.

A Review of Modern Sail Theory

By Arvel Gentry
Boeing Commercial Airplane Company

Proceedings of the Eleventh AIAA Symposium on the Aero/Hydronautics of Sailing

September 12, 1981
Seattle, Washington

Abstract
December 1999
Popular concepts as to how sails generate lift, and how two sails interact with each other are discussed in
light of modern aerodynamic research. Much of the old sail theory in the sailing references is shown to be
wrong. The origins of these old ideas are discussed and the new and correct explanations presented.
Applications of modern sail theory to practical sailing problems are discussed.

1. Introduction
explain lift. It is difficult to explain the generation of lift for
Bruce Banks and Dick Kenny in their book “Looking at laymen. The simplifications devised in attempts to do this
Sails” (Reference 1) state that “it is essential to anyone seemed logical but usually turned out to be wrong.
interested in sails, and indeed sailing, that the fundamen-
The inconsistencies in the explanations for the interaction
tal principles are thoroughly understood.” They then
between two sails (the slot effect) also bothered me.
follow with several pages of sail theory and repeat basic
However, in this case, even the technical aerodynamics
ideas that have appeared in the sailing literature for many
literature seemed to be in trouble.
years. Unfortunately, their explanations of the most
important aspects of sail theory (how a sail gives lift, the At that time I worked for a world renowned aerodynami-
interaction between the jib and main, the slot effect) are cist, A.M.O. Smith. His group was doing advanced
completely wrong. research in multiple airfoil theory (wings with flaps and
slats). I was fortunate enough to learn from these people,
Virtually all of the sailing references contain similar, but
have access to their digital computer programs and flow
erroneous explanations when they discuss the aerody-
simulation equipment and be able to try some of the new
namics of sails. References 1 through 5 are typical. These
ideas out on sails.
books and magazine articles were written by recognized
sailing authorities, class champions, Olympic sailors, and Eventually, the pieces began to fit together. The new
famous sailmakers. How could these fundamental ideas multiple airfoil theories as applied to sails made all of the
on sailing be wrong, and how could they persist for so existing literature on the interaction between the jib and
long? All of these people are certainly excellent sailors. mainsail (the slot effect) obsolete. I presented the results of
They have learned from practical experience what it takes my research at the 3rd AIAA Ancient Interface Symposium
to make a boat go fast. But when they talk or write about in Redondo Beach, California in 1971 (Reference 6).
the aerodynamics of sails, they get into trouble. Reference 7 contains some of the results for aircraft wings
as prepared by A.M.O. Smith and his group.
As a research aerodynamicist, I was dumbfounded when I
first started reading the sailing literature as a beginning However, the Ancient Interface Proceedings do not really
sailor and saw what a confused state sail theory was in. reach the average sailor. Magazine articles and books still
Even the basic explanations of how a sail generates lift repeated the old theories. I was able to interest an editor at
were wrong. This I could understand, since even the SAIL Magazine, Chip Mason, and he printed an extended
popular aviation books were wrong in their attempts to version of my material in a series of articles starting in April

C 1981 Arvel Gentry

.

