The Ghost on Artemis

One cool thing about setting your eyes on a little larger boat is that it is typically better equipped. It seems that more options boxes are ticked when the order was placed. 

If you are going to take a decent coastal cruiser and kit it out for blue-water off-shore sailing as live-aboards you will have to not only buy a lot of new hardware but also find the place for it. This is the excuse I delivered to my wife to get a little larger boat, it worked. 

You really only feel the size of Artemis when getting in and out or ports and this is still challenging us. But what would life be if we constanly did not push our limits just a little bit?

Under sail she is easy to handle and one gets very quickly addicted to the instruments, mainly the combined wind and autopilot.  Actually, the Raymarine P70 is the most recent installation by the previous owner. The autopilot itself is factory installed too, all hydraulic. I will talk later about why I will take hydraulic over the other alternatives any time. The p70 looks like this;

There are several display options. We use another one tothe one above that clearly shows true or apparent wind angle in conjunction with course and the autopilot. You can choose to sail at constant wind angle or course. Apparently you can also let the chartplotter take control for complete route hopping. This remains to be investigated. Can´t overstate how fast you get addicted to it and how useful it is when you have rookies at the helm, and they learn fast.

Now, Raymarine have gone through several generations of "networks" and as I discovered, or rather was forced to discover, was how these work.

Last year, as we left Corfu and motored north through the pass with Albania to the east to get out into the Ionian sea towards Italy, the boat icon was suddenly 180 degrees off. But only when not underway. Both chart plotters, the E80 by the helm and the E120 by the navstation showed the same error. The pic below is an E80, the E120 simply has a larger screen. One is choosen at startup as master and the other automatically becomes the slave. 

On the E80 below and in the map screen to the left, you can see the boat icon. Underway the GPS data is prioritized why the magnetic data is not used other than at very slow speed and at standstill. So the 180 degree icon issue did not affect navigation or the autopilot. 

BTW, the autopilot is used all the time. Fundamental seamanship holds that there is always a helmsman, I repeat always. If you violate that rule bad things will eventually happen. But nothing says you have to have your hand on the wheel all the time. In fact, it leaves you time to look forward, to check your surroundings and sense the ship, the trim and other systems.

So, both plotters showed the same backwards boat icon, hence not a fault with them. No it was not going to be that easy. I had no idea how all of this was interconnected. What I had was a some experience with NMEA0183. It was just an old and reliable protocol running in the RS232 standard. When we left Artemis for the season (in south Italy) I had taken some pictures of what it looked like behind the panels in the navstation and also in the starboard aft lazarette, where the autopilot controller and its adjacent gyro option was mounted. 

Above is the "Course Computer model 400/S3". From left to right is connected fluxgate compass, rate gyro (model SHS Gyroplus 2), rudder angle. Then follows the grey cable for NMEA0183 connected to "out" which probably feeds the E80/E120 plotters. Next to this are two ST1 cables connected. One of them is probably the GPS as it is located closeby on the deck. To the right in pic can be seen the ST1 hub with two plugs connected. Technically this is a non-ideal single point of failure installation. If the Course Computer goes belly up there is no compass and no GPS for the plotters. 

Factory installation of the fluxgate compass is on a bulkhead on starboard side behind couch in captains cabin. 

The current theory of what is the source are the follwing;

1. Compass is broken - still working but giving wrong data.
2. Autopilot processor introduces the error
3. Other device introduces the error
4. Calibration has gone bezerk

Out of these four the most interesting to investigate is #2 and #4. Calibration requires sea-trial, this has to wait. #2 can be tested by monitoring HDG data and detaching the fluxgate and rotating it and observing what happens. Testing #1 can be done by measuring the four wires with DVM.

Pic below is shot while visiting Artemis prior to purchase, you see the black boat icon amongst many AIS targets (purple). No problems.

Instruments at the helmstation with the later installed p70 at lower left. With hydraulic steering the rudder indicator is essential, it is the third instrument from left. It is essential in that with hydraulics very little rudder feel is given as feedback to the helmsman. There is some, but not much. Hence, you loose the feel for how the boat is trimmed. A look at the rudder indicator gives you visual feedback if your trim is being fought by the rudder or not. This is as important under manual sailing as it is with autopilot on. With it on you need to trim sails so that the autopilot does not have o work all the time. 

