Autopilot doubts??

[Corsair63]

Windows 11 Pro ARM64 ver 22631.2428 - 23H2
OCPN 5.8.4-0
Autopilot route plugin 0.4.30.0
Multiplexer ShipModul Miniplex-3E-2K (NMEA 0183/N2K/LAN)
Auto Pilot Raymarine Smart Pilot ST6002

I’ll try to be as much precise I could.
Connected autopilot via 0183 and USB but that is little old the autopilot and not transmitting latest version, so I was not able to get the rudder angle because NMEA string was quite old.
So connected to multiplexer via SeaTalk feature and multiplexer translate it to proper version of 0183 and no problem anymore with rudder angle.

But I was not able to control the autopilot with OCPN, because supposed to do a routing in the multiplexer for translate and send to SeaTalk out, routing in my device was somewhere complicated and most of the times I lost a lot of sentences, so at last I did it properly and I got everything I need.
Now I have directed output to autopilot via SeaTalk.

For spare I did a NMEA 0183 connection via USB adapter and connected to computer, defined in OCPN as COM port and filtering RSA sentences that not working because directly from pilot are obsolete and not processed by OCPN.
The data connection with multiplexer and all devices is done via UDP with output feature enabled.

Have to say that both connections are defined properly and working in OCPN, but for avoiding interference not used at the same time, only one of the features for each time. And in both are enabled the output sentences as EC.
RTFM of autopilot checked for proper function of Track mode and followed the procedure.

The procedure I’ve done with both connection in different times but same result, also, using the plugin AutoPilot route and unable it and done with the OCPN feature itself:
- Activated the route.
- In pilot display set as auto
- In pilot display set as track
- Waiting for some second and started beep and showing the proper course to waypoint.
- Click on pilot display track.
- And supposed to be following the route but the display is in track with a close course than the actual one ?????
- I cannot remember the NMEA sentences sent by OCPN to output, I guess that are in batches of 4 or 5 starting with $EC

At that time, I’m making tests with the vessel alongside, on next occasion sailing I’ll carry on the proper navigation test.

Well now in the past I was able to control the pilot with SeaClearII, and still have installed in my computer on board, so I used as IN/OUT COM port the NMEA connection direct with pilot in the first Comm device and used for inputting the second COM port from the multiplexer via (the spare data COM port in case not working the LAN for OCPN)

Procedure done in SeaClearII:
- Activated the route.
- Activated the NMEA output.
- Open the windows of Output data stream to check if is there data to pilot.
- In pilot display set as auto
- In pilot display set as track
- Waiting for some second and started beep and showing the proper course to waypoint.
- Click on pilot display track.
- And at this point I can always see the proper course to waypoint.
- The NMEA stream in this case is a repetition of this capture, only going to post one batch:

o	$ECRMB,A,0.217,L,01-ENFILACIOIN-MARINA,02-BOYA-15-RAOS,4326.039,N,00348.262,W,0.43,24.9,-0.0,V*47
o	$ECBOD,55.3,T,55.3,M,02-BOYA-15-RAOS,01-ENFILACIOIN-MARINA*35
o	$ECBWC,113036,4326.039,N,00348.262,W,24.9,T,24.9,M,0.43,N,02-BOYA-15-RAOS*00
o	$ECRTE,1,1,w,SEACLEAR,01-ENFILACIOIN-MARINA,02-BOYA-15-RAOS*5F
o	$ECAPA,A,A,0.22,L,N,V,V,055,M,02-BOYA-15-RAOS
o	$ECAPB,A,A,0.217,L,N,V,V,55.3,M,02-BOYA-15-RAOS,24.9,M,325.3,M*24

As far as I see there is slight differences in the sentences sent by OCPN and SCII, but why one is displaying the proper course and the other one is displaying some course close to actual one???

[bdbcat]

Corsair....
What do you mean by this?
"And supposed to be following the route but the display is in track with a close course than the actual one"


Course is wrong, so that ship drifts away from desired route" Or ??
Any possibility of a screen-shot or recording?



Thanks
Dave

[Corsair63]

Originally Posted by bdbcat

Corsair....
What do you mean by this?
"And supposed to be following the route but the display is in track with a close course than the actual one"


Course is wrong, so that ship drifts away from desired route" Or ??
Any possibility of a screen-shot or recording?

Thanks
Dave

Hi Dave,

moored at 286
in OCPN the course to WP is correct and, in the plugin (e.g. 313), too.
in the autopilot display when started beep receiving the signal of track (course to WP) the display is correct e.g. 313 press on track on autopilot and course displayed 288.
this happens with OCPN.

if I do the same procedure in the same circumstances, with SCII, after pressing track in the autopilot the display is 313.

????
I don't know what to say?
Carmelo

[bdbcat]

Carmelo...


Let us simplify.
Consider only OCPN base, without plugin.


Please confirm my understanding. Is all this correct?

1. You have a route on OCPN, the course to WP is 313
2. You activate route on OCPN
3. Wait for beep from autopilot.
4. Press track on autopilot.
5. Autopilot shows 288.


All correct?


We need to see what is OCPN sending to autopilot at this moment.
We can see this from OCPN NMEA Debug window.


Will you try this, and show screenshot of NMEA Debug window?


