Choosing a gateway ?

[petit bateau]

Originally Posted by stevead

A few observations/questions.

There are two sources of GPS position data:
1. Standard Horizon GX2400 built-in GPS
2. Android tablet running OpenCPN.

When the Standard Horizon detects an external GPS source over either NMEA 2000 or NMEA 0183 it gives that source a higher priority than its internal GPS

When the Android tablet is taken below, it loses its GPS fix causing the following:
1. The Standard Horizon GX2400 loses its GPS fix (Doesn't display position or date/time).
2. The Raymarine I70s displays display a Lost GPS Fix alarm


I'm assuming you have connected OpenCPN on the Android Tablet to the Actisense W2K-1 using NMEA 0183 over TCP. (From the Actisense manual it appears as though the W2K-1 does not receive over UDP)

Thank you for contributing your insight. Yes, that is a good summary of what I think is the configuration and symptoms.

Originally Posted by stevead

First thing is to filter out Position Messages emitted from OpenCPN on the Android tablet to prevent them from appearing on the NMEA 2000 network.

Either set an outbound filter in OpenCPN for the NMEA 0183 TCP network connection.
a. For the "Transmit" option set a filter "Drop Sentences" for GLL, GGA, RMC, GSV, GSA. (You may want to confirm exactly which sentences are being generated by OpenCPN from the NMEA Debug Window.)

Or set a Rx filter in the Actisense W2K-1 to filter out the conversion of the NMEA 0183 position sentences (use the same sentences as above). Alternatively set a Tx filter for the NMEA 2000 messages, 129025, 129026, 129029 & 129540.

Thank you for those suggestions. I will try experimenting with one of those options. It may be a pretty steep learning curve .... so I'll likely not be back too fast.

Originally Posted by stevead

You can also set a Data Source option on the I70s displays, I'm guessing it's currently set to Auto. If both the GX2400 and OpenCPN are providing position data, the i70s should show the presence of two sources of NMEA 2000 position data; the GX2400 and the W2K-1 and you should select a preference for the GX2400.

When I delved into the Data Source option on the i70s display I was presented with two options: a) auto; b) NMEA-2000-wifi+ a system designator that I guess is the Actisense Y2K. I've since experimented further and when the Android tablet is switched off I only get option a). With the Android tablet switched on I get both option a) and b). I have indeed set it for Auto. I do not get an option c) that is for the GX2400.

My working hypothesis is that anytime the GX sees another GPS message passing over the NMEA 2000 it ceases broadcasting its own GPS messages.

Originally Posted by stevead

All of this being said, it seems like there are potentially a few bugs or at least some unexpected behaviour. I thought it would be reasonable to expect that if the GX2400 and i70s, when two sources of GPS data are available, if one source (the Android tablet) lost its fix, that they would switch to the other source. Perhaps contact Standard Horizin and Raymarine and ask about that behaviour, although I would tend to guess that they would suggest that your boat needs a stable GPS position source.

I agree. I will in due course contact both of them. However I am expecting much the same response as you are guessing.

In an ideal world I would be able to set the GX to put its own GPS fix as #1 priority. In this system, at the moment, it is almost certainly the best GPS data source in almost all circumstances.

But I'm not holding my breath, and so I suspect the real way I will have to fix this (if I cannot implement your filtering options) is by adding in another GPS source onto the NMEA 2000 network.

Originally Posted by stevead

Regarding AIS Alarms being displayed on the i70s when the GX2400 is switched off, this makes absolutely no sense. For the instruments to display an alarm, there must be AIS messages floating around from somewhere. You need to identify that source.

Yes, this utterly mystified me. There is literally nothing else connected that can source AIS messages. The only thing I could think of is that there is some sort of 'persistent cache' of tracks that keeps on alarming (on steady courses etc) if data sources are cut off. But I couldn't think where such a cache might reside. That was what caused me to throw my hands in the air at that point and type up the day's findings.

(That evening I then figured out how to permanently silence the AIS alarms in each i70s one by one. The Raymarine design team clearly live in a plotter-centric world, given their underlying GUI/HMI design assumptions.)

Thank you for your input. In the coming days I will experiment further.

[petit bateau]

Originally Posted by stevead

First thing is to filter out Position Messages emitted from OpenCPN on the Android tablet to prevent them from appearing on the NMEA 2000 network. .

I thought about your (welcomed and helpful) filtering proposals to stop the Android tablet GPS getting to the NMEA 2000 network. Getting up the learning curve on the ways you suggested seemed tricksy given other tasks and etc in hand here. Whilst thinking about this I hit on a different approach. I disabled location services on the Android tablet for the Android OPCN. That seemed to be a much simpler - albeit bruteforce - approach than to try and do selective filtering.

This worked, inasmuch as it allowed the GX2400 VHF-AIS-GPS to push out a stable GPS position to the NMEA 2000 network. Then both the OPCN on the Win10 and the OPCN on the Android tablet would run.