1973. The23
Ranger entire set of articles later appeared in the book
Newsletter defend the technical aspects of my theories. I had not
“The Best of SAIL Trim” (8). The approach in these articles expected Mr. Barrett's type of reaction. However, I cer-
was to expand upon the technical information presented tainly found one of his points to be true, “Few successful
in the original AIAA paper, but the approach was still quite racing skippers pay much attention to the scientific
engineering oriented. At least, my ideas were now avail- articles.” One bit of evidence of this was that new books by
able to the general sailing public, although somewhat these same people kept appearing with the same old and
obscured with pressure distribution plots that were completely wrong explanations of how sails work (10-13).
probably only understood by an engineer. Obviously, not everyone shared my interest in knowing
the correct explanations for how their sails really did work.
And how were these new theories received? Stephen
Haarstick, of Haarstick Sailmakers, requested permission Finally, in 1979, C. A. Marchaj published his new book
to make copies for use by people in his loft, and stated that “Aero-Hydrodynamics of Sailing” (14). Marchaj uses much
“they are by far the best articles written on the subject for of the material from my SAIL Magazine articles, and also
laymen to read.” Similar requests were received from as far states that “many problems concerning the interference
away as Australia. There was other evidence that some between a mainsail and a jib were clarified by A. Gentry
people did read the articles and found something of who explained correctly, for the first time, the jib-mainsail
practical use. In one article I discussed the leading edge interaction effect.”
separation bubble phenomena (8), and proposed an array
With reference to the old theories, Marchaj states that “All
of short tufts starting right at the luff as an aid to windward
one can say in defense of sailing theoreticians is that these
sailing. This system soon began to appear on a few local
misconceptions concerning jib-mainsail interactions were
boats. Although I did not expect that many hotshot sailors
originally derived from the most respected and time
would pick up the idea, my mini-tufts later appeared on a
honored, authoritative aerodynamic theories and faith-
world quarter-ton champion, in the Congressional Cup,
fully reflected the state of affairs in this field.”
on a maxi-boat, and on an America's Cup boat. However,
in the sail theory area, most sailing “authorities” still stuck As you have no doubt recognized by now, this is not the
with the old ideas. usual technical paper. The approach is to review the old
theories, point out where they are wrong, and to contrast
In one case, the reaction was rather negative. After only my
them with modern accepted sail theory. The format is a bit
very first article in the SAIL Magazine series, Peter Barrett
different also. Each part of the text goes with the figure
of Yacht Racing Magazine (Olympic sailor, sailmaker, self-
printed immediately to the right. In many cases the
described as being trained “as an engineer specializing in
discussion is quite brief, and technical terms and engineer-
fluid mechanics”) had rather strong statements about my
ing plots are avoided when possible. If you want the
articles (9). He stated “that future articles will do little if
technical details of modern sail theory, then read my
anything to improve directly the performance of either a
articles in the “The Best of SAIL Trim” (8), or Marchaj's new
given class of sailboat or a reader. In fact, by implying that a
book (14).
major error in everyone's thinking is about to be corrected
(and thus we will all, of course, be able to better utilize this However, if like Mr. Barrett (9), you feel “that the average
airflow and race more successfully), I believe that a sailor, and indeed any serious racing sailor, will do himself
disservice is being done the reader.” far more harm than good by attempting to understand the
theoretical streamline flow as defined by advanced fluid
I was at first shocked and angered by Mr. Barrett's com-
mechanics,” then I suggest that you skip the rest of this
ments, especially since they were made even before he had
paper and go watch TV.
a chance to read the complete series. I was well prepared to

-2-
.

2. Generation of Lift -- Old Theories

Figure 1. Stephen Colgate states that (15), “Because of the

Figure 1
curvature of the upper surface of the wing, the air passing
over that side has to travel a greater distance than that
passing under the wing. Since it has to go farther, it has to
go faster in order to reach the trailing edge at the same time
as the air flowing past the underside of the wing.” This is
one of the oldest misconceptions about the generation of
lift and seems to come from popular aviation books (16 for
example). However, there is no physical law of
aerodynamics that requires the air to flow as this theory
states. You will not find this theory in any modern
aerodynamics text book. In fact, exact calculations show
that an air particle flowing over the upper surface gets to
the trailing edge long before its brother particle that
traveled by the lower surface route.

Figure 2. Some books (4) state that the density of the air is Figure 2
different on the two sides of an airfoil. This idea is
completely wrong. Density is defined as the weight per
unit volume (i.e., pounds per cubic foot). The density of air
does not change as it flows around a sail. The forces on a
sail are caused by other factors. The air would have to be
blowing at jet-transport cruise speeds before density
changes became important.

Figure 3. Bernoulli's principle is frequently cited in sailing

Figure 3
references to help explain the generation of lift.
Streamlines are drawn about an airfoil such as a sail. Then
the Bernoulli principle is used to explain why the air
travels faster where the streamlines are close together and
slower where they are far apart. Bernoulli's principle states
that one half the velocity squared plus the pressure
divided by the density is constant along any stream tube of
air. When the stream tube gets smaller, the air travels faster
and the pressure goes down. This equation is certainly
true. However, we first must know the stream tube shape.
The sailing literature is filled with streamlines drawn by
people unfamiliar with the laws of aerodynamics. As a
result, the explanations that go with the pictures may seem
correct, but in truth, they are usually completely wrong.
The streamlines drawn in any reference quickly tell you
how much, or rather how little, the author knows about
aerodynamics.
-3-
.

Figure 4. The first attempts by scientists to calculate the lift Figure 4

of an airfoil used what we now call Newtonian impact
theory. It was thought that the individual particles of air hit
the surface and bounced off, and thus transmitted energy
to the wing in an equal-and-opposite energy exchange
manner. However, the lift that they calculated with this
method was not very much, and the drag was high.
Clearly, man would never fly, but how did the birds do it
and why did small model gliders fly? These early scientists
were treating air as a lot of small particles hitting the airfoil
and not as a fluid flowing past the object. Some sailing
books (17) still try to use this old but incorrect idea. In
normal windward sailing, the air right at the sail surface is
always flowing parallel to the sail. The shape and reaction
of the sail is a result of the airflow and the surface pres-
sures, and not as a result of particles hitting the sail like
grains of sand. The impact effect is not useful until the
airfoil is traveling at Space Shuttle reentry speeds.