Thanks
Dave

[Corsair63]

Consider only OCPN base, without plugin.
Done, unable Autopilot route and left only O feature.

1. You have a route on OCPN, the course to WP is 313
Yes and displayed in both screenshots.
2. You activate route on OCPN
Yes, also displayed in screenshots.
3. Wait for beep from autopilot.
4. Press track on autopilot.
5. Autopilot shows 288.

done a video with it in two part and added extension .pdf, remove it for both of them in the same directory and open the .001

All correct? followed instructions.


We need to see what is OCPN sending to autopilot at this moment.
We can see this from OCPN NMEA Debug window.

attached log file done with VDR_plugin, one with all sentences and the other with stripped $EC sentences.

Will you try this, and show screenshot of NMEA Debug window?
all $EC sentences marked blue, so, it means that are accepted and sent.
Edit: attached for the sake of good order

I've had the engine out of order for some days and the weather is not fine at all by now, so I'll test this feature when we'll do the sea trials for the engine in the next days, weather permitting, and I'll see if tit is something related to movement.
but the strange thing is that SCII keeps the autopilot course as the waypoint bearing as should be.

Thanks, Carmelo

[bdbcat]

Corsair...
Thanks for the excellent report.
I will be looking and thinking about this today.


More later
Dave

[Corsair63]

The previous test is done with devices connected on this way:
autopilot via SeaTalk > Multiplexer
Multiplexer translate to NMEA > UDP > OCPN
OCPN > UDP > Multiplexer translate to SeaTalk > Autopilot.

I have a spare connection direct NMEA:
Autopilot NMEA > COM port USB > Computer (OCPN) and opposite path for sending sentences to autopilot.

I do not use both at the same time.

so, I tried the same procedure via COM port.

everything as previous message, when press on track in the autopilot the bearing is received well (as the video) it beeps and show 313, I press Track and it keeps 313 but beeps every 2 seconds, for confirming the track??.

attached NMEA string complete and stripped $EC sentences and also a screenshot of the NMEA debug window.

[Corsair63]

Hi Dave,

sent you a PM with the link of the original video that shows the right part of it.

[bdbcat]

Corsair...
Here is a theory:


1. OCPN does not send $ECRTE sentence.

2. OCPN core autopilot logic uses $ECXTE to control the autopilot.
3. The way ECXTE control works is that the autopilot steers to minimize the XTE. It makes small corrections, a few degrees at a time, and measures the XTE to make course changes.
4. I think what is happening is that the autopilot sees that your head is 286. It needs to go to a target of 313. It tries to adjust the course to 288, as a small and smooth turn. But there is no change, because you are moored. So nothing happens.


Results of this theory:
The OCPN core autopilot should perform correctly when you get underway, and autopilot course commands actually do something good.


It is a theory.
And this is the way my autopilot works.


Maybe?
Dave

[wholybee]

Originally Posted by bdbcat

Corsair...
Here is a theory:


1. OCPN does not send $ECRTE sentence.

2. OCPN core autopilot logic uses $ECXTE to control the autopilot.
3. The way ECXTE control works is that the autopilot steers to minimize the XTE. It makes small corrections, a few degrees at a time, and measures the XTE to make course changes.
4. I think what is happening is that the autopilot sees that your head is 286. It needs to go to a target of 313. It tries to adjust the course to 288, as a small and smooth turn. But there is no change, because you are moored. So nothing happens.


Results of this theory:
The OCPN core autopilot should perform correctly when you get underway, and autopilot course commands actually do something good.


It is a theory.
And this is the way my autopilot works.


Maybe?
Dave

This is how my autopilot works as well. It really only works if you are moving, and you manually get on course to get XTE to 0, and then engage the pilot. What makes this hard is that the autopilot display shows XTE in nm, so you can get it to 0.00 nm, but still be about 50 feet off course. Otherwise, the pilot turns to intersect the route at about 45 degrees (or more), and the shoots past it, and starts doing "S" turns. I think this is bad logic in the raymarine pilot, that doesn't start to straighten out until it crossed the route, instead of easing into it.

[bdbcat]

Wholybee...


Mine is a 20 year old Raymarine. It has the same "capture" logic, but my boat is heavy. So the capture works a bit better. Surprisingly low overshoot on moderate (< 45 degree) course changes.


It is a non-trivial -problem. In this mode, there are three servo loops running...


Dave

[Hakan]

Another factor may complicating this for old APs. My Simrad WP30 has of course a really outdated WMM table that seems to create trouble when a new course is initiated. Until XTE corrections are made he's off course.

[Corsair63]

As soon as I would be able to do the sea trials with the boat, I'll let you know how is the behaviour of the autopilot in track mode with O.

of course thanks for your time.

[Corsair63]

Today I've done the sea trials for the engine, and I've briefly tested the autopilot feature.
I couldn't do it in deep as I was in narrow channel and have not the track for that area, it was only an engine check.
I did it with the output OCPN 0183>Multiplexer (translate to SeaTalk) > autopilot.

as soon as I activated the route, the position was stuck in that point, but all other data (course, depth, temp, etc) were properly updated, as soon as I deactivate the route, after some seconds the position was updated in place.
I did it twice, I also checked the priorities in O inputs, etc but I didn't see any suspicious thing.
on the next trial I'll perform with direct connection USB to COM 083 > autopilot if would happens same behaviour.