Wiring and installing an isolation transformer - Page 1.

Isolation transformers are installed into boats to isolate the vessel "ground" from the actual "ground" (i.e. the earth) in order to prevent galvanic corrosion. There are further explanations of the reasons here and here.

Assuming you have made the decision that an isolation transformer is what you intend to fit, then it is necessary to install it correctly so that it, in order of priority, A) Is safe, B) Does the required job of isolating the hull from the shore earth and C) Is reliable.

Well surely everyone agrees with these points, so why should there be any confusion about the best way to wire one up?

Well, it seems to come down to risk assessment. AC mains electricity can never be made 100% safe. That is an impossibility. So we have to make it "as safe as possible". And it seems this is where the confusion lies.

Everyone agrees about how to wire up the primary (to the shoreline live and neutral), every one agrees about how to wire up the secondary (to the boat live and neutral). The different viewpoints come about with regard to what to do with the earth connections.

Firstly there is the question of whether the secondary side should recreate a neutral-earth bond (thus imitating what we get from the national grid).

This is quite simple to answer. If no RCDs (Residual Current Device - known as GFCIs in USA) are fitted, then it is safer to not bond the neutral and earth on the secondary (output) side. If RCDs are fitted then in order for them to operate, the neutral-earth bond must be made.

As many installations will be subject to regulatory guidelines/legislation it is often a requirement that RCDs are fitted. That being the case the neutral-earth bond has to be made. They simply do not operate as designed without the neutral-earth bond.

From the safety point of view, it is debatable which system is safest (and again with 2 schools of thought), RCDs and neutral earth bond, or a floating system i.e. one without the neutral-earth bond. This is a subject for a different webpage! For the record we believe a neutral-earth bond with RCDs offers the best protection.

This leaves the question of what to do with the earth connection and safety screen on the transformer.

Let's have a look at the isolation transformer with the primary and secondary already wired. Note that we have made the neutral-earth bond in this case as most installations will be using RCDs and thus require the bond.

The question is: What do we do with the safety screen/chassis connection in order to ensure the best safety?

Note that this will only ever affect the outcome in the event of a fault. In the absence of a fault, it makes no difference whether the safety screen is connected to the shorepower earth, the boat earth or indeed left disconnected.

It seems that in order to ensure we connect the safety screen to the best place, we will have to look at possible faults that may ocurr, and connect the screen to whichever point ensures the best safety, in the majority of cases. This really is the best that we can do.

So let's list some possible faults, in general, then later we can look at them in further detail and try to ascertain which are most dangerous, which are most likely etc.

1. Actual transformer faults. i.e. short or open circuits between various parts of the transformer.
2. Wiring faults. i.e. loose wires in the installation, short circuits in the wiring, open circuits etc
3. Water in the transformer or it's enclosure.

These cover more or less all possible faults that can ocurr with an isolation transformer.

Now, as a practicing engineer, I can state quite categorically that an actual transformer fault is far less likely than either of the other two. There is absolutely no question in my mind about this. I doubt, very much, many people would debate this point.

So this leaves the other two possible types of fault. The relative probability of these depend upon the type vessel. For instance water inside the transformer enclosure is probably more likely in a small ocean going vessel than a large one or a vessel on the canal system. Faulty wiring is probably more likely on a vessel owned and maintained by a private individual than by professionals (some people would debate this!).

Rather than argue this point any further, let's give them an equal probability. Water in the enclosure or faulty wiring have roughly the same probability of ocurring and both carry a much greater probability than a fault in the actual transformer itself.

So the first faults (i.e. the most likely) that we need to guard against are faulty wiring and water ingress.

Let's look at the transformer again, note that this time there are 4 points labelled A, B, C and D. We have also connected the transformer safety screen/chassis to the shoreline earth.

Now imagine a fault as a result of a wire coming loose and each one in turn shorting to the transformer chassis at each of the four points A, B, C and D. Let's take them one at a time.

1. The wire at point A comes loose from it's terminal connection on the transformer and touches the transformer chassis.

You can instantly see that this will effectively short circuit the incoming live to earth. This will cause an enormous current to flow and blow the fuse on the shorepower socket.

2. The wire at point B comes loose from it's terminal connection on the transformer and touches the transformer chassis.

This will do nothing to blow the incoming fuse. If an RCD is fitted to the shorepower socket it will trip it becasue the primary current will now be returned down the earth connection instead of neutral thus causing an imbalance between the live and neutral currents. This is what trips RCDs. Either way, there is no immediate danger presented to those on or off the vessel.

3. The wire at point C comes loose from it's terminal connection on the transformer and touches the transformer chassis.

This will do nothing other than the boat will no longer have AC electricity. The live output of the transformer is now isolated, the actual wire that has touched the case is dead.

4. The wire at point D comes loose from it's terminal connection on the transformer and touches the transformer chassis.

This will do nothing immediate. Electrically it is the same as the diagram above shows. The only problem is that the arcing may present a fire hazard.

Now let's contrast the above faults with what happens in exactly the same situtations but with the transformer chassis and screen returned to the boat earth. Here is the relevant diagram. Note that the only thing that has changed is where the transformer chassis is earthed.

1. The wire at point A comes loose from it's terminal connection on the transformer and touches the transformer chassis.

The hull of the boat now has 230 volts on it with respect to the actual ground outside. As this current has to travel through the water to return to earth, it is far from certain (particularly in fresh water) that sufficient current will flow to blow the incoming shorepower fuse. Obviously this situation is highly dangerous. If an RCD is fitted to the shorepower then this may well trip, but again it is far from certain. It is however highly likely.

2. The wire at point B comes loose from it's terminal connection on the transformer and touches the transformer chassis.

Apart from the boat losing AC power this will do nothing. If an RCD is fitted to shorepower it may trip. It may not.

3. The wire at point C comes loose from it's terminal connection on the transformer and touches the transformer chassis.

The boat will lose 230 volt power. No fuses will blow, no RCDs will trip. No danger exists.

4. The wire at point D comes loose from it's terminal connection on the transformer and touches the transformer chassis.

Loads on the boat will draw their current through this new (faulty) connection which may present a fire hazard. This is the only danger.

From the above it seems clear to me that to return the earth to the shorepower presents less danger than to return it to the vessel earth. Add up the possible dangers. There are more, and they are greater with the chassis earthed to the boat earth than when earthed to the shorepower earth.

With the transformer chassis earthed to the boat, there exists one scenario above where the hull of the vessel can become live. Obviously this is highly dangerous. No such danger exists with the chassis earthed to the shorepower earth.

To analyse the problem of water ingress, exactly the same procedure can be used but in this case instead of assuming there exists a short circuit, treat the fault as being a resistor. And also remember that the connection to the transformer remains intact.

Let's run through these.............