This article was first
published in "Catalyst" the journal of the Amateur Yacht Research Society
July 2000 volume 1, No.2
(Note: Throughout this article "propeller" refers to the underwater
driving device and "rotor" refers to the device in the air for collecting
energy from the wind).
Many yacht research projects are concerned with speed or breaking records,
and, although some useful things may be learned on the way, the main end
result is an entry in history books, and as a record is only of interest
until it is broken, the glory is but short-lived. My rotary-sailing research
is not concerned with record-breaking, (only maybe incidentally) but more
with addressing the age-old drawbacks and inefficiencies of sailing craft
which have resulted in their decline as a serious form of water-transport.
The fact has to be faced, that watercraft powered by wind are no more
than toys at present, and whilst they provide interesting challenges and
entertainment, their practicality as a serious form of transport is almost
There are two areas in which a sailing boat is not practical; its inability
to sail in a region of approximately 45 degrees either side of the direct
head-to-wind direction, and the impossibility of its making progress when
there is no wind to drive it.
The Rotary Sailing craft project is aimed at dealing with both of these
problems in the following way:
Firstly, a properly-designed Rotary Sailing craft does not have any
difficulty in sailing directly against the wind in the 90 degree "no go"
zone which is prohibited to ordinary sailing craft.
Secondly, a Rotary Sailing craft can also be viewed as an extremely
large floating wind-driven electricity generator, that can be used to generate
electrical power, which can be stored; and used to power the vessel when
there is no wind. If used in this way, power can be extracted from the
environment even when the craft is not in use, for instance, when moored,
so there is plenty of time for sufficient battery-charging.
Is it Impossible?
A word is required here about the "impossibility" of sailing directly
against the wind.
I hope those who fully understand the situation will bear with me if
I explain once more that it is possible, and, in fact, presents no special
I was quite sure that everyone understood this, but have been surprised
that there seems to be an inbuilt natural disbelief from some traditional
sailors. I will try to be brief, and will give the best example I know
to prove the case.
Consider a small sailing boat trying to get upwind and making short
45 degree tacks to windward.
Attached to the stern of the boat is a line about 1 mile long, the
other end of which is attached to a small dinghy in which a person rides.
As the sailing boat is so far away from the dinghy, the short tack
oscillations are completely damped out and all the person in the dinghy
experiences is a steady progression towards the eye of the wind.
Another way of explaining the situation, is a extract from an American
Internet news group posting, where controversy on the subject recently
took place. I cannot explain it any better:
"You are [to a third party] actually the first person so far
who has taken on board the chief difference between a sailboat and a turbine/water-propeller
boat, that, where a sailboat loses power as it heads up to windward, the
power output from a turbine remains constant, (or even increases slightly
as the apparent wind increases), the only adverse effect on the turbine/water-propeller
boat being an increase in wind-drag (agreed this will be quite substantial,
but when you consider the propulsion system is still working at full power…)"
"….there are no "theoretical" reasons why a turbine/water-propeller boat
cannot go directly to windward, however, many practical reasons there may
be, then we can leave Peter Worsley and his fellow members of AYRS to tinker
around overcoming the practical problems" - Roger Wollin.
Why try to optimise
upwind performance only?
I have concentrated almost exclusively on the direct upwind performance
of my experimental boats. This strategy may result in a less than
optimal performance in other directions, i.e. crosswind, and directly downwind.
But I feel that it does no harm to specialise in this, because the more
normal methods of wind propulsion can always be added to a vessel in addition
to the rotor whenever necessary. (I would probably consider "normal" to
mean wingsails and not traditional sails).
How it got started:
the "analogy" method.
When I started doing research into the subject, I had very little knowledge
of the previous history of rotary sailing, and so I proceeded by a process
of analogy to see if I could make a working model.
I looked at an average sailing dinghy, and checked the area of the
sails and the area of the centreboard, and worked out a ratio between these
I then transferred these findings to the rotary sailing craft model
in the following way:
My device in the air, the rotor, was deemed to be the equivalent of
the dinghy’s sails, and the device in the water, the propeller, was taken
as the equivalent of the keel or centreboard.
I ensured that the rotor area to propeller area were in the same ratio
as the sails/centreboard area.
I then made sure, by means of suitable gearing, that the rotor blades
moved through the air at the same velocity as the propeller blades moved
through the water.
In this way, I considered I would achieve the exact rotary equivalent
of the sailing dinghy, with each air-rotor blade performing a continuous
close-hauled tack, and each water-propeller blade performing the same function
as a boats keel, but in a constant way, by rotating.
The model worked perfectly, and almost leapt out of the small test
tank against the wind.
My approach is completely different from that of many before. I don’t
believe that best performance will be achieved by simply selecting the
most efficient land-based electricity turbine and putting it on a boat.
Its not as simple as that, there are some forces which affect a moving
base wind-turbine that do not apply to a fixed base land installation.
It seems to me that many previous designs have ignored these forces and
subsequently not achieved an optimum performance upwind.
I achieved some success with models almost as soon as I started and
my further research has been based on a step-by-step progression based
entirely on practical testing. The mathematical analyses and predictions
on rotary sailing I have seen in the past are very obscure and sometimes
incomprehensible. Armchair theorists delight in playing with figures, but
nothing is ever achieved!
My system use a trailing rotor which is mounted downwind of a vertical
axis pivot and the rotor is allowed to pivot freely where it likes, and
in this way aligns itself automatically to the wind. This method eliminates
the need for a wind-vane, the rotor assembly being behind the vertical
pivot axis and behaving the same way as a weathercock.
