Out of the black sliced a strong, chill wind, blowing the Swan 46 downwind at seven knots. Bermuda was 200 miles ahead as Rhode Island quickly slipped more than 440 miles from the stern. Huddled behind the wheel, spray sounded like pebbles sprinkling the plastic of my foul weather hood pulled tightly over my head. Suddenly the companionway hatch flew open spilling light into the cockpit and hammered a pounding into my startled heart.
I had been so relaxed enjoying the ride – but now had to jump to the wheel.
The crewman plopped into the damp cockpit and asked how things were going. I made it look so easy to talk and casually hold a precise course. Actually the autopilot was moving me and I was cheating by using it. It was no longer my turn but I could not let the new watch take the helm and discover my secret. Fortunately we both agreed the jib needed a slight adjustment. With that opportunity I hit the standby button and slipped the sun cover over the controller.
I was the captain but I wanted everyone on board to be fully competent at hand steering and not be enticed by the autopilot. There would be plenty of time south of Bermuda to learn how to steer a boat to the Caribbean by merely pushing a few buttons. As on most passages, this boat would never steer a straighter course than when the rudder was under control of the autopilot. As a delivery captain I have used many Raymarine autopilots, and Raymarine is what my wife and I have chosen to steer our Valiant 40, Brick House, around the world.
THE SYSTEM
There are five main components to an autopilot: The course computer, controller, drive mechanism, fluxgate compass and the rudder position sensor.
CLICK TO VIEWRaymarine's S3G course computer (see photo) is the brain of the autopilot. The course computer receives its heading information from the remotely-mounted fluxgate compass. The fluxgate compass gives precise heading information despite the rocking and rolling of the vessel due to the Gyro (stabilization) technology used with the S3G computer system. The S3G also has what they call "AST," which is short for Advanced Steering Technology, with an "auto learn" ability. The unit actually gets to know the steering habits of your particular boat, and adjusts over time accordingly. In short, we chose the S3G because it is at the top in brainpower and can handle many functions when programmed to do so.
CLICK TO VIEWThe drive mechanism (see photo) is the muscle that moves the rudder. Initially I wanted a hydraulic drive as those units always worked flawlessly on the larger boats I delivered. However, Raymarine Pre-Sales support convinced us that a hydraulic unit in a remote area would not be user-repairable, probably not even technician repairable, and that it would involve shipping the part back to be repaired in a controlled environment.
The linear unit has parts that, by a savvy boat owner, could be swapped or repaired on board. We chose the Type 2 Long Linear Drive, for boats up to 44,000 pounds. Certainly oversized but everything on our boat is built to easily go the distance.
CLICK TO VIEWThe Controller (see photo) is the small box in the cockpit with buttons for controlling the direction of the boat. We chose the simplest controller with the least user options: the ST6002.
INSTALLATION
When I looked into the big void I had made by removing a propane locker at the back of the boat, I could not visualize in what direction the tiller arm would extend from the rudder shaft or where the linear drive would mount and how the two would meet and support themselves in mid air.
CLICK TO VIEWOur technician finally stepped aboard carrying a shiny bronze tiller arm (see photo). The tiller arm is a stubby, strong, bronze lever that bolts onto the rudderpost. I had previously measured the rudderpost with a dial caliper (see photo) and ordered the 10-inch arm our technician requested. Edson tiller arms are available in 10-inch and 15-inch lengths. Edson milled the collar of the tiller arm according to my measurements to fit exactly over the rudderpost, and they allow a .006-inch clamping action onto the rudderpost. This tight clamping to the rudderpost is necessary for optimum strength. If too tight, however, the clamp could bend. Too loose and the arm could work loose or drop.
CLICK TO VIEWNormally when ordering the tiller arm, you indicate not only the exact size of the rudderpost but also mark on the order form where Edson should cut the keyway and the size of the keyway. We had a difficult installation so we first had to lay out our installation then mark the collar of the tiller arm where Edson would mill the keyway. Edson is very accommodating; the one thing Edson can't help you with is if you do not have a very low personal displacement to length ratio (unlike our technician and myself). A husky person would have a very difficult time mousing around in all the tight lockers and recesses to complete the installation.
