Problem with ver 4.0 and Raymarine autopilot

[janp391]

I have a PC running XP,OpenCPN, Raymarine a95 display and EVO autopilot. To connect these instrumen with the PC I use Actisense NGW-1-STNG. When I use the track mode on the autopilot with a route from openCPN I will get all of a sudden a quick turn of app 35 degrees (SB or port) and the boat will sail on that new course for app 10 sec. Then back on the original course. This will happen app. every 10 minutes. More often when the speed is low.
When using the autopilot no problem. Route from the a95 display in track mode, no problem. I have also ver. 3.2 on the computer in parallell. I then started the 3.2 version after shutting down the 4.o version. Then all was working normal.
I have tried different output sentence from Open (I use ENRMB, ENRMC and ENAPR) (M) or true, no diff.
Is this a known problem?
I can ofcourse use the 3.2 version but that is not a long term solution. I have used O from ver 1.3 and I want to follow the development.
I have no more ideas why 3.2 is working normal but not ver 4.0
//Jan

[janp391]

Some more findings
I see a difference how version 3.2 and 4.0 handle the track mode.
When starting to navigate a route both version do some turning to port and SB before they steer after the route line. When coming to a waypoint ver 3.2 and 4.0 give an alarm asking "Is it ok to turn" or something similar. Pressing ok on the autopilot ver 3.2 do the turn and continue to the next wpt. Ver 4.0 do some turns to port and SB and then continue to the next wpt
It looks like ver 4.0 start from scratch at every wpt in the route. Ver 3.2 do a better job in autopilot track mode without any turns when passing a wpt. What is the reason?

[bdbcat]

janp...

This is more a function of the A/P than OpenCPN.

OCPN simply sends route segments to the A/P, one-by-one. Any message you see on the A/P is coming from the A/P logic, sensing waypoint arrival.

It is possible that O4.0 is faster than O3.2, so switches to the next leg sooner than your A/P senses.

You may need to adjust the waypoint arrival setting on you A/P, or on OCPN to make them match up better.

Thanks
Dave

[janp391]

Dave
Thanks for your answer
I have the arrival circle on 0.05 on O in both versions. If 4.0 is faster is the right way to go down with that number?
Is there a way to "slow down" ver 4.0 to be identical with ver 3.2?
//Jan

[bdbcat]

jqn...

You must experiment. Every A/P has its own logic and timing.

I imagine that you want to increase the OCPN arrival distance, so that O switches to the next WP before the A/P detects arrival. Maybe...

Good Luck
Dave

[janp391]

Originally Posted by bdbcat

jqn...

You must experiment. Every A/P has its own logic and timing.

I imagine that you want to increase the OCPN arrival distance, so that O switches to the next WP before the A/P detects arrival. Maybe...

Good Luck
Dave

Have been trying "all" settings on the arrival distance and on the ap, but no luck. It seems that O givs the signal to turn and the ap makes the turn so quick that the boat over steer. There is no settings in the ap to correct this behavior. The ap is "learning" but in real life not this lesson.
My question is: Is it possible to add to O that the new information to the ap for the next wp can be slowed down by time or divided in smaller pieces?

[bdbcat]

janp....

When you arrive at a waypoint, and OCPN shifts to the next WP, how much true course change is required? I understand that the A/P oversteers at first. Does this happen when the required course change is small, or only when large? How large is large?

My Raytheon A/P has three sensitivity settings. If I make a large (say 30 degree) change in low sensitivity mode, it will oversteer and sometimes be unable to recover. However, it can handle a 10 degree change no problem.

In high sensitivity mode, it will oversteer but quickly recover on anything up to 90 degree course change.

Does your A/P have any kind of sensitivity control?

Dave

[janp391]

Originally Posted by bdbcat

janp....

When you arrive at a waypoint, and OCPN shifts to the next WP, how much true course change is required? I understand that the A/P oversteers at first. Does this happen when the required course change is small, or only when large? How large is large?

My Raytheon A/P has three sensitivity settings. If I make a large (say 30 degree) change in low sensitivity mode, it will oversteer and sometimes be unable to recover. However, it can handle a 10 degree change no problem.

In high sensitivity mode, it will oversteer but quickly recover on anything up to 90 degree course change.

Does your A/P have any kind of sensitivity control?

Dave

Dave

Thanks for your quick reply

It always oversteer but more if the change is large, say 30 degree then the oversteer are bigger and it will take longer time before the boat are on right course

Yes the ap (Raymarine EVO) has three levels but the behavior is more or less the same

Jan

[janp391]

Dave
Have been out sailing for six weeks and have used the ap and O 4.4 a lot. Has also experimented with “all” settings on the ap and in O. But it will not behave. For comparison I used the Raymarin chartplotter (a95) together with the ap. With this combo behaviour are quite different. When arriving to a wp the ap gives an alert (180 m to next wp and are counting down) when press ok the turning direction and if ok to turn comes up. When press ok the boat turns against the new steering course. The turning rate are app 1/3 compared against when O is doing the turn. If the turn is app 30 degree the turn will be done without any intermediate halt. When turning more the turning angel will be divided by 2. So first 50% of the turn, then wait a couple of sec, then the rest. With the Raymarine equipment there are no over steering. The turn is done in a controlled way and when the tun is done we are on the new course
Is it possible that O can do the same?
BR
Jan

[bdbcat]

Jan...

Thanks for the detailed explanation. I see what the RayMarine plotter logic is.
OCPN has no facility currently for splitting large A/P course change commands into smaller segments to minimize oversteer. This could be added as a Feature for OCPN5 if development resources allow.

I might be good to add this to the Feature Request list, so that it will not be overlooked in prioritization activities.

Hope you had a nice cruise.
Dave

[janp391]

Dave
Thank you for your quick reply
What Raymarine also do is that the turning speed is much slower than O. I think this is the most important differens between O and Raymarine. In that way the over steering is more under control
//Jan

[bdbcat]

Jan...

There really is no way to directly control the "turning speed" by NMEA commands to the A/P. All we can do is break the total required course change into several segments. More segments=slower turn rate.

Maybe Raymarine has a private NMEA message for rate-of-turn. Any way to capture the NMEA stream between plotter and A/P during a mark passage with large course change? Then we shall see...

Dave

[janp391]

Originally Posted by bdbcat

Jan...

There really is no way to directly control the "turning speed" by NMEA commands to the A/P. All we can do is break the total required course change into several segments. More segments=slower turn rate.

Dave

Dave
It is now in Feature Request list. I hope you will have the time, during the winter, to do a modification of the way O will "steer" the ap
//Jan