Tactics Plugin

[rgleason]

≈ ~ looks like a minus sign on low font size and old eyes, if it is causing real issues for less tactically minded end users using some platform independent special format character might be the short term work around.

Ron I think this is an excellent point. I am guessing that that is why Petri chose ">". Is there some other symbol that is universal that would be a better "calculated" indicator?


I was also wondering if there could be a "?" Help or Hover display message of some type that could have a short explanation sentence for the formula, calcs or data sources to help new users. This may be more than Petri intends to do, but it would be a potential feature request to help definitively explain how the pi values work.



-PS: It might not be that hard to accomplish. I would be willing to help with the messages side.

[RonSouthworth]

Originally Posted by FrouFrou

Hi, after installing O5 it looks like Tactics is calculating the wrong leeway direction. I use the NKE heelangle option in preference. With a heel to port the leeway is to starboard. If I change the radio button on fixed leeway the problem is over.

Hi there thanks for sharing and providing feedback I am certain the developers will appreciate the feedback.

I always find it makes it a bit easier if someone share’s there first name people don’t seem to do it often on this forum but I hope to change that just a little. mine for the record is Ron.

There has been a lot of efforts going on with tactics thread at the moment. My part mostly is in doing some testing and porting some code over to the a project to aid in testing.

Could you be so kind as to including some other basic information.

Such as
What OS Version
what version of plugin have you been experimenting with

there is the latest official release and then there are quite a few versions being tested to validate / improve different aspects of the plugin.

Also worth identifying is what other options you have selected.

It is also helpful some times if you
Have captured the NMEA message traffic for a period of time and a snapshot of some of the log file and maybe your ini file for ov50.

These all help give clues of course to tracking down any potential problems.

I am interested at the moment so that so as to build up test cases for the improving the plugin function. As a future user under passage myself of course Im keen to help a fellow user of the plugin

/Ron

[FrouFrou]

Hi Ron, my name is Joost, my boatname is FrouFrou.


The problem started, as far as I know, after installing O5.0 with Tactics 1.008 for win(10 in my case).


My heelangel sensor from NKE produces a XDR nmea sentence which in the old situation with version 4.8 calculated the right Leeway. Now I noticed that it Leeway was oposite of the heel direction and with fixed heelangel it is good.


note; NKE updated the software of my heelangel because they used a differtent nmea XDR sentence with 'heelangle' in it instead of 'ROLL' after that it worked fine.


With kind regards,


Joost

[Canne]

Hi Ron, thanks for the bug reporting, I'm in the middle of the unit tests so everything (including discussion) goes to GitHub issues pages, please, for traceability and regression testing: For your bugs see :

https://github.com/canne/dashboard_tactics_pi/issues/30

https://github.com/canne/dashboard_tactics_pi/issues/33

https://github.com/canne/dashboard_tactics_pi/issues/31

Thanks again for reporting!

Petri

[Canne]

Hi Ron and Rick,

regarding the unicode character as active calculation/instrument identifier, I've captured these secondary requirements for Tactics module which you may want to discuss in GitHub, please. I am wrapping up and need to concentrate the requirement capture in one place for unit test traceability:

https://github.com/canne/dashboard_tactics_pi/issues/25

https://github.com/canne/dashboard_tactics_pi/issues/26

Thanks!

Petri

[AedanC]

Hi Petri,


Sorry for disappearing so suddenly when I was doing some testing. First I went to launch my boat and then got sick for a few days. I'd like to do some more testing so I have two questions:


1 Where is the latest version for testing? Preferably a version that does not log lots of debug stuff as I want to put it on the boat computer and it has very little disk space.



2 Are there any particular aspects you would like tested?


Cheers,


Aedan

[Canne]

Hi Joost,

Glad to hear from you! But sorry to hear about your issues with your NKE instruments. Mea maxima culpa! The version 1.007 was sluggish in O v50 (for some, like me unusable in Win10). That version of Tactics contains NKE-specific XDR-sentences interpretation (not the reason for sluggishness). In O v50 XDR sentences are interpreted as defined here: https://opencpn.org/wiki/dokuwiki/do..._sentences#xdr and this was ported at the same time with the updates in O Dashboard for the instruments. This is beautifully structured approach to the dodgy XDR-sentence. Should it go to Tactics? That can be discussed of course.

What you can do? Several possibilities, maybe in this order:

1. Use the excellent NMEA Converter plugin to convert whatever NKE is sending (see above for inspiration) into XDR transducer type angle (A) and transducer name "ROLL" sentence according to NMEA 2.3

2. Compile yourself a version just before the XDR-sacrilege was committed, you can find last NKE-aware version from here https://github.com/canne/tactics_pi/...1a7e6154090da1

3. Move back to 4.8.8 and v1.007 of Tactics

FYI, now I have a deviation fork of O's Dashboard in which I have integrated Tactics functions (from Tactics plugin): here https://github.com/canne/dashboard_tactics_pi/releases (i.e. I am not actively working on tactics_pi plugin anymore).