3. Generation of Lift -- New Theories

Figure 5

Figure 5. It is not a simple task to explain how a sail or the

wing of an airplane generates lift. Much of what is in the
literature is an attempt to simplify a complex subject so that
the layman can understand. Unfortunately, these simplifi-
cations are usually not correct. The explanation presented
in this section sticks with modern lift theory as it is found in
aerodynamic text books, but without the advanced
mathematics usually found there. An experiment that you
can do at home will help in understanding these concepts.
The generation of lift baffled scientists many years ago. It is
no wonder that the average sailor today still does not
understand the generation of lift.

The correct theory for lift must hold for all conditions. It
must work for regular airfoils like the wing of an airplane,
for thin cambered surfaces such as our sails, and even for
flat thin surfaces such as a flying barn door. I will start with
the flat barn door airfoil since it helps demonstrate the
important principles in a step-by-step manner. Two-
dimensional airfoils will be used since the basic principles
are easier to understand. The same concepts, plus some
additional complicating factors, apply to three-
dimensional sails.
-4-
.

Figure 6. If we were able to wave a magic wand and cause

Figure 6
the air to have absolutely no viscosity, we would see some
very unusual things happen. We would find that an airfoil,
sail, wing, car, etc., would have zero drag. Unfortunately,
we would also have no lift. In real life we don't have the
magical power to check this out, but we can with our
modern computer programs. Our “zero viscosity” air
would flow about a flat plate airfoil as shown in the Figure
6. Note the symmetry of the flow. This causes all forces to
cancel each other out, and give no drag, and no lift.

Figure 7. Note that the lower surface flow is able to turn the Figure 7
corner at the trailing edge and flow upstream a little until it
meets the air on the upper surface. Even the air near the
leading edge does some strange things. Some of it turns
upstream and around the leading edge to the upper
surface. The dividing lines between the upper and lower
surface flows are called stagnation streamlines (marked S).
In this inviscid model of the flow, the air right at the surface
“slips” by the surface.

Figure 8. However, air does have viscosity and because of Figure 8

viscosity, the stream of air right at the surface will cling to it
and not move at all relative to the surface. Figure 8 shows
this phenomenon. A very short distance away, the air is
moving relative to the surface. Once we get away from the
airfoil surface, the viscosity of the air becomes
unimportant again. The portion of the flow very near the
airfoil surface is called the boundary layer. The change in
the flow speed relative to the surface in the boundary layer
is called the shear layer. It is this layer that causes skin
friction drag on the sails, hull and keel.

Figure 9. If we suddenly turn the viscosity on, the flow Figure 9

about our flat-plate airfoil will begin to change. The
viscous boundary layer will start to form on the surface. In
our magical inviscid model (Figure 8) the flow turned
around the sharp trailing edge with a very high velocity
right at the trailing edge point. Now with viscosity having
an effect, the flow is not able to make this complete turn
around the trailing edge. It at first tries to make the turn but
then separates from the surface. This initial attempt to turn
around the trailing edge causes a swirl of air to form. This
starting vortex is swept downstream with the flow.

-5-
.

Figure 10. With the air no longer turning the trailing edge
Figure 10
and flowing a short distance upstream, the upper surface
flow now continues on toward the trailing edge to fill this
void. With the upper surface air now continuing to the
trailing edge, the leading edge flow adjusts itself so that
more air is swept around to the upper surface so that there
are no voids in the flow. This is shown by the shift in the
stagnation streamline on the lower surface toward the rear
of the airfoil. In fact, the entire flow about the airfoil adjusts
itself so that much more air is flowing on the upper or lee
side of the airfoil. After we turned on the viscosity with our
magic wand, the air quickly adjusted itself so that both the
upper and lower surface flows stream smoothly off the
trailing edge parallel to each other (and at the same
speeds). In aerodynamic jargon we call this trailing edge
flow phenomenon the Kutta condition, after the man who
discovered it.

Figure 11. The airflow about the flat plate is now no longer
Figure 11
symmetrical. The summation of the forces around the
airfoil now will give a resultant force perpendicular to the
flow direction. This force we call lift. Unfortunately, the
scrubbing action of the boundary layer on the airfoil also
causes drag. Now, however, the lift is substantial (much
higher than the “impact theory” previously mentioned).
Birds and airplanes can now fly, and we have a force to
push our boats through the water! Without the viscosity of
the air, we would have never gotten off the ground or
away from the dock. In this figure we do have a problem
with the air trying to flow around the sharp leading edge.
That can be solved by bending the airfoil shape downward
so that the leading edge meets the incoming flow.

Figure 12. The flow about an airfoil really consists of two

Figure 12
separate flows added together. One of these is what we call
the non-lifting flow as illustrated in Figure 8. The other is a
circulatory flow about the airfoil that is necessary to give
smooth flow off the trailing edge. These two flows add
together just as our boatspeed and true wind add together
to give the apparent wind that we feel on the boat. The
circulation flow is not an imaginary flow, or a
mathematical trick. It is real, and can be visualized using
the experiment described on the next page.

-6-
.

