Self Steering

This article is one of two taken from a seminar on Self Steering and Emergency Rudders given by Rob Macfarlane to the Singlehanded Sailing Society in San Francisco. The purpose of the seminar was to inform and prepare entrants for the 2002 TransPac singlehanded race from San Francisco to Hanalei Bay, Kauai, in Hawaii - a distance of 2120 miles across the Pacific.

These articles are reproduced here by kind permission of the author, Rob Macfarlane. For more information, visit the Singlehanded Sailing Society's website.

Goal of self steering

To control the boats direction so you don't have to do that job. Therefore you have time to sleep, navigate, repair sails and gear. Might even have time to enjoy the trip. at a minimum - keep the boat sailing straight ahead for a short period of time so you can leave the helm and do something other than steer. Such as, say, toast a pop tart for dinner. at a maximum - sail the boat efficiently over an entire course, eg from San Francisco to Hanalei, without ever touching the helm .

Requirements to achieve this

  • Correct autopilot for boat size and handling characteristics.
  • Ability to keep autopilot or windvane functioning (spares, replacements).
  • Most importantly - you have to trust your self steering completely.

Knock-on effects

Self steering is a system, not just a bolt on box. Using a self-steering gear will have follow-on effects in electrical consumption and battery charging, sail trim and sleeping. Get the self-steering device on the boat as early as possible and use it. This is doubly true for windvanes - they are not as easy to understand as an autopilot. Do NOT decide to learn how your windvane works on your first oceanic passage.

First Decision - autopilot, windvane or both

There are two types of self steering devices:
  1. autopilot(or autohelm) - compass driven electrical unit
  2. wind vane - wind and boat speed driven mechanical device

Cost considerations play a big part in self-steering purchase, and there is no "cheap" approach. An electric autopilot will require investing in a hi-output alternator, regulator/controller and substantial battery bank. A wind vane requires none of that but is not cheap nor will it do as well on a downwind TransPac race.

Above deck electric autopilots are the least expensive way to go, but anything mounted in the cockpit is a candidate for water and mechanical damage. Therefore you need multiple autopilots to make sure one is still running 2,200 miles later (plus another 2,500 miles if you sail the boat home). Not generally repairable, carry a lot of spares. Compass is usually an undampened fluxgate.

Below deck autopilots add cost, installation complexity and power requirements. However, they are installed down inside the transom, so are less likely to get wet and short out or get snagged by an errant spinnaker sheet and flipped overboard. Not generally repairable, carry at least one complete spare. Compass is a dampened fluxgate, some with addition of solid state rate gyros. A north-pointing gyro compass is not the same as a rate gyro.

A windvane is as expensive as a below deck autopilot and weighs more. The wind vane is well suited to trans-ocean sailing and works especially well upwind. Big advantages over an electric pilot are no electricity is needed to make them work, they are silent, and will keep the boat in good sail trim upwind or reaching. Possibly repairable, but unlikely to break in the first place. Carry spare parts kit, air blades, break away tubes, and don't forget to tie a really strong line to the water paddle.

Autopilot considerations

The compass is the heart of your basic electric autopilot, and as such look carefully at the compass design before you purchase a pilot. Note that you can also interface an autopilot to boat instrumentation (wind, heading and speed) and performance polar curves via an outboard computer (e.g., a laptop sitting in the nav station) and control the autopilot to drive the boat to it's polars.

Basic compass data

A compass card does not, in fact, turn at all - it stays pointed north. The boat turns beneath the compass with the effect (when you're standing in the cockpit looking at your compass) as if the compass has turned. When I say the compass turns, this means the card is turning relative to boat heading.

A compass is effected by tilt above or below horizontal, acceleration in any horizontal direction (but not vertical), and it's own mass and momentum. The mechanical compass still operates on the same basic method as the earliest ones, with "modernizations" aimed at improving bearings, damping the card with oil or alcohol, keeping the compass level via gimbals, and adjusting a compass to indicate magnetic north via adjustable magnets or ferrous masses (the large iron spheres you see next to ship compass binnacles).

To see the effects of tilt, take a simple Autohelm flat fluxgate compass and lay it flat on a table and read the direction it is pointing. Now tilt the compass by lifting one end of the case a couple inches above the table and take another reading. They won't be the same.

The first significant change to a compass was the gyroscope inertial guidance systems used on submarines. Here was a machine that would point consistently in one direction all the time and would remain level at all times (through conservation of momentum - a spinning wheel wants to remain spinning in it's original plane and resists angular changes).

Another huge change in compass technology was the fluxgate compass.

