NAC3 Autopilot steering too good of a course

[felizcortez]

I just finished installing my B&G NAC3 Autopilot and it does a great job. However it is heavy on the current consumption particularly when it is in Navigation mode steering to a waypoint. It seems to turn port and then back to starboard to overcorrect itself. If there are waves it only amplifies this behavior.

A few questions

Is there a way to lower the sensitivity of the autopilot to enable it to be "sloppier". I'm assuming this should lower the power consumption.

I do not have my speed paddle wheel installed could this be causing an issue as well? Maybe it's not sure how fast the boat is moving?

[s/v Jedi]

Originally Posted by felizcortez

I just finished installing my B&G NAC3 Autopilot and it does a great job. However it is heavy on the current consumption particularly when it is in Navigation mode steering to a waypoint. It seems to turn port and then back to starboard to overcorrect itself. If there are waves it only amplifies this behavior.

A few questions

Is there a way to lower the sensitivity of the autopilot to enable it to be "sloppier". I'm assuming this should lower the power consumption.

I do not have my speed paddle wheel installed could this be causing an issue as well? Maybe it's not sure how fast the boat is moving?

Did you read the guide and do the sea trials? At what speed? The unit is supposed to auto-configure itself that way but you can tweak manually in the menus.

I think heading, gps and rudder position are the important data sources. Do you have an nmea2000 network with these sensors and a display that can verify they are all available without data errors?

[felizcortez]

Originally Posted by s/v Jedi

Did you read the guide and do the sea trials? At what speed? The unit is supposed to auto-configure itself that way but you can tweak manually in the menus.

I think heading, gps and rudder position are the important data sources. Do you have an nmea2000 network with these sensors and a display that can verify they are all available without data errors?

Yes, I did the dockside commissioning along with the sea trial commissioning.

Also read the manual and they talk about adjusting rudder gain if things are hunting a lot. I lowered it and it didn't seem to make a difference The weird thing is it seems to be more of an issue during navigation mode than when holding a compass course.

This is at relatively low speed say less than 5 kts.

[s/v Jedi]

Originally Posted by felizcortez

Yes, I did the dockside commissioning along with the sea trial commissioning.

Also read the manual and they talk about adjusting rudder gain if things are hunting a lot. I lowered it and it didn't seem to make a difference The weird thing is it seems to be more of an issue during navigation mode than when holding a compass course.

This is at relatively low speed say less than 5 kts.

The autopilot saves two profiles: one for when going slow and one for fast. You can change the speed at which is changes over. Try lowering that speed.

I have the older version computer AC-42 but think the software is mostly the same (it’s still Robertson algorithms).

[felizcortez]

Originally Posted by s/v Jedi

The autopilot saves two profiles: one for when going slow and one for fast. You can change the speed at which is changes over. Try lowering that speed.

I have the older version computer AC-42 but think the software is mostly the same (it’s still Robertson algorithms).

I don't have the speed log transducer installed so I don't believe it is getting and using speed information. I thought it would use SOG but that doesn't appear to be the case. I do get low boat speed alarms during the first minute or two of operation but those go away. It has been in Low Speed profile the entire time. What speed do you typically have set for the change over point.

[Pyxis156]

We experienced a similar NAC3 issue with the steering not behaving as well as we would have expected. While we have both a speed transducer and a heading sensor, during a recent extended cruise we would find the paddle wheel of the transducer would foul very quickly, sometimes even daily. If I didn't remember to clean the paddle wheel before we set out, this effectively eliminated the speed through the water (STW) and put us in the same situation you are in without the transducer.

What I believe happens is in the absence of STW, the autopilot automatically falls back to then use your GPS source to calculate SOG/COG. However, SOG/COG data has a lag to begin with since it needs to cover some distance to calculate the COG trend, and how much damping you have on those data sources can also exacerbate the problem significantly. For example, if you have the damping set to the middle setting, it may take several seconds for a COG (surrogate for heading) change to register. At this point, the AP is already late to the show and starts to correct but because of the lag thinks the boat isn't changing course fast enough so it corrects more. You enter into a lag cycle that depending on your settings can either be an irritation (in calm water) or downright uncomfortable (as in your waves example where the corrections are even larger). It is also made worse by adverse current where your actual STW may be considerably faster than your SOG, and when it uses your SOG to determine what steering settings to use (i.e., high or low speed on the NAC3 or AC42) it will likely use the slower parameters with greater steering inputs due to underestimating the speed you are going through the water.

To see if what I am describing applies to your situation, setup a short route of three waypoints with the second waypoint a 90 degree (or more) turn to the third waypoint. You will find the autopilot will make the turn and then dramatically oversteer as it waits for the laggy COG (surrogate heading) information to tell it the turn has been made, and then it oversteers back the opposite direction to try to undo the massive oversteer that it initially did. The cycle repeats itself as it oscillates back and forth hamstrung by the missing STW/Heading data.

Unfortunately setting the COG damping way down doesn't help unless you are in perfectly calm water because any pitching or rolling translates into a COG change and the autopilot goes nuts trying to adjust for heading changes that aren't really happening. The only solution I found was to get the speed transducer working. Once we got the STW back it would use the STW/Heading data to determine steering inputs instead of SOG/COG and it was much smoother. When working, we almost never saw > 3' (1m) of cross track error even on long routes with currents and wind, and watching the wheel you could hardly see the steering inputs. Usually the XTE was zero.

And a side note: We ended up changing out the autopilot computer from the AC42 we originally had to the NAC3 (long story) about 1/3 of the way through our seven month trip. There was a noticeable difference in the performance of the NAC3. We thought the AC42 was good and we still think it does a great job, but even my wife noticed how much smoother the NAC3 was, especially in following seas. I don't know that I would have upgraded just to upgrade, but now that we have the NAC3 we are glad we did.

[felizcortez]

Originally Posted by Pyxis156

We experienced a similar NAC3 issue with the steering not behaving as well as we would have expected. While we have both a speed transducer and a heading sensor, during a recent extended cruise we would find the paddle wheel of the transducer would foul very quickly, sometimes even daily. If I didn't remember to clean the paddle wheel before we set out, this effectively eliminated the speed through the water (STW) and put us in the same situation you are in without the transducer.

What I believe happens is in the absence of STW, the autopilot automatically falls back to then use your GPS source to calculate SOG/COG. However, SOG/COG data has a lag to begin with since it needs to cover some distance to calculate the COG trend, and how much damping you have on those data sources can also exacerbate the problem significantly. For example, if you have the damping set to the middle setting, it may take several seconds for a COG (surrogate for heading) change to register. At this point, the AP is already late to the show and starts to correct but because of the lag thinks the boat isn't changing course fast enough so it corrects more. You enter into a lag cycle that depending on your settings can either be an irritation (in calm water) or downright uncomfortable (as in your waves example where the corrections are even larger). It is also made worse by adverse current where your actual STW may be considerably faster than your SOG, and when it uses your SOG to determine what steering settings to use (i.e., high or low speed on the NAC3 or AC42) it will likely use the slower parameters with greater steering inputs due to underestimating the speed you are going through the water.

To see if what I am describing applies to your situation, setup a short route of three waypoints with the second waypoint a 90 degree (or more) turn to the third waypoint. You will find the autopilot will make the turn and then dramatically oversteer as it waits for the laggy COG (surrogate heading) information to tell it the turn has been made, and then it oversteers back the opposite direction to try to undo the massive oversteer that it initially did. The cycle repeats itself as it oscillates back and forth hamstrung by the missing STW/Heading data.

Unfortunately setting the COG damping way down doesn't help unless you are in perfectly calm water because any pitching or rolling translates into a COG change and the autopilot goes nuts trying to adjust for heading changes that aren't really happening. The only solution I found was to get the speed transducer working. Once we got the STW back it would use the STW/Heading data to determine steering inputs instead of SOG/COG and it was much smoother. When working, we almost never saw > 3' (1m) of cross track error even on long routes with currents and wind, and watching the wheel you could hardly see the steering inputs. Usually the XTE was zero.

And a side note: We ended up changing out the autopilot computer from the AC42 we originally had to the NAC3 (long story) about 1/3 of the way through our seven month trip. There was a noticeable difference in the performance of the NAC3. We thought the AC42 was good and we still think it does a great job, but even my wife noticed how much smoother the NAC3 was, especially in following seas. I don't know that I would have upgraded just to upgrade, but now that we have the NAC3 we are glad we did.

This does sound like what is happening, i have set up waypoints that required large course changes to go to the next waypoint and the Autopilot overshoots the desired course then has to counter to come back. We are leaving our current location in a few days so I will install the speed sensor before we leave and calibrate the system again. Hopefully that resolves the issue.

[Suijin]

I have the same autopilot. You can change the gain and other settings in the settings menu for various system devices. I recall it being way too sensitive for my tastes in it's default setting after commissioning.

It definitely will be more active when steering a course than steering a heading. Also, if you have a course entered and interrupt it, say by punching heading hold or pausing it, when you re-engage course mode it will steer straight for the course line instead of the next waypoint.

I use course mode very rarely, almost always use "no drift heading hold", in part because it appears much more energy efficient. I have not noticed any different behavior when my paddlewheel is dirty in any mode. I don't believe the AP uses STW data.

[s/v Jedi]

Originally Posted by felizcortez

I don't have the speed log transducer installed so I don't believe it is getting and using speed information. I thought it would use SOG but that doesn't appear to be the case. I do get low boat speed alarms during the first minute or two of operation but those go away. It has been in Low Speed profile the entire time. What speed do you typically have set for the change over point.

Okay we found the problem: your boat is steering to the slow profile which means more rudder, so it’s oscillating around the course to steer.

Install a sensor. I recommend the DST800 which is sold by most brands but made by Airmar. Buy it from the brand you have a display or computer interface of so that you can configure it. Mine is from Airmar soI can use their displays and USB interface. The Airmar pairs with the Actisense USB interface etc.

[felizcortez]

I have a sensor it is just unplugged from the through hull so the paddle wheel didn't foul. I have just been using speed over ground before we installed the Autopilot.

What is the speed that the low speed profile and high speed profile should typically change over on a monohull?

[Suijin]

Originally Posted by s/v Jedi

Okay we found the problem: your boat is steering to the slow profile which means more rudder, so it’s oscillating around the course to steer.

I don't think that is it. The high profile, as a default in the computer, is essentially for planing speeds. You can set transition speed during AP commissioning. I have it set to 10 knots, which would be my boat borderline surfing down a wave.

My boat cruises at 6-7 knots, up to 8.5 under sail. Rarely have I been in anything BUT the slow profile. I have STW data, and I notice a distinct difference in the "busyness" of the AP when in "steer course" vs. "hold heading".

Like I said earlier, the default sensitivity of the B&G computers is set fairly high, even after running autotune during commissioning. There are a couple of parameters, rudder gain and counter steer, that can be adjusted manually in settings (RTFM). I would dial those down and see how it performs.

Another thing to check is your steering deadband. If there is much play in the steering gear the AP is going to actively hunt much more than if there is little play in the gear.

[s/v Jedi]

Originally Posted by Suijin

I don't think that is it. The high profile, as a default in the computer, is essentially for planing speeds. You can set transition speed during AP commissioning. I have it set to 10 knots, which would be my boat borderline surfing down a wave.

My boat cruises at 6-7 knots, up to 8.5 under sail. Rarely have I been in anything BUT the slow profile. I have STW data, and I notice a distinct difference in the "busyness" of the AP when in "steer course" vs. "hold heading".

Like I said earlier, the default sensitivity of the B&G computers is set fairly high, even after running autotune during commissioning. There are a couple of parameters, rudder gain and counter steer, that can be adjusted manually in settings (RTFM). I would dial those down and see how it performs.

Another thing to check is your steering deadband. If there is much play in the steering gear the AP is going to actively hunt much more than if there is little play in the gear.

The STW sensor is used for much more than just the profile switch. I’m pretty sure the OP needs to get that sensor on-line to get good AP steering. It may well be that he needs to tune settings after that as well.

We have the switch from lo to hi at 7 knots. Our hull speed is 12 knots (we have a planing hull but hi profile works well for anything above 7 kts. I have lowered rudder gain a bit as well but don’t see a difference when switching between course, wind or waypoint mode. Unless the wind is flaky, current ripping back and forth etc. I don’t see why it would happen. Maybe maybe a sensor’s data output needs to be dampened a bit?

It seems we still need to buy a spare computer and the NAC3 may be in our future

[jmschmidt]

RTFM brother.

[felizcortez]

I put in the speed sensor and that seems to have calmed things down. However, my speed sensor is reading about 1kt too slow so I ordered a replacement and will install that when it arrives.

[Pyxis156]

Glad it has helped. I find our speed sensor also reads about 30% slow and in the configuration screens for the chartplotter there is a way for you to apply a correction factor to the speed transducer that was easy to do. I waited until slack tide in a protected cove when there was no current or wind and then did reciprocal courses to compare SOG to STW. When I was confident I had consistency and a good number, I applied that correction factor in the settings.