4. The Bathtub Experiment

Figure 13. If you have trouble understanding the circula-

Figure 13
tion idea, try this experiment and see for yourself. Fill your
bathtub with about two inches of water and let it set so that
it is not moving. Now find something to sprinkle over the
entire surface of the water so that you can better see the
movement of the water during the experiment. Fine
sawdust, talcum powder, or even pepper will work. We
now need an airfoil. A four by six inch piece of stiff waxed
paper cut from a milk carton makes a good airfoil. Bend the
airfoil slightly so that it has about half the camber of a sail.
Very carefully place the airfoil on the centerline of the
bathtub as shown in the figure. The leading edge pointing
toward the left should be slightly higher than the trailing
edge (about a half inch) to give the airfoil what we call the
angle of attack. Again let the water settle down.

Figure 14. Now, grasp the airfoil carefully so as not to

Figure 14
disturb the water. Start moving the airfoil down the
centerline of the tub toward the left end. Watch what
happens near the trailing edge of the airfoil as you first
start the movement. The flow will at first start to make the
turn around the trailing edge, then separate to form the
starting vortex described previously. The starting vortex
will stay at its starting position as we move the airfoil
toward the left end of the tub. This is illustrated in the
sketch. In this experiment in water, we have to keep the
airfoil camber and angle of attack small in order to avoid
excessive flow separation from the airfoil. However, all of
the phenomena observed in the water experiment also
happen in air.

Figure 15. As the airfoil nears the center of the tub switch Figure 15
your attention to the flow in front of and around the airfoil.
Note that the flow out in front of the airfoil somehow
“knows” that the airfoil is coming, and starts changing its
position to flow around the airfoil, even before it arrives. If
the airfoil is being pulled precisely down the centerline of
the tub, you will note some of the water in front of, and
below the airfoil will actually end up flowing over the top
of the airfoil. The upward flow out in front of the airfoil is
known as upwash. You may have to repeat this exercise
several times and concentrate on a different part of the
flow field each time.

-7-
.

Figure 16. Now comes the key part of this experiment.

When the airfoil gets within about one foot of the left end Figure 17
of the tub, suddenly lift it completely out of the water.
What you have done by removing the airfoil is to remove
one of the components of the two flow fields about the
shape (this is like stopping a boat to measure the true
wind). When you remove the airfoil all we have left is the
rotational flows that are caused by movement of the airfoil.
At the right end of the tub we see the starting vortex still
rotating in a counterclockwise direction. At the left end of
the tub we see a larger clockwise-spinning vortex. This is
the “circulation” flow about the airfoil and is responsible
for generating the lift. The forward motion of the airfoil
and the circulation field are added together to give the
final flow, just as adding the boat speed and true wind
vectors give the apparent wind we see afloat.

5. Flow Separation
Figure 17. The flow separation phenomena are very Figure 17
important to the sailor. Few sailors, however, understand
what causes separation. The current belief in the “venturi”
slot effect is evidence of this. The photo at the right shows
an airfoil in the fully separated stalled condition. Because
of the stall, the airfoil has less lift and more drag. This photo
was taken using a water channel (a fancier version of our
bathtub). The camera moved with the airfoil and a long
exposure was used to record the streaks made by the
aluminum powder on the surface.

Figured 18. Flow separation is a viscous effect. It occurs

Figure 18
when the boundary layer is no longer able to stay attached
to the surface. The ability of the boundary layer to stay
attached to the surface depends upon the local flow
conditions and what has happened to the boundary layer
previously. When the flow is accelerating along the
surface, the pressure is decreasing. The boundary layer
likes this kind of flow (it's like running downhill). When
the flow speed is decreasing, the pressure goes up. This is
called an adverse pressure gradient (the boundary layer is
flowing against the increase in pressure). When the
increase in pressure is too great, the flow will separate and
become rather chaotic and unsteady.

-8-
.

6. The Leading Edge Separation Bubble

Figure 19. Several years ago while doing some research on

Figure 19
rather thin wings for a fighter airplane, I was faced with
what is known as the leading edge separation bubble
phenomenon. As the airfoil angle of attack was increased,
the flow separated right at the leading edge, but then soon
reattached to the upper wing surface. As the angle of
attack was increased more the separation bubble grew in
length. Finally, when the flow could no longer reattach, the
separation bubble would burst and spread over the whole
airfoil causing a complete stall. It suddenly dawned on me
that this same thing must happen on the luff of a jib. That
Figure 20
weekend I taped 500 small telltales on my jib to check this
out (we call them tufts in the aerodynamics business).

Figure 20. Just as the theory predicted, I found a small

separation bubble all along the luff when the boat was
sailed slightly off the wind. Sailing farther off the wind
caused the bubble to be wider. Finally, the bubble would
burst and the entire sail would be stalled. I searched the
sailing literature but could not find any references to the
leading edge separation bubble phenomenon. The leading
edge bubble is discussed in more detail in my SAIL
Magazine articles (see pages 97 and 254 of reference 8) and
at great length in Marchaj's new book (14).

