Boris and Chris:
You can use the 3g/4g Lowrance/Simrad radar without an interface box (RL10) and without a heading sensor. (But see limitations below).
To install one you basically just need to connect the wire from the radar to (a) power and (b) an ethernet jack. The wire from the radar has a standard RJ45 plug
on it so it's much easier to NOT use the provided RL10 interface box which requires you to do some wire splicing (because the network wire out from the RL10 box has a proprietary connector on the end that you would need to cut off).
If you do not want to use a Simrad/Lowrance MFD and just want to use Opencpn (ie the cheapest and simplest solution) then you do this:
Provide 12v positive through a switch and a fuse to the red and yellow wires. Provide 12v negative to the black wire. Ignore the bare wire (screen).
Plug the ethernet plug either into a laptop
directly or a hub (better). (The short power wires are inconvenient so you will probably want/need to entend them or extend the ethernet jack (by making, for example, a wire with a female RJ45 socket on one end and a male RJ45 plug on the other end).
If you plug into a hub it will probably just all work. The 4g radome uses multicasting so your laptop will "subscribe" to the multicast and receive the radar packets. You will not need to configure your laptop at all. However, note: (a) some linux
boxes seem to have permission issues with multicasting (two of mine do). If you can't see any radar image, try running opencpn as root - this is only for linux
of course. (b) multicasting only works over the wire not wirelessly - you need to be physically connected to the hub/radome. If you need a wireless radar screen
you might consider vnc to a wired box (c) DISABLE your firewall while you play with this - mine wreaks havoc on the multicast signal.
The issue of the NMEA2000 heading is a bit confusing. You do NOT need this in the above setup. The reason why you you do need to inject a NMEA 2000
10k heading signal into the RL10 interface box if you are using a Simrad MFD is because the radome itself does all the heading calculations for chart overlay (ie the radome has the intelligence not the MFD). Also the radome must have the heading signal to do MARPA. We can't do MARPA in OpenCPN and OpenCPN does the heading corrections itself so in this simple setup the NMEA
2000 heading signal isn't needed. All you need is the data feed from the radome.
Limitations: this all works really well with some clear limitations.
(1) Heading: OpenCPN takes the radar signal and lays it onto the top of the chart. To do this it has to know which way the boat
is pointing. It can take that information from the GPS
COG sentence and if the boat
is trotting along it works well. Once the boat slows down the COG jumps all over the place and so does the radar signal. If you are anchored the radar is essentially useless unless you have a heading sentence that you provide to OpenCPN in which case OpenCPN will use the heading sentence instead and things work. So...although you don't need to provide the radome with a heading sentence you really do need to provide OpenCPN with one - ie you need some kind of electronic compass
. (A separate thread might well want to consider the cheapest way to do this - my solution is python + 1 arduino uno + a $20 compass
from Pololu Robotics and Electronics
). In general, COG filtering/smoothing becomes an issue.
(2) As the plugin clearly warns, it is alpha software
. It is not stable. When I use it OpenCPN will frequently crash, especially when changing ranges. No doubt this will improve. This is not a criticism of the developers who have done a magnificent job - it is just a factor of the maturity of the code.
(3) The radome clearly has more capabilities than we can access with the plugin (eg MARPA, dual ranges, sharpening and filtering capabilities). You give up something without a Simrad/B&G/Lowrance MFD. I don't have one to compare so I can't say how much better the image would be with a MFD doing the tuning etc.
(4) The plugin controls are very limited. You cannot do much in the way of tuning the radar. The sea clutter/rain clutter controls/gain controls are currently clumsy - again this will probably change. (I wish I had the skills to change them and I hate to sound negative about the voluntary work for which I am so grateful). There also appears to be a conceptual error in the code (I haven't bug reported this yet because I'm not certain it's not my peculiar laptops): the range correction allows you to set a factor so you can line up the distance of an object that the radar reports with the distance on your screen
... but ... the 4g appears to need two correcting factors not one: one for ranges over 2NM and one for closer ranges. Eg on my machine the former is about 1.7 and the latter about 1.3. We have been calibrating using AIS
targets so we can be quite precise about the differences. Maybe it's different on the 3g.
(5) We've been testing the radar's capabilities on simulated night approaches and it's been very enlightening. We have a person below reading the screen and calling instructions to a helmsman and a person above corroborating or correcting those instructions. So far we have been able to enter some tricky anchorages
in simulated total darkness with a fair probability of surviving the experience. (We're not ready to try this in real darkness!) One "feature" of the radar can be disconcerting. There can be a target close to the boat that clearly shows at wider ranges (say 4NM) and vanishes from the screen at short ranges (say a fraction of NM). If you are approaching the target it is disconcerting to have it vanish. And this appears to be a software issue not a hardware
issue because it will reappear when you switch out to a wider range. Of course, it is most probably the radome's software and nothing to do with the plugin because the plugin doesn't filter the data.
Overall it has been wonderful to have the radar image laid on top of the chart instead of on a conventional heads-up display.
HTH. It took weeks to work this out so maybe somebody else can save some time.