So --- there are two instances of OPCN running, which happen to pull their information from a common NMEA 2000 network via a common Y2K wifi gateway.

So .... in the last several days I did a few hundred miles and tested out this setup, which is very much a proof of concept for OPCN for me. I have some further findings and observations, and would welcome feedback.

I am pottering about in the Ionion/Peleponese. So some of the time this is marine traffic central, much like the Solent, especially when I am near the busy charter area and ferry ports. At other times I am in the less travelled bits with scarecely another vessel in sight. I am pointing this out because I have been getting various forms of OPCN crashes and I was wondering if there is any correlation between the crashes, and the external circumstances.

On my first day everything went smoothly with both OPCNs behaving fine, until just outside Preveza I noticed that the Android had stopped tracking, and soon after the Win10 had as well. I tried stop/start tracking and that did nothing. I tried rebooting both machines and also the entire NMEA network which seemed to work (but is not a great thing to be doing). It took me a while to figure this out and so the red historic track data is missing for a section. Then I looked at the Win10 and noticed that a spurious track had been generated extending a few miles ahead of my boat in a straight line. This was obviously false as it is in an area my boat had yet to get to.

Here is a screengrab I did at that moment on the Win10 device. The Android was not showing this suprious track, it had simply stopped tracking. On each occasion I found I had to shut down both devices and to power down the NMEA network entirely to restart things.



This kept recurring over the next several days, often at the most awkward moments. This one has some truly bonkers multiple false tracks going off in different directions.



I took those screengrabs on the Win10 laptop as that was easiest for me. Sometimes both devices would simply cease tracking, without these spurious tracks appearing at the time. Looking at the lower left corner of the OPCN on Win10 the /|\ rotator would keep rotating. One could seemingly interact normally with OPCN (both Android and Windows) except that there would be no tracking going on. Sometimes I would realise within a few minutes that this 'crash' had happened, at other times it might take me a little longer to realise depending on what else was happening onboard.

These are very insiduous crashes as one has no good and obvious real crash indication. On the Android it would (I think) switch from the red ownship indicator to the (stationary) grey ownship indicator, with no forwards path indication - so that was a clue that a crash had taken place. On the Win10 it would sometimes go to (stationary) greyship, and at other times stay as (stationary) redship with forwards path indication - so less reliable as an indicator. In those circumstances it is only when mk1 eyeball tells me that the realworld picture is different than the Win10 OPCN picture that I can be absolutely sure a crash has happened.

Sometimes I could go practically a whole day with no crashes. Sometimes multiple crashes in quick succession. Always seemed to be simultaneous crashes. Sometimes crashes near land/harbours sometimes seemingly in the middle of nowhere with very little traffic around me.

On one occasion there was a crash that was in fairly open water with very little traffic, except for a fast 20kts commercial roro that passes astern. That got me thinking about whether either device had tried to hop onto a wifi network on that ship (it will have had an open access wifi for the lorry drivers), and in doing so whether either / both devices had dropped ther Y2K gateway feed from NMEA. But surely that should not generate suprious tracks, and crashes, even if it were to take place.

Then I got this gobsmackingly amazing crash where I took photos of the screens as I figured nobody would believe me:























As best I can see both instances of OPCN have decided they are somewhere in the sands of North Africa (21N, 000E), yet the NMEA network is quite clearly showing correct realworld position of (38N, 020E) and so too is the actual source, i.e. the VHF-GPS-AIS receiver. I took a few sets of photos of different occasions of crashes here, sorry if that confuses anybody.

What on earth is going on ?? Where on earth did that position come from ? Why does OPCN believe it, and then carry on believing it ?

Anyway there came a moment yesterday midday that the OPCN system was crashing every few minutes. At that point I made a decision to switch off the Android tablet completely in case there was some sort of unstable interaction going on between both OPCN devices. (Actually I switched both off for a few hours, I am after al quite happy with paper charts. But then I switched the Win10 back on). For the remainder of yesterday (and all of last night) the remaining one device (the Win10 laptop) has been stable.

So ..... is there a problem with running two OPCNs on one system ?

Any thoughts as to what is going on ?

[petit bateau]

For the last couple of weeks I've been using OPCN with only the Win10 laptop connected to the ship's NMEA 2000 network via the gateway. The Android tablet has been disconnected from the gateway and although it is running OPCN it is isolated from the NMEA 2000 network and the Win10 laptop.

The Win10 laptop / OPCN has been mostly solid and stable for the last two weeks.

Here is a screengrab in the Win10-OPCN of the last 4-5 weeks. I have highlighted in red some of the tracks that were spurious and causing system crashes (heading off into the Sahara, etc). All of those tracks came from the period before I removed the Android tablet from the network. Some of them however were from the period when the Android was on the network and had its location services (i.e. its internal GPS) enabled, and others were from when its location services were disabled.