Back to the panel, in the middle, depth in feet, above it depth in meters...why? 

For the Raymarine expert it is also obvious that we have a mix of two or possibly three generations of SeaTalk protocols, Seatalk (ST1) SeaTalk 2 (ST2) and Seatalk-NG. This was not evident to me at that time. From a legacy aspect, Raymarine and Autohelm joined forces at some point so they inherited the different ways of these two manufacturers and it shows as you begin to dig in.

Little did I know that in fact there are five protocols active here, the aforementioned three plus SeaTalk HS and NMEA0183. ST-HS is in principle 802.3 or Ethernet. It is used by the radar and is exclusively fed to the E80 and E180. That´s nice, we can play with that later on. NMEA0183 looks like it is used for feeding the E80 and E120 plotters.

Artemis came well equipped with Raymarine manuals for most devices. In fact as manuals go they are excellent. Everything else you might need to know can be downloaded from Raymarine websites. What you can´t find there is on the web or can be gotten out of some talented service engineer. Invaluable.
Nobody has heard about the backward boat icon though.

One way to get a feel of the broad range of devices and how they connect is this brochure; 

https://manualzz.com/doc/22582993/multi-function-network-displays-m1500-marine-monitor-usb

(sorry can´t find it anywhere else any longer)

But, there was no drawing from Beneteau on how all these devices where interconnected.

I had that disturbing feeling growing inside that this was going to be a very deep dive. Having gotten some support by a very helpful Raymarine installer/service engineer in Stockholm by Janne at JG Marineteknik. He offered some and helpful tips on faultfinding, that is how to measure with DVM if the electrical signals where in within range. This will unfortunately only take you so far. Without knowing what devices talks to what and how, there is little chance of understanding where messages in the protocols are lost or altered. Typically you would want a complex network like this to have one core protocol and a device that interfaces and converts to all the other ones. All arranged as a star system both logically and electrically. The problem Raymarine faced during evolution (as most other manufacturers do) is that as their product range grew there had to be some simple means to allow backward compatibility. So, some of the instruments allowed for this translation. In essence they became links in a chain of communication and not just display devices. This works to a point, and it works quite well if the system onboard is properly documented. This rarely happens and Artemis is no exception. 

Raymarine also introduced a protocol converter, a grey box with ST1, ST2 NMEA0183 and NMEA 2000 inputs and outputs. It is the Raymarine DPU (TLA:d to "data processing unit") and is installed behind the panel at the NAV station. It is the second unit from the left. You would think that it would be utilized as expected, as it turns out it is not. By expected I mean that all data providers merge at the DPU and data receivers can read it. Also visible was another really annoying fact and that is the mess beneath when I took off the cover to the cable conduits. Inside where lengths of cables, odd and unprofessional connections, some hidden fuses etc. It was obvious that previos owners hade taken the short route when adding in new equipment.                       So that feeling of having no turning back soon grew out of frustration of not knowing what to look for and how to trace it, into a full realisation I had to map out the system.

Directly after christmas 2019 we flew down to Artemis to celebrate New Year italian style and have some leisure time in that region while also doing some boatwork. The weather was off and on and during a particularly howling windy night I took a look at the windinstrument (p70) at the helm but it was blank, no numbers. Neither wind direction nor windspeed. Got some numbers to verify with DVM from my friend Janne at JG Marinteknik to check the transducer up the mast, they where ok. That made som sense as BOTH wind and direction where gone. Unless the whole thing up the mast was had blown off, it would be unlikely (but not impossible) that this was the culprit. But looking up there, everything was where it should. Mind you, it had blown really bad in November, gusting to 70knots. The wind speed sensor is based on sensing magnetic pulses while the direction uses magnetic sensing of sine/cosine oriented (90 degrees apart) sensors. This is analog information sent down to a point where it is converted to data. The Raymarine device doing this job is a POD. It looks like what you see to the left. The white cable has five conductors, two for the wind speed, three for the direction. You can see that the POD is running by the green LED. The black cable is the ST-2 network cable (a CAN bus) that ends up in a junction box. I am looking at this POD with great suspicion. This is because it connects directly to the sensor via the mast cable, you only need to have a lightening storm close-by to generate impressive voltages in this cable and the first stop is this POD. 