Figure 21. Most sailors use a yarn telltale located about 12

to 18 inches from the luff on the jib to tell when the sail Figure 21
stalls. However, this only tells you when the sail is stalled.
You have already goofed, and gotten too far off the wind.
But what if you put a series of very short tufts starting right
at the luff and continuing to where you would normally
put the standard long yarn telltale? I tried this idea and it
turned out to be a tremendous help in keeping the boat at
the best angle to the wind for windward sailing. Not only
could you tell where you were between the luffing and the
stalling condition, but you could also tell how rapidly you
were changing from one condition to another. If you need
more drive to regain lost speed, you simply bear off slightly
so the first two or three tufts twirl, until the speed is back
up. I found that even the beginning sailor could be quickly
taught to sail to windward reasonably well with this new
tuft system.

-9-
.

7. Characteristics of the Single Airfoil

Figure 22. The figure at the right shows the flow stream- Figure 22
lines about a sail that has been adjusted so that the stagna-
tion streamline, S, flows directly into the leading edge.
Note that at the left side of this drawing the streamlines are
angled up slightly (upwash). You have to be several airfoil
lengths upstream, or downstream before the airflow is at
the freestream undisturbed condition. In fact, a careful
examination of the complete flow field would indicate that
the level of the stagnation streamline far downstream of
the airfoil is exactly the same as it was far upstream. These
streamlines were calculated by a computer program with
boundary layer thickness and separation effects turned
off. With modern computer programs we are able to turn
off the viscous effects (but still maintain the realistic flow
condition at the trailing edge, the Kutta condition). With
this kind of trickery provided by the program we are able
to isolate and study the various factors that influence the
flow.
Figure 23
Figure 23. This figure shows the same sail at a broader
sailing angle (35 degrees). Again, the program did not
include boundary layer or separation effects. With the
higher angle, more air passes to the upper (or lee) side of
the airfoil. We see this because the stagnation streamline, S,
starts and ends at a much lower level than it did in figure
20. The airfoil is able to generate more lift. Note, however,
that the stagnation streamline hits the airfoil on the lower
(windward) side near the leading edge. This will cause
some problems when we consider the viscous effects.

Figure 24. The calculated pressure distribution around the

sail for the 25 degree boat angle condition is shown in this
figure. The arrows pointing away from the upper surface Figure 24
represent pressures lower than atmospheric pressure
(“suction” pressures). Long “suction” arrows also repre-
sent higher velocities in the flow. The arrows pointing
toward the lower surface represent pressures that are
higher than atmospheric pressure, and the slowest
velocities. It is this pressure difference on the two sides of
the sail that give us lift. We have the highest velocities on
the sail where the upper suction arrows are the longest.
The velocities are the lowest where the lower arrows
pointing at the airfoil are the longest.
-10-
.

Figure 25. The pressures around the sail at the 35 degree

Figure 25
sailing condition are shown in this figure. Note the very
long (high suction, high velocity) arrows on the lee or
upper side of the sail near the leading edge. The velocities
are high here because the stagnation streamline is slightly
around on the windward side of the airfoil. The air
accelerates very rapidly as it goes around the leading edge
from the lower side. The velocity then decreases as it
approaches the trailing edge. The pressures are the lowest
right near the leading edge and then increase as the flow
approaches the trailing edge and slows down. The rapid
increase in pressure near the leading edge will probably
cause separation along the luff. If the increase in pressure is
not too rapid near the leading edge, only a small
separation bubble will form with the flow reattaching and
continuing on. If the angle is too wide and the pressure
increase too rapid, the entire leeside flow will separate and
the sail will be in the stalled condition.

Figure 26
Figure 26. Separation can also occur at the trailing edge of
an airfoil if the pressure increase is too great here. This will
happen if the leech is hooked to windward or if there is just
too much camber in the sail for the wind conditions. As we
will see later, the jib has a strong effect on the separation
characteristics of the mainsail and although it is not
realized by many, the mainsail also influences the
separation on the jib.

8. Sail Interaction -- Old Theories

Figure 27. The interaction between the jib and mainsail

Figure 27
(the slot effect) is probably one of the most misunderstood
aspects of sailing theory. Stephen Colgate's explanation for
the slot effect (15) is typical of what appears in most of the
sailing literature. “The jib funnels the air behind the main.
The funneling action tends to increase the speed of the air
flowing past the leeward side of the main.” The drawings,
such as that shown at the right, seem to substantiate this
theory. The jib does produce a flow header on the main,
but the effect of this on the mainsail velocities is exactly
opposite from this popular myth. In fact, if the primary
effect of the jib was to increase the velocities on the lee side
of the main, the effect would be to increase the possibilities
of separation on the mainsail, instead of decrease it as we
know happens in practice.
-11-
.

