Yves Gélinas on Self-Steering

This article is reproduced from Sailing Canada No 50, May 1988. It was first published in French in Canada in La Revue maritime l'Escale No. 10 1985 and in France, in Loisirs Nautiques no. 185, 1987

When he began making plans to sail alone around the world, Yves Gélinas wanted more in a steering system than was offered. After a year of work he had a piece of equipment which matched his expectations

On May 9, 1983, on board Alberg 30 Jean-du-Sud, I completed a 28,000 mile single-handed circumnavigation from Saint-Malo, France, to Gaspe, Quebec, by way of the Southern Ocean and Cape Horn.  In the 282 sailing days of my voyage, I did not steer more than one hour.  Jean-du-Sud stayed on course with the aid of the self-steering system I had designed, through the seas and gales of the Roaring Forties and in all other strengths of wind, even under spinnaker, square before a very light breeze. If there was enough wind to keep the sails full, there was enough for the self-steerer. It even survived two knock-downs and a 360° capsize which dismasted Jean-du-Sud. Under jury rig, it was still steering!

I had been thinking about the design of a wind-operated self-steering system for as long as I had been cruising under sail: I have always considered that there were more interesting things to do than be stuck at the helm.

Sailing historians generally credit French marine artist Marin-Marie for the first solution to this problem, in 1939: the main rudder of his pinnace Arielle was lashed, while a small auxiliary rudder, controlled by a vane, kept the boat on course. With the aid of this system, Marin-Marie was the first single-hander to cross the Atlantic under power.

This auxiliary rudder self-steerer is still in use today (Hydrovane, Auto-Helm, RVG...). I tried it, and I found that its action is slow and its efficiency decreases as the seas increase. It is more difficult, but also more efficient to keep a yacht on course with its own rudder.


If the rudder is outboard, the problem can still be solved relatively easily: in the fifties, an unknown person thought of mounting a small trim-tab, like an airplane flap, on the trailing edge of a rudder. Controlled by an average-size windvane on the same axis, this flap supplies the energy required to move the rudder. It is the simplest of all systems. Bernard Moitessier has successfully used it on all of his boats.

When the rudder-stock passes through the hull, the problem is more complex; and Jean-du-Sud, like most other modern cruising yachts, happens to have such a rudder.

For the first single-handed trans-Atlantic race, in 1960, Francis Chichester replaced the mizzen of his yawl Gipsy-Moth with a big canvas windvane linked to the tiller with control lines. The force required to move the tiller called for a very large vane, which Chichester had to reef as the wind increased.

Blondie Hasler had a better idea. He made use of the energy produced by the movement of his folkboat Jester through the water. A vertical oar blade was planted in the water behind the boat. When a vane rotated this blade around a vertical axis, like a rudder, the flow of water pushed it sideways with a considerable force. Hasler used this force by connecting the blade to the tiller through a quadrant and control lines. He called this system "Servo-Pendulum". It produced a first generation of commercial self-steerers.

I believe it was the French engineer Marcel Gianoli who thought of mounting the vane (of the self-steering system for Eric Tabarly's 1986 OSTAR trimaran Pen-Duick IV) on an almost horizontal axis (instead of vertical). This increased the power and sensitivity of the windvane considerably and a second generation of servo-pendulum self-steerers was created. Some of them are still popular to-day (Aries, Monitor, Fleming, Sailomat...).

Since then, there had been no major improvement to windvane technology.  All of the existing wind-operated self-steering systems on the market use one or a combination of these solutions. At the time of preparing for my own circumnavigation, I had not seen a system which I considered capable of taking me around the world with pride or confidence. To me, they all seemed clumsy add-ons with far too much metal. I did not want to deface the work of art Carl Alberg had created by hanging a drilling rig on its transom.

In 1975, I started to work on a better self-steering system for Jean-du-Sud.  I was spending the summer in Martha's Vineyard waters and I used the small-boat carpentry shop of Peter and Robbie Eldridge to conduct my first experiments. I did not dream, then, about sailing alone around the world, but I already felt that Jean-du-Sud, my only home in the last few months, would take me far and needed a self-steering system I could trust. I had already built and used two experimental systems and I wanted to build a final one.

