Bookmark and Share Email this page Email Print this page Print

Installing a powerful belowdecks autopilot

Jan 1, 2003
From Ocean Navigator #127
January/February 2003
We have owned several boats with cockpit-type autopilots. While they were useful for motoring on the bay in calm weather or sailing in a light to moderate breeze, they lacked the power to be useful when we needed them most: in strong winds and quartering seas. It was a revelation to us, however, when we sailed a Caribbean charter boat with a belowdecks autopilot. In the trade winds and a sea on our beam fetching from Africa, the autopilot proved itself the best helmsman aboard. We no longer wondered why people actually give these things names. We returned home convinced that a proper autopilot was essential equipment for the latest Clairebuoyant project, a Pearson 424 sloop.

Elements of the autopilot the author installed on his Pearson 424 sloop. The drive unit is attached to an independent tiller arm, allowing the boat to be steered should the manual steering experience a malfunction.
   Image Credit: Quentin Kinderman photos

There are, of course, several manufacturers of suitable autopilots. Our decision, however, was biased in favor of Nexus, the manufacturer of the instrument system we had already installed. Pleased with the performance, not to mention the price, of this system on the prior Clairebuoyant, we even carried over a couple components from that boat to this one - a fluxgate compass and a remote control. Both of these pieces were essential for the autopilot, so we already had part of our planned autopilot system.

Image Credit: Quentin Kinderman photos

Elements of the autopilot the author installed on his Pearson 424 sloop. The drive unit is attached to an independent tiller arm, allowing the boat to be steered should the manual steering experience a malfunction.

A good autopilot is versatile and an all-weather convenience. The ability of the system we chose to interface with our integrated instrument system meant that it would steer the boat not only to compass, but also to apparent wind, like a wind vane, and to a GPS waypoint. It is also an important safety device. If installed redundant to the boat's steering system, the autopilot can steer the boat in the event that the steering system fails, making it a useful alternative to an emergency tiller. I also looked forward to tucking under the shelter of the dodger in foul weather, remote control in hand, rather than braving the elements at the wheel.

The other components we needed included 1) a servo unit, 2) the "brains" of the system, which would interface with the instrument system and send power and instruction to the drive unit, 3) a rudder-angle transmitter to provide rudder-angle information to the servo, 4) a drive unit to actually turn the rudder and 5) a tiller arm or other mechanism to connect the drive unit to the rudder shaft.

Image Credit: Quentin Kinderman photos

The control box was located in the electronics closet (the former hanging locker).

Two components, the servo and the rudder-angle transmitter, had to come from Nexus to be compatible with our instrument system. Shopping the Annapolis Boat Show, we found that ordering through a large discount mail-order marine supply (with boat show discount!) was the best deal.

Picking the drive unit required a bit more effort. The recommended unit was a linear hydraulic drive with nine inches of travel. This type of drive is essentially an electric pump attached to a hydraulic cylinder. While we favored this type of drive, having heard too many stories of failures with purely mechanical drive units, the nine inches of travel required that the tiller arm be short in order to permit full rudder angle. Instead, we chose an Octopus drive unit. This unit was similar in design but provided 12 inches of travel and mounted the pump independently of the cylinder. The design allowed us to use a longer tiller arm, reducing the force necessary to turn the rudder. This, in turn, reduces the stress on the components and perhaps even means it consumes less power. The remotely mounted pump made installation and service easier in the crowded aft lazarette.

 

 

The final element was the attachment of the linear drive to the rudder shaft. Many autopilot installations simply reinforce the steering quadrant and bolt the drive to the quadrant. While some autopilot manufacturers recommend this, since it provides a compact and simple installation, we decided against it for a couple reasons.

Image Credit: Quentin Kinderman photos

The hydraulic pump unit. It was mounted separately from the drive unit in a location that makes maintenance easier.

Done properly, this was not an inexpensive approach. We examined a quadrant modified by a local (and excellent) machine shop. A stainless-steel plate was fabricated to fit on the quadrant and attached to the quadrant with machine screws. This made the quadrant body virtually unbreakable. It did nothing, however, to reinforce the part of the quadrant that attached to the rudder shaft. Both the steering system and the autopilot would fail if the quadrant came loose of the rudder shaft or the shaft key sheared off. Since the autopilot drive is far more powerful than a human helmsman is, we reasoned that this approach actually increased the possibility of steering failure.

Instead we opted for a tiller arm independent of the wheel steering system for our autopilot. Several companies offer suitable tiller arms. Fortunately, Clairebuoyant's rudder shaft is the common size of 1 1/2-inch-diameter solid shaft, and we found a tiller arm that was a good fit and available by mail order at reasonable cost. It was sufficiently robust for the autopilot, yet compact enough to mount on the rudder shaft below the quadrant. The tiller arm also provided a convenient attachment point for the rudder-angle transmitter linkage.

Clairebuoyant's electronics closet (the former hanging locker) was the logical place to mount the servo unit, since it is convenient to both the batteries and the drive unit. We "super-sized" the power wiring, with a 6-gauge to avoid even nominal voltage drop to the drive unit. The drive unit was mounted on a robust oak 2-by-6, which is through-bolted to an athwartships beam that supports the rudder shaft. The 2-by-6 is also glassed to the hull on the port side. The remotely mounted pump is bolted to a shelf near the transom, out of harm's way. The drive unit is attached to the tiller arm far enough out that the 12 inches of travel just reaches the rudder stops. Limits built into the servo prevent the autopilot from crashing the quadrant into the rudder stops.

