VentureFurther - Nmea/Depth Sharing

[rgleason]

Moo, I've got your point. Will have to try gpsgate splitter.

Conachair makes a suggestion to use Node Red with Raspberry Pi 0/1/2/3 for a nmea recording device which I'm going to make.

[conachair]

Originally Posted by rgleason

Moo, I've got your point. Will have to try gpsgate splitter.

Conachair makes a suggestion to use Node Red with Raspberry Pi 0/1/2/3 for a nmea recording device which I'm going to make.

Hi, The node red route works well if you want to get at the data yourself later, store just SOG into a .csv file or whatever. For creating just a file of nmea messages then the VDR plugin seems as easy as it can get, 2 mouse clicks and off it goes.. Why try harder?

[Moo]

Originally Posted by conachair

Hi, The node red route works well if you want to get at the data yourself later, store just SOG into a .csv file or whatever. For creating just a file of nmea messages then the VDR plugin seems as easy as it can get, 2 mouse clicks and off it goes.. Why try harder?

I find VDR unreliable it certainly seems to make OCPN unstable on my RPi based plotter so I uninstalled it completely. That is hard and frustrating to me so I find and easier, less frustrating, way.

I don't see why any one method should be treated has the holy grail. The point of my posting above was to show that there really is no such thing as VDR format and that VDR is not needed to log data for uploading to VF.

My pi plotter running Openplotter and OCPN was freezing every 8 to 18 hours on my last trip and that generally means that OCPNs data gets lost. So for reliability and duplication I am also logging the NMEA at source

As I mentioned above I don't run any plotter all the time while on passage but I want to record all the data from the instruments for later analysis so I just have a little box running that logs the raw NMEA.. it's windows and happens to use GPSGate but it could be anything else.

Here's the wind plot from my last leg.

[transmitterdan]

A lot of RPi issues are related to the power supply. The RPi board does not have sufficient bulk capacitors to handle the current surges generated by modern ARM processor cores. If the external 5V supply cannot supply minimum 2.5A current bursts (prefferably 5A) then the RPi will freeze up under certain conditions. I suspect 99% of RPi freeze up problems are power related. The reason disabling some software helps is the extra software threads causing the current spikes go away and thus the freeze problem goes away. But it wasn't the software's fault. For example, I cannot build O on RPi with make -j3 because of this problem. But it should work with make -j4.

I am working on a board patch to fix this problem as I run into it all the time. When I have it tested will post in the open plotter forum and here.

[champ8242]

Originally Posted by transmitterdan

The reason disabling some software helps is the extra software threads causing the current spikes go away and thus the freeze problem goes away.

and... be sure about the 5V you use for peripherals. Give them their own power source if they draw too much from the RPi

Bart

[conachair]

Originally Posted by Moo

I find VDR unreliable it certainly seems to make OCPN unstable on my RPi based plotter so I uninstalled it completely. That is hard and frustrating to me so I find and easier, less frustrating, way.

I don't see why any one method should be treated has the holy grail. The point of my posting above was to show that there really is no such thing as VDR format and that VDR is not needed to log data for uploading to VF.

My pi plotter running Openplotter and OCPN was freezing every 8 to 18 hours on my last trip and that generally means that OCPNs data gets lost. So for reliability and duplication I am also logging the NMEA at source

As I mentioned above I don't run any plotter all the time while on passage but I want to record all the data from the instruments for later analysis so I just have a little box running that logs the raw NMEA.. it's windows and happens to use GPSGate but it could be anything else.

Here's the wind plot from my last leg.

OK, understand, so it's more of an rpi specific issue. I can't comment on running the vdr on a pi as I haven't really tried it much, though something sounds up your end, living at anchor I'll have a Pi3 running constantly for a week or more at a time before reboot with Opencpn up with no crashes. Maybe power? (i use a buck/boost dc converter set to about 5.2v from memory) As well as Opencpn on the Pi I've a load of node red saving lots to a sqlite database and serving node-red dashboards with graphs and data.
Solid as long as I'm not fiddling even using gqrx to record wfax from a fun cube dongle the Pi can just hold up but that's without Opencpn running, but ime once openplotter is stable on the Pi it seems to just run and run.

