Steel boat compass problems

Wil Hamm, W-H Autopilots Inc.

12/3/03

All passage making vessels need to know where their vessels are heading when moving. There are two reasons for this:

The heading must be close to the heading required to get to the desired destination in the shortest possible distance.

The heading must be corrected continuously (with an autopilot) so that the vessel stays on a straight course and doesn’t wander back and forth around the desired course.

The heading for number one above need not be extremely accurate since even a simple GPS receiver will give an accurate heading over ground once the vessel has been in motion for a few minutes (but only when no wind or current is present). This GPS data can then be used to manually correct the autopilot heading. Most modern autopilots will make these corrections automatically if interfaced with a GPS receiver or electronic chart machine. Since a typical GPS measures only position, not speed, it needs time to compute speed after the vessel moves. This mode of operation makes the GPS unsuitable to provide the short time heading data required for number 2 above.

There are new GPS/GYRO systems now available that do provide very accurate true heading instantly from satellite signals. Unfortunately, their cost is in the $4,000 to $6,000 area, and they require two fairly bulky antenna domes.

A second very accurate heading source is a mechanical north seeking gyro compass system. These systems are used routinely on large commercial and military ships, but are not really practical for yachts under 90 feet in length. Their main drawback is cost, which initially runs from $10,000 to $20,000. The mechanical spinning gyro elements can fail in 3 to 5 years and cost at least $6,000 to replace. They also require 4 to 6 hours of warm up time and have a moderate power drain, which normally is never turned off. They always read true headings to better than 1 degree. They can also "tumble" in severe storms making them unusable until the storm is over.

There are a number of non-north seeking gyro systems available that give excellent short time heading data, but drift off as much as 10 or 15 degrees per hour. The mechanical versions of these gyros were used in most WW2 military aircraft. I personally have used these in steering yachts, but abandoned the project a long time ago because of the high initial cost and the frequent rebuild cycle required for these aircraft components. They would correct the gyro (called caging) frequently using a magnetic compass or fluxgate sensor as the heading standard. They also made corrections using radio direction finders tuned to known radio beacons or even AM radio stations. These radio direction finders could be off as much as 10 degrees. Before GPS was invented, sailing yachts had been known to sail from Seattle to Hawaii using a radio direction finder tuned to an AM entertainment station in Honolulu.

In the last 10 years, there have been several new developments that have made short time gyro heading sensing far more practical. They are the development of both fiber-optics-laser and Coriolis effect rate gyro systems. Both these technologies involve no moving parts, and appear to have long lives without any maintenance. Both these technologies come up to full opertion in just a few seconds, and require very small amounts of electrical energy when operational. Since they are inertial devices, they can be mounted almost anywhere in a steel vessel. They do not respond to magnets or steel masses. They are being mass produced and used in many machines including missiles, aircraft and even passenger cars.

The major method of determining heading is still sensing the magnetic north pole (located near the earth’s NorthPole). The magnetic North Pole is not exactly at the geographic North Pole, so the magnetic heading is never exactly the true heading. The heading must be corrected by a given number of degrees (called the magnetic deviation), which, unfortunately, varies from location to location. The dip angle of the magnet field lines get worse as one goes north, thus it becomes impossible to use near the North Pole. Fortunately, most yachts don’t go that far north, and fisherman normally don’t go beyond Bristol Bay, where magnetic compasses still work.

Magnetic compasses have been used on ships for a long time, but most magnetic sensors today are second harmonic fluxgate magnetometers. This magnetic sensing device measures magnetic field direction, just like the magnetic compass, but has the drawback of being even more sensitive to tilt than the magnetic compass. In the Pacific Northwest a 1 degree tilt on a fluxgate creates a 3 degree heading error! This means that the fluxgate systems must be gimbaled to avoid large errors due to the pitch or roll of the vessel.

