@Radio Flyer yes, rudder position sensor is wired direct to the course computer. I think it’s a variable resistor.
Best sailing autopilot?
- Thread starter Radio Flyer
- Start date
Why would that matter? The AP is not for tight maneuvers. It’s for setting and holding a course.
Are you certain it was ethernet. NMEA only released the Ethernet protocol about a year or two ago. In 2000 there were 2 NMEA protocols, 0183 and 2000 both based on Canbus which was developed in the auto industry. As the ancient 0183 matured it allowed faster transmission rates, from around 4800 baud to 38Kbaud. NMEA 2000 now allows fairly rapid transmission rates, however, it is not as fast or able to support as much data as ethernet. Which is why all the digital radars and sonars use ethernet, but they are proprietary protocols.
Sounds like the AP is not calibrated correctly. As part of the calibration the AP should go through a lock to lock routine to find the max rotation of the wheel. It can then calculate center and it knows how far to turn the wheel. Wheel pilots are turn much slower than below deck pilots which are slower than panicked humans. The rotation rate will make a difference in how tight a circle the AP can steer. On the B&G the turn rate can be set in the calibration process.
It is likely that the hunting you are concerned with is really just the AP responding to the conditions on the water. Waves, wind, wake, current all affect the boat the greater these forces are the more the AP has to work and compensate for them. APs are reactive, they can not anticipate what the boat will do, they can only respond to what the boat has done. A good human helmsman can anticipate, the human can see the wave action and can see the wind and make subtle course changes to compensate for their effect.
On a 27 foot boat you likely have a wheel pilot. Wheel pilots are not as computationally nor physically as fast as a below decks AP. This makes it harder for the AP to make the kind of course corrections you are expecting. Simply put, the problem may be with your expectations and not the AP.
Well, I didn't describe the circle properly. I see the confusion.
When holding a line, a target directly ahead, the sight line over the bow moves in a circle. With each wave you ride up and down, left to right. At best you draw a tight circle around your target.
That is how I judge the accuracy of sail trim and AP together.
The problem is rougher seas push the boat to be less accurately on point. The circle is drawn bigger until it becomes unstable.
The degree to which the line becomes unstable is a function of the APs calculation speed, i.e., how fast can it process the data sent to it and the drive speed. As I mentioned earlier, the AP can not anticipate the boat's motion, a human can this allows the human to hold a tighter line.
The more powerful the AP, the better it can hold a line. Also the setting, is it trying to hold a compass course or a wind course. In boisterous conditions, holding to a wind course is a much smoother ride than trying to hold to a compass course. Holding to a wind course does require closer monitoring of wind shifts and human course adjustments than holding to a compass course.
You haven't mentioned (or I missed) the type of AP you are concerned with. If it is a wheel pilot, your expectations my exceed its capability.
Dave, I don't think that's correct. I believe NMEA 0183 used the EIA-422 physical layer protocol, and transmits ACII , "8, none, and 1" (bits per character, parity, and number of stop bits), up to 4800 bps, 'though some vendors have pushed it to 9600 and beyond, as well as using UDP to encapsulate NMEA ASCII sentences over IP networks, over ethernet.
NMEA 2000 uses CANBUS, which runs at 250kbps, and uses a binary message format, not ASCII.
You are right with the NMEA 0183 & 2000 I had both systems on different boats. Never "ethernet" I misspoke.
2000 was faster with the feedback of the electronic compass unit. It did hold a better line.
You are again right in my lofty expectations. I want the AP to anticipate the regular wave forces.
Your description of the B&G software suggests be different/better than the Raymarine controls. All I had from the Raymarine was the integral (response rate) value to change.
In my understanding the AP control we have is the "I" or integral control. I'd like to have the differential to experiment with.
I'll go after a B& G manual to see if the tools are available.
You know, the performance of your autopilot is influenced by the location of your compass and gyro sensor. It's best to have it at the center of motion of the boat, meaning, the place of least movement with yaw and pitch and roll, typically near the CG, which is likely low down near the mast.
I wonder, does your C27 have a tiller or a wheel? A below-decks AP might be a bit much for a boat that size, and certainly very expensive, considering the alternatives. A wheel pilot would do very well on a C27. A tiller pilot if you have a tiller.
The C27 and a Hunter 325 in Florida are the last 2 boats. In 2024 will be my next ship. I've always refit whatever boat I bought. From engine parts, controls, to the macerator. (There is a yucky job).
In answer, on the C27 it was a tiller. That controlled by the raymarine ST-1000. The 325 controlled by a Raymarine EV-100 wheel pilot.
I just want to move up in function this time. Also to a larger boat. 36 to 42 is my target. I'm eyeballing a couple local boats and figure a plan is better than a reaction. I'm in the 15 to 20 year old sailboat market. At that point original equipment is pretty dated and begging to be refreshed. If for nothing else the map software and firmware need to be updatable. The garmin chartplotter with its unique footprint on the 325 was locked in the past. Sometimes you have to start fresh.
In answer, on the C27 it was a tiller. That controlled by the raymarine ST-1000. The 325 controlled by a Raymarine EV-100 wheel pilot.
I just want to move up in function this time. Also to a larger boat. 36 to 42 is my target. I'm eyeballing a couple local boats and figure a plan is better than a reaction. I'm in the 15 to 20 year old sailboat market. At that point original equipment is pretty dated and begging to be refreshed. If for nothing else the map software and firmware need to be updatable. The garmin chartplotter with its unique footprint on the 325 was locked in the past. Sometimes you have to start fresh.
Have you experienced AP with and without rudder feedback? I wonder about the performance difference.
No. It’s part of the system and setup process. I would assume the system would not be fully capable without it.
For a boat in your size range you will need a below decks AP. Currently the B&G H5000 is probably the most advanced AP computer.
Thanks, the B&G system was in my planned review, but it is great to have experienced opinion from the people who use it. I really appreciate kindred sailors insights. Marketing obfuscates the performance we seek under a glitz of shiny features. Our pursuits are often a one man quest. Glad to hear from the raft of opinions.
It would be cool if the vendors would have online simulators so you could try them out!
Without a rudder feedback, all autopilots will hunt like a dog. The only way they know they have made a rudder change is to monitor the directional change of the boat itself. This leads to a spiraling feedback loop where the AP is fighting itself trying to chase the wind. Yes, a rudder feedback will make a world of difference.
While it is true that in the past most AP's were purely reactive, that is not true now for higher end ones like B&G and NKE. These have very good steering algorithms that go beyond simple PID and are anticipative. They cannot anticipate one-off stuff like a single breaking wave from a strange direction, but they are very good at learning steady sea state and steering in anticipation like a human. I enjoy watching our B&G start to apply rudder well before the boat has gone off course going up a wave, then reverse the rudder near the top well before the boat has begun to turn going down the wave, then beginning a correction before the boat hits the trough and begins to turn again. It is at least as good as me, and definitely more attentive and not prone to fatigue.
We have both current B&G NAC3 and Raymarine EVO pilots. They share system data like nav data, speed, gps, wind, etc, but have separate compasses and rudder feedback units. They drive separate, but identical, drive units. The Raymarine does seem to use simple PID steering, and is not as predictive or steady in larger seas than the B&G. However, it is still lightyears ahead of past Raymarine products. The B&G steers significantly better in changing conditions, as well as automatically learns conditions and changes tuning, where the RM needs to have manual intervention and tuning to changing conditions.
If all else is the same as far as boat data available wrt speed, wind, and gps, then the difference among AP's falls into three categories: steering algorithm, useable compass data, and rudder feedback precision. From comparing ours, it is obvious that the B&G steering algorithm is better than the RM, and the B&G rudder sensor is more precise than RM's (the RM rudder sensor is the same old one they have been using for decades, while the B&G is a newer N2K sensor). I also suspect that the B&G uses more of their compass data in their algorithms than the RM, as it outputs more data such as heave, sway, and surge. If the RM uses these, they keep it entirely internal with no way of knowing, and they only advertise using roll, pitch, and yaw. I can watch both rudder sensor outputs simultaneously, and the RM is always 2-3* lagging or leading the B&G, or can just be 0-2* off with the rudder still. B&G lists theirs as +/- 0.5* accuracy but RM does not list specs for theirs. My experience is that the RM is +/- 2* and reacts slower than the B&G.
For precise sailing to wind, one needs very well calibrated speed and wind sensors. If on a monohull, then one also needs to have a calibration table for these in relation to heel angle. Otherwise, expect the boat to hunt around a bit, but not near as much as using a mechanical windvane. If no rudder sensor, then don't expect much at all steering to wind. One difference between the RM and B&G sailing deep angles is the B&G knows where the gybe is and can be set to prevent jibing. It can also be set to automatically change between AWA and TWA depending on point of sail. This is useful if one is lazy, or experiences a sudden dramatic windshift, or needs to bear off quickly in a wind gust. The RM can be manually set to AWA or TWA, but cannot switch automatically. It also has no gybe prevention.
To my knowledge, only the B&G H5000 and NKE systems allow calibration tables. However, these are expensive systems, pretty complicated to setup correctly, and probably overkill for most cruising boats. They are go-to systems for racing boats - and for good reasons. Otherwise, the B&G NAC systems contain the steering algorithms of the H5000 minus the use of calibration tables and some other more esoteric things.
