Sea Talk Mystery

[ALPHORA]

Strange happenings with our electronics. The system is all raymarine, with a C120 plotter linked to ST60 instruments via Seatalk. The instruments are log, depth wind, and 4000+ wheel pilot, radar, and an ST70 repeater. The heading alignment on the autopilot has taken to resetting itself to apparently random directions. This also causes the boat icon on the plotter to wander off the true boat heading, which I suppose makes sense. I can realign the heading on the wheel pilot display and it stays put for varying lengths of time before wandering off. When it does wander I have seen the heading displayed count up or down by varying amounts until it settles on a new value, where it will stay, again for unpredictable times. Changing the boat heading is mirrored by the display, even if the numbers are rubbish, ie a change to port or starboard by a given amount is seen on the autopilot display. On its own this would point to a fault with the 4000+ control unit which I have no idea how to check. Yesterday the depth display changed to look just like the autopilot display with heading, standby etc. Pushing buttons on the depth display had no effect, turning all the instruments off and on again changed it back to a depth readout. The wind instrument also gives wayout readings occasionally, over or under reading by large values. The speed display is also prone to giving odd values, and this appears to be caused by the calibration factor being reset. I have checked all the connections on the system several times and all appears sound, visually and with a multimeter. I have also isolated each instrument and they appear to work fine , but it is difficult to be sure as the times between the odd values appearing can be a few seconds to weeks, but the autopilot errors do seem to be cropping up more often, at shorter intervals. None of the instruments has been added recently, all worked happily together for about 4 years. The batteries are all good and the voltage has never been below 12.8v when the faults appeared. Charging is by a combination of wind and solar when we are underway. I have tried turning everything else of an electrical nature off and it seems to make no difference. Is there something obvious I have missed? Time to get out the Walker log and sextant? Help.

 

[AngusMcDoon]

It sounds like one of the items on the network is going mad occasionally. If the problem is reasonably repeatable then you could try disconnecting items one at a time, but if it only happens every few weeks, it's tricky. May be I could knock together a Seatalk logger with timestamps so when it goes bonkers you can have a look at the logs and see what messages you have going round.

 

[ALPHORA]

Angus, the idea of a data logger sounds superb, if you have time to oblige I will be forever indebted. Were you thinking of something to transfer data to a laptop for analysis? I have gone as far as I can with my limited knowledge of digital electronics. I was getting to the point of swapping the heads for new, starting with the autopilot, the idea of spending more money than necessary has caused more than a few sleepless nights.

 

[AngusMcDoon]

Well that would be the easiest. Read it in from a serial port into a terminal program and use the terminal program to lob it into a file that is large but untimately not a very thrilling read. There are terminal programs (better versions of HyperTerm) that will timestamp and are free to download. There are some sample circuits on Thomas Knauf's site that allow raw Seatalk data to be read by a PC serial port. The bits can be bought for a few £ from RS and made up on Veroboard.

Once you have it all recorded, in particular when it goes titzup, you can look at the data using Thomas Knauf's site to see what is ok and if there is any rubbish. If you have the ability you could write a quick VB app or Python script that parses the data into a more human readable form.

Of course, after all this, it may be that all you know is that there is junk on your system, but not where it is coming from.

 

[Plevier]

Your initial fault sounds very much like I had recently. Cured by cleaning all the connections to the fluxgate compass including the ribbon inside it that pushes into a connector.
You can do resistance checks on it - values from Raymarine website look up compass faultfinding or similar in the FAQ's or I can dig up the values if you can't find.
No idea about the further weird happenings sorry.

 

[laika]

Well that would be the easiest. Read it in from a serial port into a terminal program and use the terminal program to lob it into a file that is large but untimately not a very thrilling read. There are terminal programs (better versions of HyperTerm) that will timestamp and are free to download. There are some sample circuits on Thomas Knauf's site that allow raw Seatalk data to be read by a PC serial port. The bits can be bought for a few £ from RS and made up on Veroboard.

Once you have it all recorded, in particular when it goes titzup, you can look at the data using Thomas Knauf's site to see what is ok and if there is any rubbish. If you have the ability you could write a quick VB app or Python script that parses the data into a more human readable form.

Of course, after all this, it may be that all you know is that there is junk on your system, but not where it is coming from.

