We can do it either by shaping the mast or by shaping the sail, (using a pocket luff.). Figure 16-4a shows a practical cross-section for a mast and 16-4b a shape for a pocket luff. The mast could be grooved to accept the sail and the pocket luff simply fits round the mast. From the standpoint of fluid flow these are good shapes for our purpose. However there are snags with both. The shaped mast is useless if it does not rotate to maintain its alignment with the sail. For the mast to rotate we have to accept mechanical complexity, which can be overcome, but only at the expense of the facility to bend the mast. The pocket luff must slide round the mast if this profile is to be achieved at all points of sailing. This means that there is a friction force to interfere with the set of the sail on the mast. Further, no matter how the luff is cut, it cannot fit properly at all positions on the mast because, as we have seen, the mast will bend under the essential force in the fore stay. The result will be unwanted creases. The pocket luff is probably not an option for the main sail. We must either use a rotating mast or look to other systems
For preference the system should keep the leading edge of the sail radial to the mast. Our choice is now between a fixed grooved mast that could be round or profiled, a jackline or mast rings. Only the mast rings can satisfy our requirements. However the usual methods of using loops of fishing cord or wire rings hooked together at the ends are not really good enough because there is so little control of either the gap between the luff and the mast or of the position of the sail relative to the mast. The use of cord involves unnecessary friction. (The cord round a mast is just like a rope round a capstan which only works because of the friction.) The simple metal ring slides easily but tends to hang obliquely on the mast under its own weight. We can achieve this, solve the friction problem and get rid of the tendency to hang by spending a little more time on the rings. The best solution I can come up with is in Picture 16-5. The ring is made from soft brass wire of 0.024² diameter and, in order to make a joint, at one end five turns of the wire are wound round a drill of 0.025² diameter. The other end can then be threaded through these turns to form the equivalent of a tie-wrap. Eventually it can be secured by soldering. Figure 16-6 shows how it is fitted through an extra hole cut in the tape reinforcement of the luff. Once it has been shaped and fitted, polyester sewing thread can be used to bind it through the eyelet. It is easy to get a good sliding fit on the mast and the gap can be very small. The ring is always at right angles to the luff as a result of the binding. The result is a fair approximation to the arrangement in picture 16-4a.
The use of mast rings of any sort permits us to design main sails to make use of mast bending to alter the shape. This requires some way of deciding how much initial bend we need. The basic idea is to provide the sail with a curved luff and some shape which is thought to be desirable. When the curved luff is attached to the mast the back stay can be tightened to make the mast match the designed shape of the sail and then it will have the designed profile. From this condition the camber will decrease if the mast is bent more and increase if the mast is allowed to straighten. We need to have an idea of the movement required to change the camber. Graph 16-7 is for a circular arc and for a 10² chord. It shows the relationship between the camber expressed as a percentage and the movement of the clew to produce it in inches from the zero camber position. The movement will be proportional to the chord. At half way up the main sail for a metre boat the chord of the sail has a chord of just over 10² and for this profile we can use the graph directly. This shows that a movement of the clew of about 0.2² will change the camber from 8% to 12%. If we started with a bend in the mast of about 0.5² with the luff of the sail cut to match an increase in mast bend will reduce the camber and permit an increase in fore stay tension for heavy weather and vice versa.
So how should we draw the profiles? The most simple way is to use two arcs of circles which join tangentially as in Figure 16-8 which is drawn to scale for a profile of 10% camber with the maximum offset at 40% of the chord. It is helpful to know the radii and these are given in Graph 16-9 for a range of cambers for the same profile and for a 10² chord. For other chord lengths increase the radii in proportion.
The fore sail.
It is evident that for a model yacht the fore sail provides most of the drive when the yacht is beating. We must consider how to design it to maximise the drive that it generates.
We do not have as much control over the shape of the fore sail as we have over the main sail simply because, when it is mounted, its luff must be straight. It follows that if we want to the camber to change from foot to head we must do it by cutting the luff of the sail to a curve. On the face of it this seems to be straightforward but there are other considerations.
Somehow we must make it insensitive to the veering of the wind. Our consideration of the main sail tells us that we really need a rounded luff that will be less sensitive to the small and rapid wind shifts that are inherent in the wind. Even the rules for metre boat do not currently preclude the use of a pocket luff. A pocket luff over say a length of spiral cable tie fitted over the fore stay would give a sail which had a rounded luff and could be rolled. This adds a complexity which makes the manufacture of the pocket luff more difficult but, if the luff can be made, the seams can be placed to give the desired shape although with only two seams permitted on the fore sail of a metre boat the scope there is not great.
Where greater freedom is given the fore sail can be quite complex with shaping to reduce the downward flow and shaping to place the maximum offset nearer to the luff. A sail with say 5% camber at the foot changing progressively to 12% at about one third and then back to 10% for the upper part of the sail looks to be a good starting point.
 The gas turbine engine is a classic example of the application of this process. The range of size that now exists is quite astonishing as is the efficiency achieved by the large ones.
 This appears to be a serious problem for the users of high tech sails for full sized yachts. They have little choice but to fold them.
 The fore stay is always straight and even if the luff is cut to a curve we have no further control over its shape.
 This apparent disadvantage is offset by the better performance down wind where both sails are certain to be presented to the wind which is not the case for a conventional rig. It would still be better if we could increase the area of the fore sail.
 The roach is an area of sail which would be free to flutter unless it is supported by battens which extend well inside the basic triangle for support and then across the roach. Wind surfing sails are fully battened from luff to leech.
 On some full sized yachts the foot of the main sail has no camber.
 We might fit the foot of the sail with an end plate as is done in some aeronautical applications or with an inflated foot as in full sized sailing. We are not allowed an inflated end plate on models but flat booms are permitted and it may be that a boom mounted flat ways would be more effective in directing the flow than the frequently-seen boom on edge.
 On ocean racers the sails are made in jigs and mono-filaments are laid to a predetermined pattern and bonded to give a sail which will not undergo a serious change in shape under load.
 This is not the sort of thing that everyone will want to do but it only requires some ingenuity as the wire winds easily.
 The rig which is shown in Chapter 11 was originally fitted with rings which were designed to preserve the flow over the convex side of the sail by letting the sail set tangentially to the mast. When tested most of the tell tales near to the luff went through the gap and stayed there. This did not happen with the new rings.
 The movement is too small to justify calculating it for other likely profiles.