CLICK TO VIEW
If you intend on hiring a professional to do the installation, you should be well informed of each step that needs to be performed. As in any business an installer can make mistakes or simply cut corners, which could affect the ultimate operation of the autopilot.
Our technician slipped into the large hole and I handed him the tiller arm. He loosely clamped the tiller arm onto the rudderpost. Then I lowered to him the Raymarine linear drive unit. Balancing the tiller arm and linear drive into their rough positions, the layout and placement of parts took shape, and I could see what the steps of my project would be. He was the layout man; I was the labor (see photo).
CLICK TO VIEWBecause of the enlarged cockpit drains I installed on Brick House, there was no room to mount the tiller arm and drive assembly forward of the rudderpost and above the steering quadrant, which would normally be the most likely place to do so (see photo). The only place to install the assembly was aft of the rudderpost and below the steering quadrant. The steering quadrant is the big metal bracket bolted to the rudderpost, where the boat's steering cables attach. In some boats, the "quadrant" is not one quarter of a circle but a full 360-degree disk and is called a radial drive.
It could be possible to mount the autopilot directly to the steering quadrant as long as it is a bronze case quadrant. Never attach a drive unit directly to an aluminum radial or quadrant; this could cause the quadrant to fail. If you want your autopilot to be a truly independent back up steering system then use an Edson tiller arm.
ASSEMBLING THE PIECES
With the rudder centered, the tiller arm loosely clamped to the rudderpost and the push rod of the drive unit halfway extended, our technician set the drive unit perpendicular to the tiller arm and found where the base of the drive unit would be mounted to a bulkhead. This relationship also gave us the hole along the tiller arm for attachment of the drive.
CLICK TO VIEWAdditionally, when setting this up we had to plan for the rudder movement to be "limited by the steering system end stops before the push rod reaches its end stop. Failure to do this could damage the drive." (See photo.) Our technician left one inch of potential movement in the push rod in both directions before the steering hit its end stops. Now we were ready to mark on the collar the spot corresponding with the keyway on the rudderpost. This is where Edson would mill the keyway in the collar. All this set up left the tiller arm to be positioned slightly to port of the rudderpost rather than directly astern of the rudderpost. It does not matter where the tiller arm attaches as long as it is perpendicular to the drive when the steering wheel is in the centered position.
CLICK TO VIEWIf it is not possible to locate a tiller arm over a keyway, two pinbolts could be threaded in the collar and holes drilled into the rudderpost for the bolts to set into. The holes in the rudderpost should be as deep as the diameter as the bolts used. Drilling into stainless steel or Monel with a hand drill, while contorted over bone bruising hard spots, is not an easy proposition. The drill bit must turn slowly and be cooled with oil. Allowed to heat up it will further harden the metal and make it more difficult to drill into.
CLICK TO VIEWThe drive is not waterproof so it must be mounted in a dry location. The drive base can be mounted flat, sideways, upside down, it does not matter. The important thing is that the push rod of the drive must operate in the same plane of travel as the tiller arm. So it is very import, vertically, where the drive base mounts. If there is an error, it cannot be more than five degrees out of plane, otherwise the ball end fitting could bind with the tiller arm (see photo). The next important thing is for the drive to rest parallel with its mounting base (see photo). This is structurally the best way for the mounting surfaces to absorb the tremendous forces exerted on them.
There are the instructions then there is reality. The forces exerted by our installation would not be a sheering force but would be applied at an angle to the bulkhead. My job would be to beef up the mounting bulkhead so the drive will not be ripped out of its footing or rotate unexpected parts of the boat. Our technician had the easy part and went away to have Edson mill the keyway on the tiller arm. We are fortunate to have the Edson factory located nearby, though any machine shop can mill the keyway.
The existing bulkhead we are using for mounting the base of the linear drive is three-quarter-inch plywood, glassed to the hull and supported by adjoining bulkheads. To make sure the plywood does not flex and to further spread the loads to the adjoining structure, I sandwiched the plywood with quarter-inch aluminum plates.