Andrew G has posted about some windows software he's written which reads seatalk: Maybe he has something to help. I have some stuff for linux/MacOSwhich will parse seatalk and add a timestamp (though prints the raw seatalk, not human-meaningful): Could be handy if you want to stick a raspberry pi behind the bulkhead and forget about it until you have a problem. I believe Andrew G's stuff might need a "real" serial port. Mine doesn't, but if not using a "real" serial port does require a USB to serial converter which handles parity errors correctly (not al lof them do). Another curve ball: According to Frank Wallenwein, the circuit on Thomas Knauf's site has problems at higher data rates. Presumably when things go a bit mad high data rates may be experienced I believe that NigelMercier posted an "improved" circuit a while back.

 

[AngusMcDoon]

Another curve ball: According to Frank Wallenwein, the circuit on Thomas Knauf's site has problems at higher data rates. Presumably when things go a bit mad high data rates may be experienced.

I wonder what he means as the data rate is fixed at a very lowly 4800 baud. This won't be the OP's problem though.

 

[laika]

I wonder what he means as the data rate is fixed at a very lowly 4800 baud. This won't be the OP's problem though.

Fair enough: He's contributed here before so maybe he could clarify? "Presumably" should have read "Possibly": If something is going wrong it may be pumping out data fast, ignoring collisions...or not

 

[ALPHORA]

Thanks for all the suggestions . I will have a look at the fluxgate compass next time I am on board and see if there is anything amiss then try to understand the replies regarding data capture. The one that really confused me was when the depth display changed into an autopilot display. I know all the ST60 units look the same apart from the button labels, they must have a lot in common on the inside as well. Presumably something tells the unit to behave in the correct mode to match the labels on the buttons?

 

[Plevier]

Here are the test instructions. Mine was at the top of the resistance range after cleaning everything when measured from the instrument end of the connecting cable. I think to see the values below you would need to measure on the ribbon cable in the compass.
Be careful inside the unit - the ribbon cable is delicate!
Best way to clean its contact surfaces is with a good quality pencil eraser.

Raymarine Fluxgate Compass Test (Multi-meter required)

Multi-meter should be set to:
(a) Resistance scale
(b) 200 ohm scale
The following procedure should then be followed:
1. Disconnect the transducer from the system.
2. Connect one multi-meter lead to the red wire and the other lead to the green wire and a reading
as indicated in the test data table should be present.
3. Follow the above procedure for all the color combinations indicated in the test data table.
4. If all readings match those indicated assume the compass transducer is functioning correctly.

Compass Test Data:

Compass
Cable Colors
Resistance
Measurements
Comments
Red to Green
4 ohms
Measurement is taken between the two colors



Red to Yellow
4 ohms
As above
Green to Yellow
8 ohms
As above
Screen to Blue
8 ohms
As above
All other combinations
Open Circuit


You are looking for a short or open. A +or - 2 Ohm reading is allowed.

If this check indicates a fault with the compass then you will be best served by replacing the compass. If no problem is indicated then you should send the compass and the course computer to a repair facility for evaluation.

 

[AngusMcDoon]

The one that really confused me was when the depth display changed into an autopilot display.

I didn't know they could do that. If they can I would have expected that the mode they should operate in would be flashed in at the factory and should never change. I can't think why it should be a configurable setting at a later date.

 

[AngusMcDoon]

Update to this after a bit of asking around: Yes, they all contain all the software to be any of the instrument heads. There is no network command to change a device from one type to another - it is set permanently in the factory what a particular instrument head should be. If it is changing from one type to another on occasions then the hardware is failing and it should be repaired or replaced. This might point you where to look first for your problem.

I didn't know they could do that. If they can I would have expected that the mode they should operate in would be flashed in at the factory and should never change. I can't think why it should be a configurable setting at a later date.

 

[ALPHORA]

Thanks for that Angus. I will have a look at the depth unit as well as the compass. I have gone back through the log and the only electrical issue of note (that I had forgotten about) was a couple of years ago when I saw that all the instruments and the plotter started flashing on and off rapidly, shortly after leaving the berth. I eventually traced this to the main battery earth cable being fractured inside the still intact insulation. I wonder if the rapid turning on and off for a few minutes upset things?