Signal K Implementation

[bdbcat]

All....
Thanks to big effort from Hakan, we now have signalK support for the standard OCPN Dashboard plugin. We have implemented support for most existing Dashboard instruments, using the appropriate sensors visible to, and as reported by, the signalK server.


Pre-built OCPN images are available as usual. Testing requested on all platforms, with as much variety of sensors as are available.


Hakan will describe a test technique in which the signalK server on rPI may be driven by a VDR file to simulate high volume and diverse signalK traffic. A neat idea for testing...


Kudos to Hakan
Dave

[Hakan]

All..
Dave are to generous. Thanks.
My method to test Signal-K streaming is as follows:
All devices are on the same network router.

Signal-K:
It's installed on a RPi4 using Openplotter v 2.0.22 (Installation tips: https://pysselilivet.blogspot.com/20...all-guide.html )
Sign in to Signal-K server and add a new connection like my example:
Type: NMEA0183, ID: localhost, NMEA0183 source: UDP, Port: 10112 (Shot 1)
Restart the server.

On a Win10 with OCPN v. 5.0.522:
To feed SK with data I can't use VDR on OCPN. That would create a data mix mismatch.
Instead I use TDan's nice Python script to play a VDR file. See: https://github.com/transmitterdan/VDRplayer
This is, thanks to TDan, a handy tool.
I use any of my saved VDR files where most common NMEA data are available and transfer the data stream to Signal-K. Here are a couple of VDR files I've used. Welcome to Sweden and Croatia. (The latter I found here on CF years ago. Thanks to the originator): https://www.dropbox.com/s/989kmuz5lo...files.zip?dl=1
Save the txt file where you want and use a cmd-window to start the replay. My example to broadcast the entire network:

C:\Documents\Python\VDRplayer-master\VDRplayer.py C:\Documents\VDR\VDR_Hakefjord.txt 192.168.10.255 10112 0.1 UDP

Check on the Signal-K server for incoming data, Shot 2.

Then open OCPN Connections tab to automatically find and connect to the Signal-K server. (One tip so far - Beta-test-issue! The Signal-K server connection procedure needs some more tweaking. Take it easy when activate or deactivate the connection. Especially when deactivate and activating another source. Let the "Apply" function finalize before check and activate another source. The config file may be corrupt if you hurry to much.)
Open Dashboard and add instruments of you choice for the test. I've no Mac so a MacOS test would be appreciated. Pls report any issues.


Funny stuff:
Open a forward port on your router to Signal-K's LAN-IP port 3000 and you can monitor you data stream from any Internet connected OCPN-PC. Shot 3.
Share your WAN-IP address with friends?
Having a wlan router on board? Then you can share your data with your close buddy boat also having a wlan.

Thanks
Håkan

[verkerkbr]

Originally Posted by bdbcat

All....
Thanks to big effort from Hakan, we now have signalK support for the standard OCPN Dashboard plugin. We have implemented support for most existing Dashboard instruments, using the appropriate sensors visible to, and as reported by, the signalK server.

Pre-built OCPN images are available as usual. Testing requested on all platforms, with as much variety of sensors as are available.

Hakan will describe a test technique in which the signalK server on rPI may be driven by a VDR file to simulate high volume and diverse signalK traffic. A neat idea for testing...


Kudos to Hakan
Dave

Tested here on my Linux Mint Tricia AMD 64 bits. Linux Sparky and Tricia on 32 bits and the Raspberry Pi 4. All work great with the SK instrument panel.

However I have not so much external sources available tot test here at home.

Distance total gives no indication. But it is still early days.

Progress all around.

Regards,


Bram

[Hakan]

Originally Posted by verkerkbr

Tested here on my Linux Mint Tricia AMD 64 bits. Linux Sparky and Tricia on 32 bits and the Raspberry Pi 4. All work great with the SK instrument panel.
However I have not so much external sources available tot test here at home.
Distance total gives no indication. But it is still early days.
Progress all around.
Regards,
Bram

Bram..
If you like you could play TDan's VDR player on Linux as well. (Python3....)


What instrument did you use for distance?
Which NMEA message included your distance data?
My example: shot
Thanks
Håkan

[conachair]

Originally Posted by bdbcat

Pre-built OCPN images are available as usual. Testing requested on all platforms, with as much variety of sensors as are available.

Sorry, not sure of the link - is the SigK dashboard included in the V5 release download?
https://opencpn.org/OpenCPN/info/downloadopencpn.html

[CarCode]

No, that is from March 2019. We have 2020 already.

[transmitterdan]

Originally Posted by conachair

Sorry, not sure of the link - is the SigK dashboard included in the V5 release download?
https://opencpn.org/OpenCPN/info/downloadopencpn.html

No, it is only available in the unstable testing downloads. Use at your own risk...

https://cloudsmith.io/~david-registe...able/packages/

[conachair]

Originally Posted by transmitterdan

No, it is only available in the unstable testing downloads. Use at your own risk...

https://cloudsmith.io/~david-registe...able/packages/

[conachair]

Looks very good so far, well done gang!!Most of this is from the SKsim SigK app.

Barometer seems to expect maybe hPa instead of SI units?

VMG looks like it works, set Opencpn on the laptop to send to a port set up in signalk on a Pi on same network.

Any specific SigK paths you'd like to check? QUick to set up node-red to send as SigKpretty much anything.

[transmitterdan]

Please post bug reports here:

https://github.com/OpenCPN/OpenCPN/issues

[CarCode]

It does not matter if bugs are posted to flyspray or to github.
They don't get fixed either.
Since a long time there is a bug with course directions (180 degrees off) and nobody cares: https://github.com/OpenCPN/OpenCPN/pull/1668


Showing a course with 180° off is a severe bug for a navigation software in my humble opinion...

[cagney]

Originally Posted by CarCode

It does not matter if bugs are posted to flyspray or to github.
They don't get fixed either.
Since a long time there is a bug with course directions (180 degrees off) and nobody cares: https://github.com/OpenCPN/OpenCPN/pull/1668


Showing a course with 180° off is a severe bug for a navigation software in my humble opinion...

That's just garbage...
Flyspray is for bugs in stable and beta releases.
Github is all about developing the next version, hardly even at alpha stage just now.
The author who made the "180 degrees off" error has pointed it out himself.
Nothing to worry about....
You are maintaining your own version of OpenCPN so I'm sure you could contribute to the code development, if you wanted to....
I'm looking forward to your first PR.

[transmitterdan]

Originally Posted by CarCode

It does not matter if bugs are posted to flyspray or to github.
They don't get fixed either.
Since a long time there is a bug with course directions (180 degrees off) and nobody cares: https://github.com/OpenCPN/OpenCPN/pull/1668


Showing a course with 180° off is a severe bug for a navigation software in my humble opinion...

Get up on wrong side of bed again today?

That PR is a whole 12 days old. It is going through review and testing.

All github issues are answered as fast as time permits. We could use more programming help but shade like this isn’t helpful.

[CarCode]

Originally Posted by transmitterdan

Get up on wrong side of bed again today?

That PR is a whole 12 days old. It is going through review and testing.

All github issues are answered as fast as time permits. We could use more programming help but shade like this isn’t helpful.

There is no reason to insult people.
Better have a look to the enormous list of orphaned issues and bugs.
This was only one example found by a user on 28 Nov 2019.

[teppokurki]

Please, very little to do with the thread topic.