As you can see that is a reasonable amount of pottering about, enough for testing purposes.



So I seem to have a stable OPCN-Win10 proof-of-concept solution.

I am still unsure why the addition of the second instance of the OPCN-Android causes these crashes.

This morning the Android tablet ( a Lenovo tab2 A10) died. It had been running OPCN in standalone mode, and nothing else at all, for the last two weeks. It was proving quite useful for anchorwatch purposes as it meant I could power down the laptop and radio and NMEA network overnight on anchors, thereby saving a 3+ Amps draw. But for whatever reason the tablet died, investigations still ongoing, I suspect it has a battery failure. It was brand new, bought just for proof-of-concept testing of OPCN.

Neither OPCN device was able to run the Watchdog app. Nothing, nada. I've put up a separate post asking for help on that but received no comments. I did notice that when I went to close down OPCN on Win10 after trying to run the WAtchdog app that I then got an unstable shutdown, and an invitation by Windows to restart OPCN in safe mode next time. I ignored that and have carried on using OPCN as usual, but Watchdog is just not starting. I wonder if that has anything to do with running OPCN 5.9.0-beta.

I still got no access to the Android OPCN support forum.

So, overall something of a mixed bag. There is a promising piece of software here with OPCN, but it is really struggling to get to market. And the lack of reference designs / implementations mean that any newbie is inevitably in a world of one-off unique pain unless they happen to have a pal with a functioning stable system they can copy.

The irony of OPCN running stable on Win10 is not lost on me.

[bdbcat]

petit bateau...


"I still got no access to the Android OPCN support forum."


What is your login name on that forum?


Thanks
Dave

[petit bateau]

Originally Posted by bdbcat

petit bateau...


"I still got no access to the Android OPCN support forum."


What is your login name on that forum?


Thanks
Dave

I sent you a message, regards, pb

[bdbcat]

pb...


Sorry, no message received. How sent?


Dave

[petit bateau]

Originally Posted by bdbcat

pb...


Sorry, no message received. How sent?


Dave

resent properly, my bad !

[rgleason]

@petit bateau

Neither OPCN device was able to run the Watchdog app. Nothing, nada. I've put up a separate post asking for help on that but received no comments. I did notice that when I went to close down OPCN on Win10 after trying to run the WAtchdog app that I then got an unstable shutdown, and an invitation by Windows to restart OPCN in safe mode next time. I ignored that and have carried on using OPCN as usual, but Watchdog is just not starting. I wonder if that has anything to do with running OPCN 5.9.0-beta

You seem to have a misunderstanding about this program. It is opensource and powered completely by Volunteer Programmers and Users lending their skills and time. You mention something about "struggling". OpenCPN has been available for quite some time and is used by many knowledgeable voyagers. Indeed you as a User are helping with your posts about issues you have. Welcome to the community of contributors!

Regarding Watchdog. I have been trying to maintain it for Sean, with others help. Recently I have been busy on other personal things and more important OpenCPN efforts such as rewrite of the User Manual. So I have been ignoring the 20 plugins plugins that I have been trying to maintain for others. Sorry.

Please try to be patient with us, we'll get to it eventually.

What version of Watchdog is this? and, sorry, I have not read the full thread, what is the OS and OpenCPN version please? See https://opencpn.org/wiki/dokuwiki/do...support-report

[petit bateau]

Originally Posted by rgleason

@petit bateau


You seem to have a misunderstanding about this program. It is opensource and powered completely by Volunteer Programmers and Users lending their skills and time. You mention something about "struggling". OpenCPN has been available for quite some time and is used by many knowledgeable voyagers. Indeed you as a User are helping with your posts about issues you have. Welcome to the community of contributors!

Regarding Watchdog. I have been trying to maintain it for Sean, with others help. Recently I have been busy on other personal things and more important OpenCPN efforts such as rewrite of the User Manual. So I have been ignoring the 20 plugins plugins that I have been trying to maintain for others. Sorry.

Please try to be patient with us, we'll get to it eventually.

What version of Watchdog is this? and, sorry, I have not read the full thread, what is the OS and OpenCPN version please? See https://opencpn.org/wiki/dokuwiki/do...support-report

1. I'm well aware it is open source and has a long history, that is why I am taking the trouble to a) ask questions before blundering ahead; and b) document the issues I find; and c) make suggestions; and d) be patient ...

2. If I had been able to get stuff to run smoothly then maybe I'd have helped with the basic user manual. But I couldn't, so didn't. I do help in other places with other stuff.

3. Both are latest revs from download catalaogue. Draw = 1.8.49.1 and Watchdog = 2.4.110. Note Win10 OPCN = 5.9.0-0-beta. This affects both Watchdog on Win10, and Watchdog on Android. The Android OPCN is whatever the latest paid download rev is.