Mark
While it is true that in the past most AP's were purely reactive, that is not true now for higher end ones like B&G and NKE. These have very good steering algorithms that go beyond simple PID and are anticipative. They cannot anticipate one-off stuff like a single breaking wave from a strange direction, but they are very good at learning steady sea state and steering in anticipation like a human. I enjoy watching our B&G start to apply rudder well before the boat has gone off course going up a wave, then reverse the rudder near the top well before the boat has begun to turn going down the wave, then beginning a correction before the boat hits the trough and begins to turn again. It is at least as good as me, and definitely more attentive and not prone to fatigue.
We have both current B&G NAC3 and Raymarine EVO pilots. They share system data like nav data, speed, gps, wind, etc, but have separate compasses and rudder feedback units. They drive separate, but identical, drive units. The Raymarine does seem to use simple PID steering, and is not as predictive or steady in larger seas than the B&G. However, it is still lightyears ahead of past Raymarine products. The B&G steers significantly better in changing conditions, as well as automatically learns conditions and changes tuning, where the RM needs to have manual intervention and tuning to changing conditions.
If all else is the same as far as boat data available wrt speed, wind, and gps, then the difference among AP's falls into three categories: steering algorithm, useable compass data, and rudder feedback precision. From comparing ours, it is obvious that the B&G steering algorithm is better than the RM, and the B&G rudder sensor is more precise than RM's (the RM rudder sensor is the same old one they have been using for decades, while the B&G is a newer N2K sensor). I also suspect that the B&G uses more of their compass data in their algorithms than the RM, as it outputs more data such as heave, sway, and surge. If the RM uses these, they keep it entirely internal with no way of knowing, and they only advertise using roll, pitch, and yaw. I can watch both rudder sensor outputs simultaneously, and the RM is always 2-3* lagging or leading the B&G, or can just be 0-2* off with the rudder still. B&G lists theirs as +/- 0.5* accuracy but RM does not list specs for theirs. My experience is that the RM is +/- 2* and reacts slower than the B&G.
For precise sailing to wind, one needs very well calibrated speed and wind sensors. If on a monohull, then one also needs to have a calibration table for these in relation to heel angle. Otherwise, expect the boat to hunt around a bit, but not near as much as using a mechanical windvane. If no rudder sensor, then don't expect much at all steering to wind. One difference between the RM and B&G sailing deep angles is the B&G knows where the gybe is and can be set to prevent jibing. It can also be set to automatically change between AWA and TWA depending on point of sail. This is useful if one is lazy, or experiences a sudden dramatic windshift, or needs to bear off quickly in a wind gust. The RM can be manually set to AWA or TWA, but cannot switch automatically. It also has no gybe prevention.
To my knowledge, only the B&G H5000 and NKE systems allow calibration tables. However, these are expensive systems, pretty complicated to setup correctly, and probably overkill for most cruising boats. They are go-to systems for racing boats - and for good reasons. Otherwise, the B&G NAC systems contain the steering algorithms of the H5000 minus the use of calibration tables and some other more esoteric things.
Mark
That in. A nutshell is what I was looking for. I never had a rudder feedback. And frankly was unassured the RM would do much with it. Thanks for your insight. I believe the BandG H-5000 is the correct direction to go. With rudder feedback.
That will feed my winter 2013 & 2014 thing to do.
Apparently your Raymarine unit was improperly installed, or was a wheel or tiller pilot. You didn’t have a valid basis to experience what a below decks properly configured Raymarine AP does.
I‘d get the same brand as your chart plotter so they can interface as designed.
Last edited:
WOW the B & G H-5000 looks nice... I started looking at what appeared to be a basic H5000 on a $40,000 sailboat.
Here is the B&G Example of a H5000 system. I am sure you can cut and paste some of the units. But if you are talking a highly sensitive system.
H5000 system example
An example of a typical H5000 system. At the centre of the system is the Central Processor Unit (CPU). All sensor information is fed back to the CPU and can be easily controlled and configured via the Graphic Display or webserver.
- H5000 Graphic Display $1199
- B&G Hydraulic Ram - RAM-T1-12V $2199
- B&G Precision-9 Compass $699
- B&G RF 25N Rotary Rudder Sensor $349
Here is the B&G Example of a H5000 system. I am sure you can cut and paste some of the units. But if you are talking a highly sensitive system.
H5000 system example
An example of a typical H5000 system. At the centre of the system is the Central Processor Unit (CPU). All sensor information is fed back to the CPU and can be easily controlled and configured via the Graphic Display or webserver.