Once we programmed the remote control to recognize the autopilot and purged the air from the drive unit, it was time to try it out. We ran through the setup protocols for both the fluxgate compass and the servo unit, then engaged the autopilot. The electric pump made an audible buzz from under the helm seat, and the wheel moved slightly. It worked.

Compass and waypoint driven, it is no contest. The autopilot is the better helmsman. Going to windward, especially in light air, I like to think that I'm still a little better than the autopilot. However, like me, the autopilot still could use a little fine-tuning. Come a stormy night, he might get the job without competition from me.

Although not exactly "plug and play," installing the autopilot was not a difficult job. A bit of research and planning were required. Choosing locations for the components and configuring a strong enough mounting arrangement for the drive unit required the most effort. Already having a fluxgate compass and a remote control saved $500 or so. The Octopus drive was a bit less expensive than the proprietary unit, but more importantly provided utility and convenience with its longer travel and remotely mounted pump. Seamless integration with our existing instrument system was a major benefit, as the autopilot can now do about everything a human helmsman can do except keep a lookout.

Quentin Kinderman is retired from the federal government and lives near Annapolis, Md. He has restored several boats and is working on Clairebuoyant with the aim of long-distance voyaging.

The other components we needed included 1) a servo unit, 2) the "brains" of the system, which would interface with the instrument system and send power and instruction to the drive unit, 3) a rudder-angle transmitter to provide rudder-angle information to the servo, 4) a drive unit to actually turn the rudder and 5) a tiller arm or other mechanism to connect the drive unit to the rudder shaft.  

Image Credit: Quentin Kinderman photos
The hydraulic pump unit. It was mounted separately from the drive unit in a location that makes maintenance easier.

Two components, the servo and the rudder-angle transmitter, had to come from Nexus to be compatible with our instrument system. Shopping the Annapolis Boat Show, we found that ordering through a large discount mail-order marine supply (with boat show discount!) was the best deal.

Picking the drive unit required a bit more effort. The recommended unit was a linear hydraulic drive with nine inches of travel. This type of drive is essentially an electric pump attached to a hydraulic cylinder. While we favored this type of drive, having heard too many stories of failures with purely mechanical drive units, the nine inches of travel required that the tiller arm be short in order to permit full rudder angle. Instead, we chose an Octopus drive unit. This unit was similar in design but provided 12 inches of travel and mounted the pump independently of the cylinder. The design allowed us to use a longer tiller arm, reducing the force necessary to turn the rudder. This, in turn, reduces the stress on the components and perhaps even means it consumes less power. The remotely mounted pump made installation and service easier in the crowded aft lazarette.

The final element was the attachment of the linear drive to the rudder shaft. Many autopilot installations simply reinforce the steering quadrant and bolt the drive to the quadrant. While some autopilot manufacturers recommend this, since it provides a compact and simple installation, we decided against it for a couple reasons.

Done properly, this was not an inexpensive approach. We examined a quadrant modified by a local (and excellent) machine shop. A stainless-steel plate was fabricated to fit on the quadrant and attached to the quadrant with machine screws. This made the quadrant body virtually unbreakable. It did nothing, however, to reinforce the part of the quadrant that attached to the rudder shaft. Both the steering system and the autopilot would fail if the quadrant came loose of the rudder shaft or the shaft key sheared off. Since the autopilot drive is far more powerful than a human helmsman is, we reasoned that this approach actually increased the possibility of steering failure.

Instead we opted for a tiller arm independent of the wheel steering system for our autopilot. Several companies offer suitable tiller arms. Fortunately, Clairebuoyant's rudder shaft is the common size of 1 1/2-inch-diameter solid shaft, and we found a tiller arm that was a good fit and available by mail order at reasonable cost. It was sufficiently robust for the autopilot, yet compact enough to mount on the rudder shaft below the quadrant. The tiller arm also provided a convenient attachment point for the rudder-angle transmitter linkage.

Clairebuoyant's electronics closet (the former hanging locker) was the logical place to mount the servo unit, since it is convenient to both the batteries and the drive unit. We "super-sized" the power wiring, with a 6-gauge to avoid even nominal voltage drop to the drive unit. The drive unit was mounted on a robust oak 2-by-6, which is through-bolted to an athwartships beam that supports the rudder shaft. The 2-by-6 is also glassed to the hull on the port side. The remotely mounted pump is bolted to a shelf near the transom, out of harm's way. The drive unit is attached to the tiller arm far enough out that the 12 inches of travel just reaches the rudder stops. Limits built into the servo prevent the autopilot from crashing the quadrant into the rudder stops.

Once we programmed the remote control to recognize the autopilot and purged the air from the drive unit, it was time to try it out. We ran through the setup protocols for both the fluxgate compass and the servo unit, then engaged the autopilot. The electric pump made an audible buzz from under the helm seat, and the wheel moved slightly. It worked.

Compass and waypoint driven, it is no contest. The autopilot is the better helmsman. Going to windward, especially in light air, I like to think that I'm still a little better than the autopilot. However, like me, the autopilot still could use a little fine-tuning. Come a stormy night, he might get the job without competition from me.

Although not exactly "plug and play," installing the autopilot was not a difficult job. A bit of research and planning were required. Choosing locations for the components and configuring a strong enough mounting arrangement for the drive unit required the most effort. Already having a fluxgate compass and a remote control saved $500 or so. The Octopus drive was a bit less expensive than the proprietary unit, but more importantly provided utility and convenience with its longer travel and remotely mounted pump. Seamless integration with our existing instrument system was a major benefit, as the autopilot can now do about everything a human helmsman can do except keep a lookout.

Quentin Kinderman is retired from the federal government and lives near Annapolis, Md. He has restored several boats and is working on Clairebuoyant with the aim of long-distance voyaging.