And lately, advancements in microgravity experiments yielded the solid state accelerometer and then the solid state rate-gyro. Consider these to be motion indicators in a computer chip.

The are four basic types of compasses used in autopilots:

Oil filled mechanical compass
Rarely used now, as manufacturers have switched to fluxgate compasses. Information from the compass was passed to the pilot via string wrapped around the compass card - when the compass card turned, the string was pulled and the pilot could "see" the movement of the string.

Flux gate compass
A fluxgate compass is a set of wire coils wrapped around a pair of ferrite bars. Applying a signal to one wire will result in a voltage in the other wire. The signal undergoes a phase shift due to the earth's magnetic field; the phase shift is indicative of the coil angle relative to magnetic north.

Oil-dampened fluxgate compass
A fluxgate compass in an oil bath, the oil is intended to dampen the core's swinging and provide a more stable reading.

Powered north-pointing gyroscopes
A powered, spinning gyroscope. As an alternative, KVH Azimuth Digital Gyro - about $3,000. Very inexpensive for a true gyro compass, very interesting - appears to be using a trio solid state rate gyros to determine tilt angles and mathematically back out errors induced by the tilt(s).

Things to learn with self-steering

Sail trim is critical to efficiency and reducing power consumption. You probably sail with too much sail up, therefore too much heel angle, therefore too much autopilot work. Reducing sail area will usually make the boat flatter, the autopilot happier and the boat faster.

The autopilot will steer better than you can, especially if you're down below making dinner.

Special Considerations

The TransPac is essentially a heavy air reach followed by light spell followed by the thousand mile downhill ride. You know exactly where you're going when you exit the Gate, and you want to minimize distance sailed to get there. An electric pilot will point the bow straight at Hanalei and then it's up to you to trim the sails as the wind shifts; you are sailing shortest distance at the expense of electricity and attention to the sails. In the opening windy two or three days of the race you had better be sure your diesel fuel is in good shape, the Racor filters are clean and you have a lot of spares. If you lose the motor due to clogged filters you will lose your autopilot too.

Ideally the autopilot should have adjustable compass gain and deadband. Deadband describes how far the boat will turn before the compass senses the turn, and gain refers to how rapidly the autopilot turns the rudder to get the boat back on course. Upwind set deadband high and the gain high - you want small corrections quickly. Downwind the deadband is low and the gain is low - you want large corrections executed slowly.

Alternatively, a wind vane will maintain excellent sail trim (provided you set the sails up that way to begin with) but you will alter course as the wind shifts. Wind vane will not sail shortest course but will keep the sails trimmed for you.

Spinnakering under autopilot

Some people are amazed that singlehanders are dumb enough to go to sleep with a spinnaker up and drawing. The truth is it's not that hard, you just don't sleep for very long periods. Autopilot will correct quickly (if it has a fast ram) and keep the boat on course while surfing down waves. It will not notice if the wind shifts unless you've connected it to wind instrumentation and the instruments are working. It just keeps you headed the same direction all the time. You need to pay attention to what the wind is up to and look out for squalls and associated wind shifts. One thing that gets mentioned by the Vendee and BOC folks is that autopilots have no fear - they will happily launch the boat down the face of a 40 foot wave. Fortunately, that's usually not a problem for the TransPac.

Wind vanes are problematic for boats that surf easily. A vane is working with apparent wind angle (AWA) and knows nothing about true wind angle or swell height or any of that stuff.. When you start down the wave on a surf the AWA moves forward as the boat accelerates and the wind vane will see this as a wind shift forward and reacts by turning downwind to restore the AWA. As you go faster, it turns downwind more. This can lead to catastrophe in the form of a round down - I did that in a matter of seconds when the boat went from 7 knots to 11 knots on a 3' wave in 20 knots of breeze of Los Angeles. Wind vane spun the boat into a perfectly executed highly professional round down followed by an amazing round up, then we rounded down again and that was the end of that spinnaker.

Hint - you can use bungee cord tied to the counterbalance weight of a wind vane to dampen the vanes sensitivity to wind shifts, and make the vane return to center course rapidly. Play around with it, you'll find out the nuances quickly. This is a great trick for running DDW with poled out headsails - something you might not want to do with an autopilot that is NOT connected to wind instrumentation.


  • Oversize below deck autopilot connected directly to the rudder stock independently of the quadrant.
  • Hi-output alternator with a large battery bank.
  • Lots of spare fuel filters.
  • A solid windvane:
  • Navik for smaller boats (less than 30')
  • Monitor for larger boats (more than 30')