Another Seatalk question? " Sorry "

[xeitosaphil]

I am looking at my ST60 instrument cable feeds and the Chart Plotter, because I want to add my Auto Pilot into the system. They seem a bit random at present and it has made me look at the overall picture, which in turn has made me come up with a query.

As all my ST60 instruments and Auto Pilot data feeds will be daisy chained, if one unit goes down will I lose all the interface between all the other existing units down line from the failed unit?

Is there any way to protect all the existing units from going down if one unit in the daisy chain fails?

Would it be possible to create a ring main as such to prevent failure of all the units exising after one in the daisy chain goes down, or is this just a no no, and something you just have to except?

I know it is possible to ring main the power side, if over a certain number of units, but is it possible to ring main the Data side for protecting the continuity to all the existing units.

As usual thanks for any replies


Philip

 

[prv]

In general you can wire original Seatalk "chaotically" with branches all over the place and it will still work. I don't know about rings specifically, but I don't see why not. It doesn't have terminators as such, just treats the whole data line as one common conductor.

Seems at least worth a try.

Pete

 

[xeitosaphil]

In general you can wire original Seatalk "chaotically" with branches all over the place and it will still work. I don't know about rings specifically, but I don't see why not. It doesn't have terminators as such, just treats the whole data line as one common conductor.

Seems at least worth a try.

Pete

That's actually what I mean Pete, " treats the whole data line as one common conductor. "

So if there is a break in the line, the continuity is lost between units. I was thinking if it was possible to feed the data from both ends of the daisy chain, one unit going down wouldn't then be a problem, as the supply data would come from the other end supposedly ?

Perhaps I'll give it a try like you say. I don't suppose it can do any harm can it?

Philip

 

[Niffler]

Three wire basic Seatalk is like your 12v system on board, it doesn't mind if you have units all over the place. I use a three bus bar system, all units including the autopilot run back to the three bus bars except a couple of ST60 units that are daisy chained.

 

[nigelmercier]

There are two scenarios that I can think of that could cause failure:

A wire falls out of a socket, so the rest of the chain goes down.
A unit develops a fault that holds the data line high or low.

The first can be made less likely by completing the ring for the power connectors, but don't do it for the data line.
The second is unlikely, but I have a unit to fix with this very fault.

 

[Billjratt]

I don't remember seeing posts on here about failed Seatalk circuits (Although Nigel has just come up with one!) In general it seems to have been a robust system for years.
However, if you want to chop up the nice moulded cables you could connect them in a 'star' configuration, electrically it's the same, as the input and output are linked in the units, not re-driven. It would add an unprotected joint to the circuit though. A ring circuit would also work if you've got the extra cable and join them after the fuse. In any of the above cases, a single instrument failure can still bring down the whole circuit, so I don't see any advantage in messing with something that works.

 

[xeitosaphil]

" but don't do it for the data line "
.

Is there any special reason for this Nigel?

 

[prv]

I don't remember seeing posts on here about failed Seatalk circuits (Although Nigel has just come up with one!)

I also had one fail when somebody (not me ) put a screw through the cable . The Compass needle span round a few times and then caught fire, fortunately nothing else was damaged. As long as the broken Compass was on the bus, though, nothing else worked. I guess it was holding the line high or low as Nigel suggests.

Pete

 

[nigelmercier]

Is there any special reason for [not completing a data ring] Nigel?

Raymarine say not to, but don't explain why. It's probably because doing so will cause the two data paths to conflict due to pulse timing.

 

[xeitosaphil]

Thanks for all the replies,

Ring, and Daisy Chain it is then

Cheers Philip :encouragement: