Off course, how coarse is the track?

[johnphilip]

Using a Raymarine wheel pilot linked to our Garmin plotter we do have a strange phenomenon.
Using the wheelpilot just by pressing auto to keep the boat in the same direction seems fine. When however I tell the pilot to track to the next waypoint I find that initially the boat moves steadily to a cross track error of hundreds of feet, it then turns to correct and crosses the track to a large error the other way. However after this it settles on an almost perfect track to the waypoint. Does anyone else see this effect? It might help explain a situation that occured with another boat a couple of years ago ( we were not the ones using the autopilot in this instance.)

 

[johnalison]

Mine depends to some extent on the cross-tide. In steady conditions it will start on the course indicated and wander between narrow limits. If there is much current, it will start on the compass course indicated before realising its error and back-tracking, sometimes with some overrun. I don't know to what extent it can take COG into account.

 

[knuterikt]

Using a Raymarine wheel pilot linked to our Garmin plotter we do have a strange phenomenon.
Using the wheelpilot just by pressing auto to keep the boat in the same direction seems fine. When however I tell the pilot to track to the next waypoint I find that initially the boat moves steadily to a cross track error of hundreds of feet, it then turns to correct and crosses the track to a large error the other way. However after this it settles on an almost perfect track to the waypoint. Does anyone else see this effect? It might help explain a situation that occured with another boat a couple of years ago ( we were not the ones using the autopilot in this instance.)

Steering a compass course is different than steering to a waypoint.

First you should know if your plotter use a greate circle or a rhumb line when calculating the cours(es) to steer.
If it is using great circle the plotter will (for longer distances) calculate the different course to steer (unless you are going N/S or E/W at Equator).

If you set a compass course on the autopilot, you are setting the heading. If you are drifting - the point you are heading towards will change (but the compass course is the same).

If you have drift when steering to a waypoint the plotter will start to correct your heading when the crosstrack error reach a set value (you can probably set the maximum cross track error you will accept on the plotter).
It's possible that plotter try to "learn" the course to steer to keep you on the track, maybe there is something about this in your manual. (You didn't say what model you have


Edit:
In the middle of a 200nm stretch the difference between Great Circle route and Rhumb line can be around 2nm

 

[DaveS]

Using a Raymarine wheel pilot linked to our Garmin plotter we do have a strange phenomenon.
Using the wheelpilot just by pressing auto to keep the boat in the same direction seems fine. When however I tell the pilot to track to the next waypoint I find that initially the boat moves steadily to a cross track error of hundreds of feet, it then turns to correct and crosses the track to a large error the other way. However after this it settles on an almost perfect track to the waypoint. Does anyone else see this effect? It might help explain a situation that occured with another boat a couple of years ago ( we were not the ones using the autopilot in this instance.)

I experience something similar with my Raymarine tiller pilot. If, in "auto" mode, heading straight towards the next waypoint, I switch to "track" mode, the pilot immediately moves the tiller, sending us off course. After a few minutes it gradually brings us back then tracks perfectly with XTE never more than 0.01 nm. So I don't see the second, other side, course deviation you report, but in other respects very similar.

I also find that if I use track mode to follow a multi waypoint route, that after passing a waypoint it allows a big error to develop before slowly condescending to close with the new route line.

I should add: no cross tides causing issues, and my coastal sailing involves legs far too small to worry about GC errors.

 

[knuterikt]

I also find that if I use track mode to follow a multi waypoint route, that after passing a waypoint it allows a big error to develop before slowly condescending to close with the new route line.

Different plotters have different ways of switching from one waypoint to another on mine (Furuno MFD8) I can set how this is to be done
There are three different modes

Cross Line: Change the waypoint when the vessel moves through an imaginary cross line (vertical line) that moves through the center of the destination point.
Circle: Change the waypoint when the vessel comes within the circle.
Cross Line & Circle: Change the waypoint when one of the above two conditions is met.

The radius of the circle is also adjustable.

Depending on the plotter this can explain what you see when switching waypoint.

I should add: no cross tides causing issues, and my coastal sailing involves legs far too small to worry about GC errors.

The OP didn't say anything about the length of his legs, thought it was worth mentioning the GC issue.

Since you have no cross tides giving drift it must be something else.

I'm guessing now but one possibility is that the compass course (magnetic) used by the autopilot and the desired COG is not aligned.
The only course to steer information the plotter & autopilot can use is the compass (remember that COG is calculated after the fact, based on the last POS and the previous POS).

So I'm thinking that the initial course is set based on the direction to the target and the AP steers this course on the compass.
The plotter detects the difference between COG and compass course and learns how to correct. This could also influence the tracking after switching waypoint.

Note that when I write compass I mean the heading sensor used by the autopilot (some of these is supposed to give true heading others give magnetic but all can give wrong heading as they all use earth magnetism to measure the heading).

You could try the following:
-Steer manually towards a waypoint
-Set the autopilot (AP) to steer on compass (current course to waypoint)
-Compare COG from the plotter and course displayed on AP, note the difference
-On the plotter set the AP to steer to the same waypoint
-Note the new COG and compass course (when wearing off the track)
-Note COG and compass course after the boat has settled in on the correct track

Compare your observations.
Would be interesting to know the results

COG is the direction the boat moves over the ground, if the boat is running backwards that will be your COG
Heading is the direction the bow is pointing

True heading and COG is not necessarily the same and the plotter don't know why they are different.
Since the only heading information the plotter has is the compass it must start the track by steering by the compass.
After a little while the plotter "see" that the boat is moving in the wrong direction and change the heading the AP is to steer.
The AP does not know the deviation and/or drift so for each new course the process has to be repeated.

 

[john_morris_uk]

I have no idea about why the electronics are playing up but my first thought is that steering to keep the boat on the course to the waypoint using cross track error can be a slow and inefficient way to get there... For short passages with no great change in the tidal stream it's ok, but for longer passages where the tidal stream changes its not optimum.

 

[knuterikt]

I have no idea about why the electronics are playing up but my first thought is that steering to keep the boat on the course to the waypoint using cross track error can be a slow and inefficient way to get there... For short passages with no great change in the tidal stream it's ok, but for longer passages where the tidal stream changes its not optimum.

Using a waypoint and cross track error work fine when you must follow a narrow "road".
For longer crossings with tidal stream nothing beats proper passage planning and steering a to a heading and change the course during the voyage.

 

[johnphilip]

Collecting thoughts

Thanks forthe interest, different replies have raised new points.
Normally when using the autohelm, which is either for short periods to release the helm for other duties such as eating or else when under power (when we prefer not to sit with little interest tied to the wheel) I just jab the "Auto" button on the Wheelpilot to maintain current direction which it does accurately and immediately.
However if when heading to a waypoint I choose the "Track" button the boat veers a distance off to one side then the other (as seen on the "distance to track" on the plotter) before settling on an accurate course up the line. The only time this has been an issue is when in restricted waters when the wanting to follow a route up a channel say between the mud and the shipping channel. Ok so I don't need to use the function but I do wonder whether others have noticed it happening. I don't know either what sentence from the plotter the Wheelpilot is using which might give rise to the effect.
There was an instance when we were manual helming and a boat to our side that had been parallel veered steadily but not violently towards us soon after switching to autopilot, although we were both aiming for the same buoy.
Anyone else?

 

[macd]

The OP's experience is normal but perhaps suggests the compass is not accurately calibrated. When 'go to waypoint' is pressed the autopliot (tiller or wheel) turns to what it believes is the correct compass heading. Obviously it cannot use cross-track error at this point as none has been established. Only a little while later will the system default to course-correcting by XTE, whether this has arisen due to faulty calibration or cross-streams. This is when it settles down to the more accurate course-keeping the OP decribes.

Obvious correct calibration is the answer, but there's nothing to stop you pressing the "+10" or "-10" buttons when the boat initially veers away from the correct route to bring it closer to the course you want.

 

[knuterikt]

To be able to dive deeper into this to find an explanation one have to know the model and SW version of both the plotter and the autopilot.

My autopilot supports 4(5) different ways of controlling from the plotter, what my plotter output is RMB

From the autopilot manual

AUTOMATIC SENTENCE SELECTION
For navigation inputs, the system looks for groups of sentences in this order:
RMB
APB
APA
BOD and XTE
The search stops when the first sentence in the list is found. If it reaches BOD/XTE,
and only one of these is present, the autopilot operates from the data in that sentence.

Description of the RMB - found on internet

RMB - The recommended minimum navigation sentence is sent whenever a route or a goto is active. On some systems it is sent all of the time with null data. The Arrival alarm flag is similar to the arrival alarm inside the unit and can be decoded to drive an external alarm. Note the use of leading zeros in this message to preserve the character spacing. This is done, I believe, because some autopilots may depend on exact character spacing.

$GPRMB,A,0.66,L,003,004,4917.24,N,12309.57,W,001.3,052.5,000.5,V*20

where:
RMB Recommended minimum navigation information
A Data status A = OK, V = Void (warning)
0.66,L Cross-track error (nautical miles, 9.99 max),
steer Left to correct (or R = right)
003 Origin waypoint ID
004 Destination waypoint ID
4917.24,N Destination waypoint latitude 49 deg. 17.24 min. N
12309.57,W Destination waypoint longitude 123 deg. 09.57 min. W
001.3 Range to destination, nautical miles (999.9 max)
052.5 True bearing to destination
000.5 Velocity towards destination, knots
V Arrival alarm A = arrived, V = not arrived
*20 checksum

I'm convinced that it is the difference between "True bearing to destination" and the compass course the autopilot uses to set the initial course that cause this to happen.
The software in the autopilot need some time to detect and correct this error. Note that there is no COG from the GPS going to the autopilot.

 

[earlybird]

...When however I tell the pilot to track to the next waypoint I find that initially the boat moves steadily to a cross track error of hundreds of feet, it then turns to correct and crosses the track to a large error the other way. However after this it settles on an almost perfect track to the waypoint. Does anyone else see this effect? It might help explain a situation that occured with another boat a couple of years ago ( we were not the ones using the autopilot in this instance.)

Are you setting the correct initial conditions?
To acquire track, my ST4000 handbook says that XTE should be <0.1 M and heading within 5 deg. of BTW, although, TBH, I rarely use the track function.

 

[AntarcticPilot]

One little "gotcha" I've found with my Raymarine C80 is that the compass deviation set in the software sometimes resets itself without warning back to the factory default. This may be connected with fairly long periods of disuse. The factory default is about 7E; the actual value where I am is about 5W! Resetting the deviation usually sorts out any problems such as that noted by the OP. I usually spot it pretty quickly when I see that the ship's head and course vectors differ wildly on the screen.

 

[johnphilip]

Are you setting the correct initial conditions?
To acquire track, my ST4000 handbook says that XTE should be <0.1 M and heading within 5 deg. of BTW, although, TBH, I rarely use the track function.

Yes I always use this when within the margin or the wheelpilot just errors and bleeps at you.

 

[malcp]

Yes I always use this when within the margin or the wheelpilot just errors and bleeps at you.

A solution to this is to command RESET CROSSTRACK ERROR (In Raymarine parlance) on the plotter immediately before TRACK mode on the A/P. What this does is to re-draw the rhumb line between your current position and the next waypoint. If you don't do this the first thing the A/P does after entering TRACK mode is to seek the rhumb line between the previous and next waypoints on your route. This command will zero your cross track error; but don't wait too long before hitting TRACK mode or it will build up again.

To the OP, I'd suggest trying reducing the WPT arrival warning on the plotter. This means you will be closer to the actual WPT before the A/P asks you to accept the new course.

 

[DaveS]

Different plotters have different ways of switching from one waypoint to another on mine (Furuno MFD8) I can set how this is to be done
There are three different modes

Cross Line: Change the waypoint when the vessel moves through an imaginary cross line (vertical line) that moves through the center of the destination point.
Circle: Change the waypoint when the vessel comes within the circle.
Cross Line & Circle: Change the waypoint when one of the above two conditions is met.

The radius of the circle is also adjustable.

Depending on the plotter this can explain what you see when switching waypoint.

I understand that, but this is not the problem. On reaching a waypoint the Autohelm recognises the next waypoint (distance and XTE are displayed sequentially on the Autohelm display) but the boat continues on its previous course with XTE steadily increasing for a few minutes before it slowly turns in the right direction.

Your other, much more in depth, points I'll look at later, I just wanted to clear a misunderstanding.

 

[johnphilip]

Still not the explanation, still puzzled

A solution to this is to command RESET CROSSTRACK ERROR (In Raymarine parlance) on the plotter immediately before TRACK mode on the A/P. What this does is to re-draw the rhumb line between your current position and the next waypoint. If you don't do this the first thing the A/P does after entering TRACK mode is to seek the rhumb line between the previous and next waypoints on your route. This command will zero your cross track error; but don't wait too long before hitting TRACK mode or it will build up again.

To the OP, I'd suggest trying reducing the WPT arrival warning on the plotter. This means you will be closer to the actual WPT before the A/P asks you to accept the new course.

The zigzag before settling on a course occurs if I am heading to a single waypoint with the plotter not just when following a multi-point route and by the time I am reaching the waypoint the boat is spot on the rhumb line. Even if I reset the course by starting a "goto waypoint" immediately before setting the wheelpilot to "track" the boat is taken a long way to one side then the other of the track before getting the precise direction.

 

[knuterikt]

I understand that, but this is not the problem. On reaching a waypoint the Autohelm recognises the next waypoint (distance and XTE are displayed sequentially on the Autohelm display) but the boat continues on its previous course with XTE steadily increasing for a few minutes before it slowly turns in the right direction.

Are you always seeing this behaviour or is it different depending on old/new course and the size of the course change.

I still suspect that it is the autopilot compass that is causing this.

When setting the WP the plotter send "True bearing to destination" to the AP.
The AP starts steering to WP using it's compass, if this heading is not directly to target the XTE starts to increase and new "course to steer" is calculated.

If there are no drift you should be able to compare heading (from the AP compass) with the COG and "True bearing to destination"
In the "wearing off course" phase I guess you will see that these are not aligned, would expect to see:
COG <> Heading = "True bearing to destination"

After settling in on new course this will probably change, expecting:
COG = "True bearing to destination" <> heading

 

[johnphilip]

I experience something similar with my Raymarine tiller pilot. If, in "auto" mode, heading straight towards the next waypoint, I switch to "track" mode, the pilot immediately moves the tiller, sending us off course. After a few minutes it gradually brings us back then tracks perfectly with XTE never more than 0.01 nm. So I don't see the second, other side, course deviation you report, but in other respects very similar.

I also find that if I use track mode to follow a multi waypoint route, that after passing a waypoint it allows a big error to develop before slowly condescending to close with the new route line.

I should add: no cross tides causing issues, and my coastal sailing involves legs far too small to worry about GC errors.

It sounds as thoughyou have as you say similar issues. I think it would be useful if I monitor some more trials on the water. The boat has been launched but I am not yet fully set up to go out and with this morning's weather it may be a while yet.

 

[johnphilip]

Are you always seeing this behaviour or is it different depending on old/new course and the size of the course change.

I still suspect that it is the autopilot compass that is causing this.

When setting the WP the plotter send "True bearing to destination" to the AP.
The AP starts steering to WP using it's compass, if this heading is not directly to target the XTE starts to increase and new "course to steer" is calculated.

If there are no drift you should be able to compare heading (from the AP compass) with the COG and "True bearing to destination"
In the "wearing off course" phase I guess you will see that these are not aligned, would expect to see:
COG <> Heading = "True bearing to destination"


After settling in on new course this will probably change, expecting:
COG = "True bearing to destination" <> heading

I will try to get more data from on the water trials soon, according to the Wheelpilot manual it just uses cross track error from the plotter.