Figure 28. If the jib did cause a high speed jet of air on the Figure 28
lee side of the main, what would happen? First, the airflow
off of the leech of a single sail is very near the freestream
velocity well out in front of the sail. This means that the
pressure near the leech is near the atmospheric pressure. If
the jib caused higher speed flow near the leading edge of
the mainsail, we would have lower pressures there also. If
the pressures on the forward lee side of the mainsail are
lower, then there must be a more rapid increase in pressure
as the flow approaches the trailing edge. The flow would,
therefore, be more prone to separation. Also, higher speed
flow in the slot might help the main, but it would certainly
harm the jib since the pressure would also be acting on the
windward side of the jib.

Figure 29. A sail, being a flexible surface, reacts directly to Figure 29

the air pressures around it. When the pressure is higher on
the windward side than it is on the lee side, the sail will
take the familiar cambered shape and exert a useful force
on the boat. When the pressures are the same on both sides
of the sail, the sail will shake. If the jib creates a higher
airspeed on the lee side of the mainsail, then why does the
mainsail shake along the luff when the jib is sheeted a bit
too tight? If the smaller slot caused higher speed and
therefore lower pressures on the mainsail, then the
mainsail should have even higher pressure differences
between the two sides of the sail, and it couldn't possibly
shake. Only if the jib created lower velocities instead of
higher, and higher pressures closer to the windward side
values, could the mainsail react as we all know it does
when we oversheet the jib.

Figure 30. The figure at the right shows an effect that we Figure 30
often see on some boats. The trim of the two sails is such
that the mainsail actually has a reverse camber in the luff
area. Often, the “bubble” in the mainsail is quite stable. In
this situation, the pressures over a portion of the lee side of
the mainsail are actually higher than the pressures on the
windward side of the sail. The old slot effect theory cannot
explain this situation. We should by now begin to realize
that the jib must actually cause the velocities on the lee side
of the mainsail to be lower, instead of higher. We will now
see why this is the case.

-12-
.

9. Sail Interaction -- Modern Theories

Figure 31
Figure 31. One of the main reasons why the slot effect has
been misunderstood for so long is that we are unable to
separate the various aspects of the flow while we are
actually sailing. With modern computer programs we are
able to separate each of the effects and to study them in a
systematic manner. This allows us to understand the basic
physics of the flow. Later, we can study how the boundary
layer reacts to the various pressure distributions. These
computational experiments tell us for sure that the old slot
effect theories are wrong.

Figure 32. The figure at the right shows the streamlines

about a mainsail alone. The stagnation streamline that Figure 32
divides the flow that goes on each side of the sail is marked
as Sm. The streamline that passes through the headstay is

marked with an H. The H with the circle around it indi-

cates the headstay itself, and this is where we will later
position the leading edge of the jib. The air that passes
between the mast and the headstay is indicated as the “slot
air” at the left side of the figure. Note that the stagnation
streamline hits the mainsail on its windward surface. This
means that there will be very high velocities, and low
pressures, on the forward lee side of the mainsail. The
pressure will then rapidly increase along the lee side of the
mainsail. If viscous effects were present, the sail would
stall.

Figure 33. In this figure the jib is included. The solid

streamlines are for the flow with the jib present. The Figure 33
dotted streamlines are for the mainsail only. The stream-
line that went through the headstay in Figure 32, H, now
goes well to the lee side of the jib. The line that divides the
flow passing on the two sides of the jib, Sj, we call the jib
stagnation streamline. The stagnation streamline for the
mainsail, Sm, has been shifted (headed) by the jib. The slot

air is marked at the left side of the diagram. The amount of

air passing between the headstay and the leading edge of
the mainsail is now much less than was the case with the
mainsail alone. We can see this because the width of the
“slot air” stream tube marked in this figure is much smaller
than that in Figure 32. Much of the “slot air” in Figure 32
with the mainsail alone now passes to the lee side of the jib.

-13-
.

Figure 34. The effect of various sail angles is shown in the

four separate drawings at the right. The number at the left Figure 34
indicates the changes in the amount of slot air. In case B,
A. Basic settings
with the jib sheeted in 5 degrees tighter, the amount of slot
air is reduced by 60 percent. Look at the position of the
mainsail stagnation streamline in case B. It hits the main on
its lee side. Since we have taken most of the slot air and
caused it to flow on the lee side of the jib, we do not have
much left for the slot. Because this small amount of air
must spread out to fill the region between the jib and the
main, it will have to slow down (Bernoulli's principle
again). The pressure in this region will be high. In fact it
o
will probably be higher than the pressure on the wind- B. Jib in 5
ward side of the main. If our sail was flexible, it would
adjust its shape (carry a stable bubble), or certainly shake in
what we have always called backwinding.
-6O%
In case C with both the jib and mainsail sheeted in 5
degrees, the slot air is reduce by 30 percent. The stagnation
streamline for the jib now hits the jib well around on its
windward side. This means that there will be very high
velocities as the air negotiates the sharp turn around the
leading edge of the jib. The jib would probably immedi-
o
ately stall. To prevent this we would either have to let the C. Jib and main in 5
jib out some, or as we usually do, head the boat closer to the
wind so that the jib stagnation streamline again comes in at
the luff wire.