Maybe Thomas, the author of tactics_pi reads this and can give you more advice; me ,I have but one NKE display so testing is a bit difficult to me in order to help more than just reading from the code. Like this: in my current code I could not forget my NKE-owning competition in the sake of the fair play and I wrote

#ifdef _TACTICSPI_H_
                        else if ((m_NMEA0183.Xdr.TransducerInfo[i].TransducerName == _T("ROLL")) ||
                                 (m_NMEA0183.Xdr.TransducerInfo[i].TransducerName == _T("Heel Angle"))) {
#else
                        else if (m_NMEA0183.Xdr.TransducerInfo[i].TransducerName == _T("ROLL")) {
#endif // _TACTICSPI_H_

Try it, it may work for you without the NMEA converter... Let me know if it does. Or even if it does not.

Cheers,

Petri

[Canne]

Originally Posted by AedanC

First I went to launch my boat...

Hi Aedan, there can't be a better reason than launching your boat, can there?

Originally Posted by AedanC

1 Where is the latest version for testing?

This beta release https://github.com/canne/dashboard_t...s/tag/v0.2.019 is the one which fixes the issue with missing BoatPolar causing crash (the fix made the laylines disappear altogether if no polar, which is not needed but I've fixed that for the next release https://github.com/canne/dashboard_tactics_pi/issues/7 so no need to report that). Other open bugs and bugs which will be fixed for the next version you can find from here:
https://github.com/canne/dashboard_t...one%3Av0.3.001

Originally Posted by AedanC

Preferably a version that does not log lots of debug stuff... as I want to put it on the boat computer and it has very little disk space.

This version has some debug printing both for true wind calculation and for missing polar but they've done in way that they don't tell the same thing twice so it is typically only a 10 - 20 lines per run or per disconnection / reconnection of your instruments. So if you have an issue, search for dashboard_tactics_pi lines from your log file, it may give a hint what is happening.

Originally Posted by AedanC

2 Are there any particular aspects you would like tested?

Just use it and enjoy your sailing! I am doing now the unit tests (i.e. the boring testing of every single instrument and function, wondering how I could automate that in Travis... ). No reason to spoil your day at sea by staring at the dials - for now, I guess, it is important to have feedback on the usability (ergonomics) and long term run reliability. For example, if you arrive back to harbor late at night, you could try the night mode in O, all instruments should follow! If you have a polar file, you can activate some ^active instruments (i.e. calculated values) and tell what do you think about them.

Thanks for your help, have a nice day out!

Cheers,

Petri

[RonSouthworth]

Hi Joost,

Sounds like version 1007 might be worth a look, having said that 1008 looks to be the oldest version easily found?

I've not had any experience with your instrumentation type unfortunately,
however below is taken from Tom's document that I have uploaded hopefully it might help ...

NKE style NMEA Performance Records and export to the instruments:

NKE supports the upload of specific performance data to their instrument bus, which can be shown in their displays then. These records are polar based and unless you’re using their (quite expensive) regatta processor, this gives us an easy way to display e.g. the “Target Polar Speed” outside in the cockpit on the standard instrument displays.
Due to the lack of information on other manufacturers capabilities, I implemented this for the NKE system right now.

Be aware that OpenCPN can only import/export NMEA183 right now, but not NMEA2000 or SeaTalk.


The following 5 records are implemented :
Speed and performance target
$PNKEP,01,x.x,N,x.x,K*hh<CR><LF>
| \ target speed in km/h
\ target speed in knots
Course on next tack
$PNKEP,02,x.x*hh<CR><LF>
\ Course (COG) on other tack from 0 to 359°
Opt. VMG angle and performance up and downwind
$PNKEP,03,x.x,x.x,x.x*hh<CR><LF>
| | \ performance downwind from 0 to 99%
| \ performance upwind from 0 to 99%
\ opt. VMG angle 0 à 359°
Angles to optimise CMG and VMG and corresponding gain (available but to be verified)
$PNKEP,04,x.x,x.x,x.x,x.x*hh<CR><LF>
| | | \ Gain VMG from 0 to 999%
| | \ Angle to optimise VMG from 0 to 359°
| \ Gain CMG from 0 to 999%
\ Angle to optimise CMG from 0 to 359°
Direction and speed of sea current
$PNKEP,05,x.x,x.x,N,x.x,K*hh<CR><LF>
| | \ current speed in km/h
| \ current speed in knots
\ current direction from 0 à 359°
There is a new tab now in the Preferences where you can define up to 5 performance NMEA183 records to be created.