Although a drive system using bevel gears with a vertical shaft has
proved attractive to many in the past, I decided not to use this
system for several reasons.
Firstly, the torque of the vertical shaft would make the automatic
weathercocking action difficult or even impossible, and then the only way
to align the rotor to the wind would be manually, an extra complication,
and tiresome for the helmsman.
Secondly, its likely that bevel gears would absorb some power, and
you need every bit of power you can find to penetrate the wind!
Thirdly, the bevel gearing system would be expensive because it would
probably have to be specially made, and it may be heavy, too.
Fourthly, on such a system there is no easy way of changing the gear
ratios, as there would be when using pulleys and belts, or sprockets and
So, in the light of these drawbacks, I elected to use a pulley and belt
system, and, for the models, I managed to devise a virtually frictionless
and non-slip drive, using rubber-covered pulleys and string.
The general arrangement of the complete system is as follows:
(See diagram below).
The top pulley on the rotor assembly is positioned directly over the vertical
pivot point, the rotor-assembly is allowed to pivot 90 degrees each way.
The belt, or chain, descends to a fore-and-aft driveshaft in the boat to
which the bottom pulley is fixed. A belt or chain-drive can accommodate
this degree of movement, if the tension in the drive is adjusted automatically.
The driveshaft is then connected through another belt of chain to the
The rotor is not used for downwind sailing.
This general arrangement I have patented.
How the rotorhead pivots is better shown in the overhead view of the
various points of sailing below:
The above system refers to the two, man-carrying small catamarans I
have put together. For the models, a simpler system is used with one drive
belt which goes directly from the pulley on the rotorhead to the pulley
on the propeller shaft.
I will now describe the latest fullsize version, called "Twice Lucky"
- (no particular reason for this name, only that it happened to be painted
on the side of the hulls when I acquired them) (Click
here for picture).
A pair of Hawke Surfcat hulls are connected by their normal aluminium
tube structure, less trampoline, and two wooden fore-and-aft supports are
bolted between the front and rear transverse alloy tubes. The wooden supports
are boarded up top and bottom and they form a box which supports the transmission
shafts, bearings, rotor tower and prop-shaft, and of course the "driver".
The rotor tower is set well to the rear of the boat and the vertical axis
pivot of the rotorhead coincides with the CLR of the whole boat. The rotor
assembly, which is a certain distance behind the vertical axis pivot, is
balanced forward of the pivot by an arm with a counterweight, on any windward
course this arm always points directly to where the wind is coming from.
Transmission starts with the pulley on the front of the windshaft, (which
is directly over the vertical axis pivot). The belt drive is led downwards
behind the drivers’ seat to the intermediate shaft which projects forwards,
within the box (which the driver sits on) to a forward position where it
is led through a six-speed cycle gear system and then to the propeller
shaft which in turn, projects aft, through a universal joint to the
propeller shaft which is supported by an underwater bearing near the prop
and has the ability to fold up sideways for beaching purposes.
The earlier boat, (click here for
picture) had a slightly different system involving an underwater skeg
with a toothed belt inside with the propeller mounted on a small shaft
at the bottom. The skeg could be folded sideways for beaching . The
boat sailed well, but only had one gear-ratio, the present use of a six-speed
variable gear has great advantages because it is simple to test different
gear ratios, whereas previously, different pulleys, would have to have
been fitted to achieve this.
Ideas from different fields have been used on the present boat, main
ones being the use of aeromodelling experience for the rotor blades and
cycle technology for parts of the drive-train.
The six rotor blades are standard model aircraft wings using built-up
construction in balsa on a spruce mainspar. Each wing has about fifty different
parts including 32 ribs with sheet balsa covering. Each blade is covered
with plastic "solarspan" heat-shrink material and took about two weeks
to make. I probably would use foam with obeche veneer covering next time!
The pitch of the blades can be varied from a fully "feathered" position
to a maximum drive setting. The inspiration for the pitch-control system
came from that used on a fly-ball governor on a steam engine, and was made
by an engineer who specialises in model railways. Control input to the
pitch-control mechanism is by cycle "bowden" cable. Using a standard cycle
In practice, the blade setting is either full on or completely off.
Latest move in development is the construction of a stationery-adjustable
variable pitch underwater propeller, on which different pitch settings
can be tried easily. There is still some room for further optimisation,
although the performance of the boat is never going to be impressive with
such a small blade area (about 12 sq foot) which is used at the moment.
An interesting discussion would be whether to count the swept area of
the blades or just the area of the blades themselves. As the blades are
slow-rotating, it might not be so appropriate to use the swept area.
At the present time, work on the man-carrying version is shelved and
attention is being given to measuring the thrust of models directly against
the wind, similar to the "bollard pull" test used on outboard motors.
The plan is to use a simple data-logging system with a video camera recording
the pull and windspeed simultaneously. In this way I hope it will be possible
to produce some figures which will relate thrust to blade area/windspeed.
This kind of test is particularly appealing because it takes everything
into account and produces an answer of what is actually achievable in real
life, as opposed to hypothetical figures. It might be possible to extrapolate
a graph and predict the performance with much larger rotors (most likely
this would be on the pessimistic side as one would expect better efficiency
with larger scale).
All the foregoing refers to performance directly into the eye of the
Comments on the above are always welcome.
Return to main page