Looking in the phone book under "Steel," I found a metal supply company that could sell the aluminum sheets cut to my rough dimensions. To make the plates fit exactly to the contours of the hull I had made cardboard templates. Tracing the templates onto the sheets of aluminum left a line to be cut with a jig saw. For fine fitting the aluminum, a disk sander ground down the projections for a perfect fit. Before installing, the aluminum was primed on all sides. This was done for cosmetics, as well as oxidation protection, since this was not anodized aluminum.
CLICK TO VIEWSetting the plates in position, I applied a heavy bead of 3M 5200 sealant to the perimeter of the aluminum. Although this is supposed to be a dry area, I wanted to make sure no water ever gets between the plywood and the aluminum. The sandwich was held together with clamps till holes were drilled and the stainless steel bolts were screwed down tight. When our technician saw the mounting bulkhead (see photo) his comment was, "Bullet proof!"
THE ELECTRONICS
Ideally, the fluxgate compass would be mounted in the boat where movement is the least. That would put it in the middle of the cabin floor near the mast. Since we can't be stumbling on the thing, the floor gets wet and there are bilge pumps and other wiring nearby, the search for the best mount expanded out from that spot.
All electrical wiring, stereo speakers, pumps, motors, keel bolts and ferrous metal must be at least three feet away from the compass. Suddenly our options were greatly reduced to two locations. We had the plate cabinet high in the galley near the centerline of the boat, or low in a forward hanging locker. We chose the hanging locker location simply because it was far easier to run concealed wiring from that point to the course computer. Although this is a good location, we now cannot use metal hangers in this locker.
The course computer was mounted near the rest of the electronics in the chart table area. It too needs to be kept at least three feet away from speakers, motors, pumps and high-current-carrying cables. These items could have a magnetic field that can influence the Gyro Sensor in the course computer. Cables from the rudder position sensor, fluxgate compass, controller and drive all connect to well-labeled and color-coded connections on the bottom of the course computer. This is a real plug-and-play operation.
The 6002 Control Head was set into the cockpit coaming, within easy reach of the helm. Factory-supplied, color-coded wires simply plugged into the back of the control head then snaked through the boat to plug into the bottom of the course computer.
Next came the wiring to power the linear drive. The power source begins at a new 40-amp circuit on the breaker panel. A radio frequency noise suppressing ferrite, supplied with the autopilot, was installed on the new power wire at the course computer. The linear drive consumes an average of 48 to 72 watts at 12 volts, or four to six amps. There are simple tables to determine the gage wire needed for the round trip length from the course computer to the linear drive.
CLICK TO VIEWWhenever in doubt go with the heavier wire. The only consequence for this is financial. Using too small of wire, the consequence could be a fiasco. An undersized wire may not carry the required amperage resulting in melted wires or energy starved, burned-out equipment. Near the linear drive, the cables were terminated at bolt terminals. This is where the smaller wires from the linear drive will connect. At these junctions (see photo), a moisture drip loop should be made in the wires to shed water away from the connections. Since these cables will carry a lot of electricity, they should be located at least three feet away from any VHF radios or radio cables. For SSB transmitters or their cables, the autopilot power cables should be no closer than seven feet. Sometimes this just is not practical, so the autopilot might have to be turned off when transmitting on the radio, or you could install an expensive and complex grounding system throughout the vessel.
RUDDER POSITION SENSOR
CLICK TO VIEWOnce the freshly milled tiller arm was back on the boat, I was able to install it, bolt the drive base to the bulkhead and complete all the mechanical and electrical connections. The only installation left to do was the rudder position indicator (see photo). Using plastic (PVC), I made a shelf for the rudder sensor to mount on and screwed the shelf to a bulkhead. The pivot point of the sensor has to align with the rudderpost, so the arm of the indicator will move parallel with the tiller arm. A threaded rod connects the indicator to the tiller arm. The electrical wire from the sensor plugs into the course computer.
After five days of work the installation was complete. We switched on the power and the most amazing thing happened, it all worked perfectly! Besides our technician, this job was successful because of the full, patient support from the people at Edson and Raymarine.
The boat was ready to move out of the slip and put the autopilot through the short commissioning process, which includes compass calibration and auto learn. We were ready to sail out of the marina, take two rights and head for Panama, 2,400 miles south.
A version of this story was previously published in Blue Water Sailing.