So, what do these Raymarine connectors look like and how are they wired up? Pic below shows them.

From this we can draw some conclusions. ST(1) is single data wire and RS-232 like. By "like" I mean it is not following RS-232 as it is not a  point-to-point format but a type of bus format where you can have a number of talkers (transmitters) on same line. The protocol ensures that talkers listen on the line before talking (something humans can have issues with). 

ST-2 is spelled out clearly as being CAN and hence is RS-485 (or possibly RS-422). RS-485 is a proper bus format and can have many talkers and listeners.  CAN is a robust and widely (industry and automotive) used format. 

Then we have ST-NG and this too is CAN but now with a twist in the connector, it has an extra wire that carries ST-1.

So while ST-2 is not backwards compatible with ST-1 electrically, ST-NG is essentially a physical upgrade that carries both formats. ST-NG is also in essence the same as NMEA-2000, or N2K. In fact, the protocol and messages should (not verified)

be identical between ST-NG, ST-2 and N2K.

The blue plug titled "ST Backbone" is for connecting between nodes or between the splitters. Note that it does not include the ST-1 data in the physical transport any longer. 

Based on single handed pulling of wires and the logic from the Raymarine manuals the current version of Artemis system looks like this; 

It is for sure incomplete but probably not completely off. With the help of som friends pulling and tugging on the unmarked cables plus opening up the helm-station, I am confident we will map out the current system good enough. This was planned to happen in early April, ticket booked for the 6:th and home 11:th. From the news it became apparent the Italians where having problems. Hence, the week before we where due to fly down there, we cancelled. That was a decision not regretted, Italy closed down completely on the 10:th.

The question arises following all this if the system should just be repaired or if it is time to clean it up?

The reply to this question became obvious looking at the system and what it could benefit from. One thing is the embarrassingly messy wiring. I really can´t sleep well when it looks like this. Then one must ask the question if these ghosts are appearing is it not time for some exorcism? 

Before any of this can happen I need to know what has gone wrong. The odd boat icon is bearable, but the loss of wind data is not. What is required is to be able to snoop the various protocols? How can I do this? Electrically two things can be done. For the RS-232 like ST1 a serial-to-USB dongle is used. It looks like this; 

All I have done is attach wires with hooks to a male DB-9 connector that plugs into the dongle and I have a tool to snoop ST1 and NMEA0182. For ST-NG and N2K there are a few aftermarket devices and I have chosen the YDNU-02 from Yacht Devices. It plugs straight into the ST-NG network and has a panel type female USB at the other end. It looks like this;

There is a reason for the choice of this device, it allows interface to the ST-NG network for other purposes and I have plans to run Open-CPN on a separate PC at some point. This will let me do so. Another reason is that Yacht Devices also provides a complimentary CAN viewer application, link is here;

CAN Log Viewer Software

Open-CPN has the capability to set up your own "dashboard, that is, to pick incoming data from the USB/serial port and choose what and how it is presented. After getting the initial communications setting correct it works perfectly (to my surprise). A short visit to Artemis in February revealed that reading out NMEA0183 data from the DPU showed that heading was indeed wrong and that no wind was present either. 

The decision was made to modernize the network and make it ST-NG/N2K centric. This means the the system components that are capable to do ST-NG are used as such and other devices are routed to interfaces that uses ST-NG. No more permanent jury-rigging. 

The first step was to buy this, the iTC-5;

All of the transducers for speed, depth, windspeed/angle are wired to POD´s and these output ST-2. By using the iTC-5 all POD´s can be removed resulting in this data being directly fed to the ST-NG network. Looking at the Artemis schematic this should rid us of most of the ST-2 network.

Actual opening up of the helm-station displays and understanding what the E80 and E120 uses will define the next step. But for that we have to be onboard and for that to happen it must be possible to travel there.  

To be continued