-3O%
In case D, with the jib at its basic position and the main
sheeted in 5 degrees, the flow in the slot is 20 percent more
than the basic setting. The stagnation streamline for the
main has shifted slightly around toward the windward
side. The lee side velocities on the forward part of the
mainsail would increase, and the pressure would go
o
down. The flow on the lee side might separate at the mast. D. Main in 5
To avoid this we would have to sheet the jib in tighter to kill
the higher velocities, or head the boat closer to the wind.

From these studies we see that the primary effect of the jib
+2O%
is to reduce the velocities on the lee side of the mainsail.
This reduction in velocity will give an increase in pressure,
a reduction in the pressure gradient on the sail, and
therefore, lessen the possibility of lee side separation.

-14-
.

Figure 35. In the bathtub experiment we learned about

Figure 35
circulation on a single sail. When we have two sails, we
have two circulation fields that must be added together.
Note in the figure at the right how the two circulation
fields tend to oppose each other in the slot between the jib
and the mainsail. This is further evidence that the jib really
causes a decrease in the velocities on the lee side of the
mainsail. The two circulation fields also add together to
cause more upwash out in front of the sails, and give our
boats better pointing ability.

Figure 36. One primary effect remains to be investigated,

and that is the effect of the mainsail on the jib. We all know Figure 36
what a powerful sail the jib is relative to the mainsail. This
is sometimes blamed on the separation from the mast, but
this is not the whole story. We have already seen how the
mainsail can cause an increase in the upwash at the leading
edge of the jib.

One effect not covered previously is the fact that the

trailing edge of the jib is located in a high speed region of
flow created by the mainsail. I have mentioned that the
airflow at the trailing edge of the mainsail must return to
conditions near the freestream conditions. This is not the
case with the jib. The airflow at the trailing edge of the jib
only has to return to the speed of the air present on the lee
side of the mainsail.

Figure 37. The single sail has a circulation field that, when Figure 37
added together with the non-lifting field, satisfies the
condition for smooth flow off of the trailing edge. When
two sails are present we have two circulation fields that
add together to create the total flow picture. We have seen
how the mainsail causes more air to flow on the lee side of
the jib. These higher velocities to lee of the jib, plus the fact
that the jib trailing edge flow returns to a higher velocity
than the mainsail, means that the jib will have a much
higher pressure difference across the fabric and higher lift.
The higher trailing edge velocities on the jib also mean that
there is less possibility of flow separation on the jib. All of
this can be turned into more forward drive, or to better
pointing ability.

-15-
.

the jib and main are used, and this contributes to the high
10. Summary of Sail Interaction efficiency of a jib.
4. The higher lee-surface velocities on the jib mean the jib
Although they are for the most part interdependent, the can be operated at higher angles of attack before the jib lee-
major jib-mainsail interaction effects will now be segre- side flow will separate and stall.
gated into the effect of the jib on the mainsail, and the 5. Because of all this, proper trim and shape of the mainsail
effect of the mainsail on the jib. significantly affects the efficiency of the overlapping jib.
Anything that causes a velocity reduction in the region of
10.1 Effects of the Jib on Mainsail the leech of the jib (such as some separation on the aft part
of the main) results in a lower driving force contributed by
1. The jib causes the stagnation point on the mainsail to the jib.
shift around toward the leading edge of the mast (the 6. The trim of the main significantly affects the pointing
header effect). ability of the boat for it directly influences the upwash that
2. As a result, the peak suction velocities on the forward lee approaches the luff of the jib.
side of the main are greatly reduced. Since the peak suction
velocities are reduced, the recovery adverse pressure 11. CONCLUSIONS
gradients also are reduced.
3. Because of reduced pressure gradients on the mainsail, The lift generation for a sail is shown to be caused by the
the possibility of the boundary layer separating and the combination of two flow fields. One is the flow that would
airfoil stalling is reduced. be present if the air had no viscosity. The other is a circula-
4. With the jib present, a mainsail can be operated effi- tion field about the airfoil. This circulation field is an
ciently at higher angles of attack without flow separation indirect result of the fact that air has viscosity. The initia-
and stalling than would be the case with just a mainsail tion of the circulation flow field occurs simultaneously
alone. This is caused by a reduction in velocities over the with the formation of a starting vortex off of the trailing
forward lee part of the mainsail rather than by a speed-up edge. The old theories that rely on impact theory, density
in the flow (which is the popular theory). changes in the air, or the different distances on the two
5. Much less air goes between the headstay and the mast sides of the airfoil are all wrong.
when the jib is placed in the flow with the main. The
circulations of the main and the jib tend to oppose and The interaction between the two sails is much more
cancel each other in the area between the two sails. The complicated than the old theories imply. The flow about
two circulation fields are in the same direction out in front the jib and mainsail are a result of two circulation fields
of the sails and on the lee side of the jib. This forces more air which add together. These two circulations oppose each
to flow over the lee side of the jib. other in the slot between the jib and mainsail. The primary
6. As the jib is sheeted closer to the main, there is a continu- effect of the jib is to slow down the flow on the lee side of
ing decrease in suction pressure on the lee side of the main the mainsail, reduce the pressure gradients and, therefore,
(exactly the opposite of the popular venturi myth). When prevent separation on the mainsail. If the effect of the flow
the pressures on the windward and leeward side of the was to speed up the air, as the old venturi theory states, it
mainsail are equal, there is no longer the pressure differ- would increase the possibility of separation, rather than
ence across the sail fabric to maintain the airfoil shape, and decrease it. The old venturi explanation for the slot effect is
the sail begins to luff. completely wrong.