I decided that Jean-du-Sud's system would have a Hasler-type servo-pendulum and a Gianoli-type horizontal- axis windvane. It would be installed permanently and become an integral part of the boat.  Regardless of the strength of the wind or the state of the sea, I would not have to worry about its resistance or performance. It would be discreet and would not spoil the pleasing lines of my boat. It would be sensitive enough to steer on any point of sail as soon as there is enough wind to move the boat. Finally it would be as inexpensive and easily repaired as possible.

From then on, I always kept the problem in the back of my mind, even when I was not working at it actively. I can say now that I worked the equivalent of more than one year, on a full-time basis, either at the drawing board or at building and testing prototypes. If my application of the principles discovered by Hasler and Gianoli is better, it is probably not because I am more gifted, but because I worked at it longer.

Throughout the whole design period, I had this constant preoccupation with simplifying the device. By eliminating useless metal, I achieved the advantages of less weight, simpler operation and cheaper fabrication.

In order to install a wheel steering system in a yacht, you have to punch a hole through its cockpit sole. To really integrate Jean-du-Sud's self- steering system to the hull, I did not hesitate to drill a hole through its transom for a horizontal tube. Inside this tube,  another tube; near its aft end is welded a smaller vertical tube holding the stock of the servo-pendulum; at its forward end, inside the lazarette, a steering quadrant is bolted. I can't imagine a simpler or more robust installation.

The main problem to solve, in designing this type of self-steering system, is to transform the vertical movement of a connecting rod coming from a vane, to the rotary movement of a servo-pendulum blade. The existing systems use gears, heavy and expensive to manufacture, or rods and joints, lighter but more fragile.

The system I invented solves this problem with a single piece made with bent stainless steel rod: first two 90° elbows form a crank (4) that transforms the vertical movement of the rod to rotary, then in the form of a horizontal "Z" (5) . The central branch of the "Z" goes through a slot (6) cut through the stock of the servo-pendulum blade.

To set the course, the turret (2) is oriented so that the apparent wind hits the vane (I) on its edge. As long as the boat stays on course, the wind pressure is equal on both sides of the vane and it stays vertical. Linked with a crank and a connecting-rod assembly (3) to the bent rod, it maintains its "Z" part (5) in a vertical plane. The slot in the stock of the blade is thus kept in the fore-and-aft plane, so the servo-pendulum blade stays vertical.

As soon as the yacht wanders off course, the wind tilts the vane. (The axis of the vane is not quite horizontal, so this tilt is greater, yet remains proportional to the variation in course.) The crank and connecting-rod assembly (3) and the handle (4) rotate the "Z." Part (5) which is no longer in the plane of the boat's movement, but makes with it an angle also proportional to the course variation. As it passes through the slot (6), the "Z" rotates the stock, the blade is pushed sideways by the movement of  water until it is once again in the plane of the "Z". The tilt of the servo-pendulum and the rotation of the horizontal tube are thus proportional to the tilt of the vane and course variation. A quadrant bolted at the forward end of the horizontal tube, inside the lazarette (8), and connected to the tiller or wheel through control lines, turns the rudder to an angle proportional to the course variation.

I made an effort to analyze all stress-points and to conceive a very strong mechanism; yet one capable of coming apart if a load above normal is imposed.

The servo-pendulum blade is held locked to its stock by an elastic link that keeps two notches in its mount in contact with a pin on both sides of the stock (10). If the blade hits an obstacle, this assembly comes apart instead of breaking the blade or bending its stock.

I made two vanes: a large one, very light, for increased performance in light air and a small and stronger one, for total dependability in heavy weather. I change to the small vane when I take a first reef.

This third generation system offers a double integration : it is not only more integral to the hull, it also integrates all steering modes : if I want to use an electric autopilot, I simply remove the vane and connect the forward end of the rod, which I made to emerge in front of the quadrant, to a small electric autopilot, which drives the servo-pendulum. Since this autopilot supplies only the information (the servo-pendulum provides the power), the smallest autopilot on the market can steer even a large boat with a minimum expense of power.

For steering manually, this control rod can be driven through light lines that provide remote power steering.

Before I left for my voyage around the world, I had already put ashore the factory-installed wheel steering, the tiller being all I need for the few times I would need to steer manually

I must admit I was a little worried about possible wear of the rod as it passes through the slot and rubs stainless on stainless. This was the only criticism a mechanical engineer had made of my system. I was pleasantly surprised, after 28,000 miles, to find no apparent wear.