Ps. Really pretty graphs! What did you use to graph the data?

[Moo]

Originally Posted by conachair

OK, understand, so it's more of an rpi specific issue. I can't comment on running the vdr on a pi as I haven't really tried it much, though something sounds up your end, living at anchor I'll have a Pi3 running constantly for a week or more at a time before reboot with Opencpn up with no crashes. Maybe power? (i use a buck/boost dc converter set to about 5.2v from memory)

Ps. Really pretty graphs! What did you use to graph the data?

Yes the Pi is stable usually. I was using openplotter 0.14 as that was the latest version when I sailed. I posted on the OCPN forum with my experiences but I believe that there was a memory leak or something in the compass/sensor routines as the memory usage would slowly rise until things ground to a halt. Seems much better in the latest versions but I'm not sailing and generating a track at the moment. VDR has never really worked for me. I keep coming back to it and trying it but ...

Power supply is a buck boost of my own design/construction and more than enough power for the Pi and peripherals.

The plots were produced with a couple of simple scripts in SED/AWK to extract the NMEA data and plotted with GNUplot running under the Ubuntu BASH environment that is now available on Windows 10 in developer mode.

I can post the scripts here if anybody is interested, they are very basic and crude but work for me. Basically I make up my own NMEA sentence from the raw NMEA with DATE TIME, COG, SOG, TWD, TWS, AWS, AWA, DPT etc....

[conachair]

Originally Posted by Moo

Yes the Pi is stable usually. I was using openplotter 0.14 as that was the latest version when I sailed. I posted on the OCPN forum with my experiences but I believe that there was a memory leak or something in the compass/sensor routines as the memory usage would slowly rise until things ground to a halt. Seems much better in the latest versions but I'm not sailing and generating a track at the moment. VDR has never really worked for me. I keep coming back to it and trying it but ...

Power supply is a buck boost of my own design/construction and more than enough power for the Pi and peripherals.

The plots were produced with a couple of simple scripts in SED/AWK to extract the NMEA data and plotted with GNUplot running under the Ubuntu BASH environment that is now available on Windows 10 in developer mode.

I can post the scripts here if anybody is interested, they are very basic and crude but work for me. Basically I make up my own NMEA sentence from the raw NMEA with DATE TIME, COG, SOG, TWD, TWS, AWS, AWA, DPT etc....

Started the VDR recorder few hours ago on openplotter 15.1 / opencpn 4.8.0, I'll leave it on for a few days and see how it does.
If I don't crash the Pi fiddling with other stuff...

[conachair]

Originally Posted by conachair

Started the VDR recorder few hours ago on openplotter 15.1 / opencpn 4.8.0, I'll leave it on for a few days and see how it does.
If I don't crash the Pi fiddling with other stuff...

Well VDR seems stable, I can't get it to fall over anyway. Both on a raspberry Pi and Win10 laptop.
The saved file does get big quick though so for bandwidth limited cruising sailors some sort of message type/rate of saving filter additions to the VDR would be needed to keep the file size down for boats living away from fast wifi. Or (yet!) another plugin. It's doable quite easy in node red but I doubt if most boats really want to roll their sleeves up and get involved with that.

[rgleason]

Jon, this is a good description by Sean of his scripts for mbtiles. I think it might give you some good ideas?

http://www.cruisersforum.com/forums/....php?p=2175605

[rgleason]

Also for recording, as suggested below by Moo.

http://gpsgate.com/products/gpsgate_client

Also for windows users NavMonPC is excellent and can save a recording file.

I wonder if there are some good standalone opensource programs that can be used in linux and mac, of course there is red-node, as conachair pointed out.