GPS-Hdg Sensor Shoot-Out

[Pizzazz]

Dockhead, the KVH Autocomp is spec'd to 0.5 deg at up to 30 deg tilt and in all conditions. It is not like a cheap Raymarine fluxgate compass. It does not have a gyro but it is a really good, gimballed fluxgate compass with variable level of damping. Plus, it has autocompensation so you never have to worry about calibration. I am just sharing my experiences.

I understand the fascination with new technology but it is not always better. If you read up on the Coastguard electronics update (during the Furuno decade at least, they have moved on to Raymarine now), they were positively surprised by the accuracy of quality fluxgate compasses and issued a statement somewhere in the mid-2000s that they were good enough to use as the main ship compass (which is what I do as well).

My issue with the GPS compass is that conceptually it is quite difficult to make it work. You have two antennas, they measure their positions, then you average them and get the heading. Great. Now imagine the boat is dynamic, waves, etc. The electronics need to take the 3D orientation and adjust the GPS input for movement in three dimensions. It is a lot of calculations and this is why the dynamic accuracy of these compasses is not so good. If you want stabilized output, you have to average over a longer period of time. It is just a difficult way to calculate heading and it requires quality accelerometers (expensive) and proprietary algorithms. Hence the prices we see. I would rather use the Earth's magnetic field (which is not easy to jam on my boat. Plus, my boat is mostly nmea 0183.

I have tested against the best in class H2183 and the KVH works better for me. The specs confirm it. I can see why vendors have moved to solid state autopilot compasses (easy calibration, less sensitive to magnetic fields, cheap and good enough for autopilots). Has the Coasguard switched though?

One day I will get a real gyro or an INS on the boat and settle this debate .

[Dockhead]

Originally Posted by Pizzazz

. . . My issue with the GPS compass is that conceptually it is quite difficult to make it work. You have two antennas, they measure their positions, then you average them and get the heading. Great. Now imagine the boat is dynamic, waves, etc. The electronics need to take the 3D orientation and adjust the GPS input for movement in three dimensions. It is a lot of calculations and this is why the dynamic accuracy of these compasses is not so good. If you want stabilized output, you have to average over a longer period of time. It is just a difficult way to calculate heading and it requires quality accelerometers (expensive) and proprietary algorithms. . . .

Well, except that this is not at all how satellite compasses work at all. It has nothing to do with comparing positions, and they have far better dynamic performance than fluxgate compasses. You are right that greater distance between the antennas of a satellite compass increases the accuracy, but heading is calculated by a satellite compass not from the difference in positions, but from the PHASE DATA from the satellites. The wavelength of GPS signals is a few centimeters, so it is possible to measure exquisite differences in RELATIVE position to the satellites by PHASE, without knowing the ABSOLUTE position within even orders of magnitude of the same accuracy. That's the whole trick.

Satellite compasses do not require accelerometers at all for their primary processes. Their dynamic performance is almost exactly the same as their static performance, unlike the case with fluxgate compasses. They have accelerometers only for the sake of doing dead reckoning in case the signal is briefly blocked by a bridge or cliff or something.