10.2 Effects of the Mainsail on the Jib These modern theories have proven useful in developing
aids for sailing to windward (the mini-tufts). The sail
1. The upwash flow ahead of the mainsail causes the interaction ideas should help to understand the require-
stagnation point on the jib to be shifted around toward the ments in achieving optimum sail trim to windward. The
windward side of the sail, and the boat can be pointed sail interaction concepts should also be useful in selecting
closer to the wind without the jib stalling or luffing. and setting staysails to help the flow around the mainsail.
2. The leech of the jib is in a high-speed flow region created The proper understanding of the sail interaction played a
by the mainsail. The leech velocity on the jib is, therefore, large part in the author's research that led to an improved
higher than if the jib alone were used. mast section shape that was used on the America's Cup
3. Because of the higher leech velocity, velocities along the boats Courageous, Enterprise, and Freedom (18 & 19).
entire lee surface of the jib are greatly increased when both
-16-
.

12. References BIOGRAPHY

(1) Bruce Banks and Dick Kenny, LOOKING AT SAILS, Arvel Gentry is presently a research supervisor in the
SAIL Books, Inc., 1979. Aerodynamics Research Department at the Boeing
(2) Jeremy Howard-Williams, SAILS, John de Graff, Inc., Commercial Airplane Company. He has raced his own
1967. boats very successfully (primarily in Southern California),
(3) Wallace Ross, SAIL Power, Alfred A. Knopf, 1975. and has extensive crewing experience on larger ocean
(4) Alan Watts, WIND AND SAILING Boats, Quadrangle racing yachts. He has authored numerous magazine
Books, 1965. articles on sailing aerodynamics and sailboat performance.
(5) Tony Gibbs, PRACTICAL SAILING, Motor Boating & He has conducted research efforts in support of America's
Sailing Books, 1971. Cup projects, and designed the mast section shape used on
(6) Arvel E. Gentry, The Aerodynamics of Sail Courageous and Freedom. He has also developed special-
Interaction, Ancient Interface III, AIAA Symposium, ized sailboat performance recording equipment and
November 1971. served as a sailing performance test engineer on Jim
(7) A.M.O. Smith, High Lift Aerodynamics, AIAA Paper Kilroy's Kialoa maxi-boats.
No 74-939, Los Angeles, 1974
(8) Anthology, THE BEST OF SAIL TRIM, SAIL Books,
Inc., 1975.
(9) Peter Barett, Carte Blanche, Yacht Racing Magazine,
June 1973.
(10) Stephen Colgate, FUNDAMENTALS OF SAILING
CRUISING & RACING, W. W. Norton & Company,
Inc., 1978.
(11) Peter Cook & Barbara Webb, THE COMPLETE
BOOK OF SAILING, Doubleday & Company, Inc.,
1977.
(12) Bob Bond, THE HANDBOOK OF SAILING, New
York Alfred A. Knopf, 1980.
(13) Alan Watts, BASIC WINDCRAFT, Dodd, Mead &
Company N.Y., 1976.
(14) C.A. Marchaj, AERO-HYDRODYNAMICS OF
SAILING, Dodd, Mead & Company, 1979.
(15) Stephen Colgate, COLGATE'S BASIC SAILING
THEORY, Van Norstrand Reinhold Company, 1973.
(16) Richard L. Taylor, UNDERSTANDING FLYING,
Delacorte Press, 1977.
(17) Jeremy Howard-Williams, RACING DINGHY
SAILS, quadrangle Books, 1971.
(18) Arvel E. Gentry, Design of the Courageous Mast,
Yachting Magazine, February 1975.

-17-