These 5 NMEA183-records all begin with $PNKEP and are created on the fly using the data calculated in the plugin and are sent to OpenCPNs NMEA stream. To send the records to your instruments, you have to define an outgoing connection in you Interface connections, e.g. like this :

• Set an output filter as shown above, filtering for PNKEP.
After set up, you should see records beginning whith $PNKEP, in your NMEA debug window.
NKE exports the $PNKEP sentences as soon as they’re available on the topline bus. Normally they’re calculated in their regatta processor and then exported to the PC.
• Therefore ignore all incoming $PNKEP sentences !
Click on Input filtering (see screenshot above), select Ignore sencences and add PNKEP.
• Don’t forget to re-init your NMEA data stream in your instruments, to make sure the new records are accepted.

Failing this being of any use, next step to help us help you will be if you can to data record some traffic off the network to help folks analyse what is happening, maybe a capture of the log file to assist in finding where the problem is.

Maybe transmitter Dan might have some clues if he is available, I think he did the performance improvements that resulted in V 1008 ?

/Ron

[Canne]

Dear Joost, a small correction from my side:

I was mistaken about v1.008: I allowed myself to peek into the tactics_pi author's repository (from which I presume the v1.008 is coming from) and I found from here:
https://github.com/tom-r/tactics_pi/...b3e71d1d7df939

in tactics_pi.cpp, this:

					// NKE style of XDR Heel
					if ((m_NMEA0183.Xdr.TransducerInfo[i].TransducerName == _T("ROLL")) ||
						(m_NMEA0183.Xdr.TransducerInfo[i].TransducerName == _T("Heel Angle"))){
						if (m_NMEA0183.Xdr.TransducerInfo[i].MeasurementData > 0)
							xdrunit = _T("\u00B0r");
						else if (m_NMEA0183.Xdr.TransducerInfo[i].MeasurementData < 0) {
							xdrunit = _T("\u00B0l");
						}
						else
							xdrunit = _T("\u00B0");
						SendSentenceToAllInstruments(OCPN_DBP_STC_HEEL, xdrdata, xdrunit);

If I look at https://github.com/tom-r/tactics_pi/...d7052708c51d57 which is the last one with patch level 7, it is exactly the same. Nothing has changed in this regard, between patch level 7 and 8, I was mistaken, sorry, this is not current Dashboard-like XDR intrepretation. It should work both with "ROLL" and "Heel Angle" transducer names, in both patch versions. You may also observe in the above code the logic of heel angle left (<0) or right (>=0) which has not changed, either. I remain puzzled about the issues you reported with v1.008.

Best regards,

Petri

[FrouFrou]

Hi Petri,

I tried it yesterday on another computer at home with the same setup (Ov5.0 and tactics 1.008) with a logfile I recorded played on VDR.

And quess, I could not reproduce it.

So it looks like the problem only occurs with input directly from the NKE system.

I use a NKE box USB datalog wifi 90-60-538 to get the nmea into Open CPN, maybe there is someting wrong there.

to be continued

Joost

[Canne]

Originally Posted by FrouFrou

I use a NKE box USB datalog wifi 90-60-538 to get the nmea into Open CPN, maybe there is someting wrong there.

Hi Joost, it is not the first time than a box fiddles with the streams passing through. FYI, you an upload here you recording: compress it first as .zip then rename it as myrecording.zip.pdf, upload it as attachment and we can try it out (and analyze the sentences).
Cheers,
Petri
P.S. of course, a recording which make it failing...

[RonSouthworth]

Originally Posted by FrouFrou

Hi Petri,

I tried it yesterday on another computer at home with the same setup (Ov5.0 and tactics 1.008) with a logfile I recorded played on VDR.

And quess, I could not reproduce it.

So it looks like the problem only occurs with input directly from the NKE system.

I use a NKE box USB datalog wifi 90-60-538 to get the nmea into Open CPN, maybe there is someting wrong there.

to be continued

Joost

Petri,
Ive not used the data recorder
Does the recorder plugin capture the raw stream?

Starting to suspect it is only capturing frame types it recognises and mot a raw capture hence why it worked on replay ?

Thanks for the feedback too Joost. If it is working on what was captured with the OpenCPN it might also be worth capturing the raw network traffic with wireshark ?

/Ron

[AedanC]

The VDR captures all received NMEA sentences regardless of whether OpenCpn can process them or not.

[RonSouthworth]

Originally Posted by AedanC

The VDR captures all received NMEA sentences regardless of whether OpenCpn can process them or not.

Thanks,

As stated I have not used the data recorder. Does it capture malformed invalid checksum messages ?

At times it is worth capturing raw traffic as this can find other problems that can occur on any network.

/Ron