The end result is that fluxgates do not respond as rapidly as magnetic compasses (due to the electronic filtering required for limiting the errors from tilting the sensor). Since the fluxgates have no moving parts, except for the gimbal components, they have a longer life than most mechanical compasses, which have delicate jewel bearings and are filled with oil. The other big advantage of the fluxgate is that one can devise electronic compensation schemes to make corrections due to nearby soft and hard iron errors. Since fluxgates require electronic circuits just to operate, they can easily provide digital data and have internal filters with adjustable time constants. All of these features make fluxgates ideal for operating autopilots with GPS interfaces.

Steel yachts may be floating magnets (hard iron problems) or just have a lot steel paths to deviate external magnetic fields(soft iron problems).

The hard iron effects can come about when the vessel is moored in a North/South direction for long periods of time (and by other, less understood reactions, from welding and cutting). Hard iron magnetism can be removed by applying a large AC magnetic field to the entire vessel. This "degaussing " technique was used on large commercial and military vessels about 60 years ago. The local degaussing station was located under the water at Jefferson Head (just north of Bainbridge Island, WA). The shore control building would apply huge AC currents to a giant underwater coil just as a vessel positioned its self directly over the coils. The entire steel vessel would hum at 60 Hz! I know of no existing degaussing station still in operation. Another , solution, though not too practical, is to change the vessel’s direction at the dock by 90 degrees every month or so.

The soft iron errors cannot be removed, except by avoiding close proximity of large steel structures around the magnet heading sensor. One standard technique is to place sensors in line with glass windows in the pilothouse.

Most modern fluxgate sensors (including those built by W-H and KVH) have auto-compensating schemes to correct some of the hard and soft iron errors. One just selects the compensating mode and turns the vessel very slowly for two complete turns. This should be done in very calm water and with no wind. This technique can improve a sensor’s accuracy from 3 or 4 degrees of error to within 1 degree or better. It cannot totally eliminate really large errors.

Placing the sensor on an aluminum or plastic pipe far above the deck is another possibility. The latter solution has one large problem associated with it. Not only does it swing though a large arc when the boat rolls, but it swings over a large steel mass beneath it, The boat going due north creates a large "heeling effect error". This solution will make an autopilot turn the vessel to port and then starboard as it rolls!

It will steer the boat fairly well in dead calm.

The only real reasonably priced solution is to use a compensated fluxgate sensor mounted 5 to 7 feet above the deck, and have the output averaged over 10 to 30 seconds. One then must use a fast responding solid state gyro mounted low in the vessel to steer the boat from moment to moment. The averaged fluxgate signal is then used to correct the slow drift of the solid state gyro sensor. W-H Autopilots has just such a system available.

After a steel boat is finished and launched, a quick magnetic survey should be attempted. One should have a portable or hand held magnet compass in hand as he comes down the dock to the vessel. This will determine the heading of the dock, which should be parallel to the subject steel hull. Then enter the vessel and check the heading in various locations that are farther than 3 feet from large steel sheets or walls. If a spot is found that seems to read within 5 degrees or so of the actual boat heading, and is otherwise suitable as a home for the sensor, try attaching the magnetic compass with duct tape. (It may have to be mounted on a cardboard box.)

Now take the boat away from the dock and try reading the compass in at least the four major headings. (N,NW,S,SW) If the sea is calm and there is no current or wind, one could use one’s GPS receiver to determine the headings. Be sure that the GPS is set to read magnetic headings, not true heading. Compass adjusters usually use landmarks (roads etc) that have been built with exact magnetic headings.

These readings will be a test of both the hard and soft iron magnetic effects in your boat. If everything seems to be within 5 to 8 degrees, you may be able to use a standard, electronically compensated fluxgate to steer your boat and point you in the right direction. If the results are more than 10 degrees out, then a solid state gyro system with fluxgate sensor on a pole, might be recommended. If both your boat and your budget are large, a north seeking mechanical gyro or GPS gyro might also be an option The reliability of the GPS system in bad seas is still an unknown (at least in the opinion of the author, who has installed one in a 75 ft vessel).

A simple magnetic standby compass has the advantage of operating without electricity and is also able to survive a lighting strike. It will, in general, need a compass adjuster to make it read accurately. The latter procedure may make this compass almost as costly as an electronic sensor. In addition, the most convenient place to mount it is often the worst place for accurate readings.