Let's take a look at the galvanaic isolator again when installed in a typical installation from some years ago and see what happens.

This diagram shows shorepower operating some form of load. The load is the bit within the green box. It doesn't matter what, it could be an old battery charger, a table lamp, more or less anything. Nothng is connected to earth. Nothing changes. The galvanic isolator continues to do it's job of protecting against galvanic currents just like it was designed to do.

The problem these days is that most equipment isn't built like this any more. A lot of equipment uses switch mode power supplies which generate lots of radio frequency interference (RFI). Under CE regulations this RFI is not permitted, it has to be got rid of. One of the easiest ways to get rid of it is to divert it to earth. And this is what much equipment these days does.

So now we have to modify the diagram to this:-

The load remains within the green box.

Note the red box. This now contains two capacitors. These are components that conduct AC but not DC. The higher the frequency of the AC, the more they conduct (this is a very basic summary of what a capacitor does)

Their purpose in this case is to conduct any unwanted RFI down to ground where it is safely got rid of instead of it radiating everywhere and interfering with TV, Radio, mobile telephones etc.

The problem is that in order for this to happen, they can cause the galvanic isolator to conduct the RFI to ground, so the galvanic isolator is almost permanently conducting.

This almost certainly won't happen in saltwater as the parallel path of the water will conduct most of this AC to earth. However in freshwater it is a possibility.

The capacitors are thus trying to place an AC voltage on the earth conductor. Normally they cannot do this as it is earthed. But in the case where a galvanic isolator is fitted they can put a small AC voltage across the galvanic isolator. Just sufficient to make it conduct. This is a mixture of line frequency AC (50Hz in Europe) and the RFI from the switch mode power supplies.

These AC voltages are not so much of a problem for galvanic corrosion. But it means that the galvanic isolator is almost permanently conducting. And therefore it is not blocking the DC galvanic currents.

This has become more of a problem as more and more equipment has started to be built with these components in place. And today this includes just about every piece of equipment you can buy. Please note that for many years certain items have always had these capacitors in place. However in recent years just about everything uses them and they are also of a much higher value than they used to be which causes greater currents to flow, the equipment attempts to put higher voltages across the galvanic isolator and thus force it harder into conduction.

And the more equipment of this type is installed, the greater the problem becomes.

The capacitors used are of a very low value and the currents introduced are very small. In fact there are limits defined in CE, UL and CSA Standards for each item of equipment. But this doesn't prevent 10 pieces of equipment causing 10 times more current!

(As an aside it is interesting to note that the problems caused by these capacitors [i.e. AC currents flowing in the earth conductor] has become so great that in many instances, with several items of equipment, they are sufficient to cause RCDs to trip in the absence of any fault!)

For this reason, some galvanic isolator manufacturers have put a capacitor inside the galvanic isolator so that the capacitor conducts the AC currents thus preventing the diodes from conducting. This idea would work, if the capacitor was the correct size.

However, whilst this could be done to cope with the currents produced by one piece of equipment, it cannot be done for those produced by, say, 15 separate items installed round the boat. For technical reasons, there is a limit to the size that this capacitor can be. And it would be necessary to exceed this limit in order to keep the AC voltage below the "conduction" voltage of the galvanic isolator. Unless the AC voltage is held below this level, there is no point in fitting the capacitor to the galvanic isolator. It will not achieve anything.

The only conclusion is that if an installation has enough of this "modern" equipment to cause AC voltages across the galvanic isolator that are sufficient to cause it to conduct then the galvanic isolator is not doing it's job properly.

An important note is in order here. Some equipment is manufactured to a standard known as "Double Insulated" or "Class II". This equipment has no ground conductor in the power cord. They use a 2 core cable. This equipment cearly cannot put any RFI voltage on the earth conductor. Or can it?

Tests show that it can actually do so due to the capacitance in the AC wiring. The amount of RFI introduced by such equipment is very much less, but with enough of them connected could still be sufficient to cause a problem. Admittedly there would have to be a lot of equipment.

Further, if this equipment is also connected to the DC system (dual voltage equipment, battery chargers etc) then it is highly likely that a similar set of capacitors are fitted to divert the RFI down to the DC ground which will (or rather should) also be bonded to the AC ground. Again the effect will not be as great but it can still cause a problem and perhaps more than class II equipment that is not connected to the DC system.

Now someone is going to ask "is there a way to measure whether this is happening? How do I know if my galvanic isolator is doing it's job?"

And that's a perfectly fair question. Yes there is a test, of course there is. It isn't a difficult test however you do need to know what you are doing, you would need an oscilloscope and due to the fact that it involves testing on the AC system whilst is it live I cannot go into it here. Anyone with the knowledge and equipment to do such a test will know from this paragraph what needs to be done.

Finally there is one more issue that can cause problems for galvanic isolators. That of the incoming earth lead conductor having a voltage on it. Tests show that this is actually far more common than many people would believe. Not a dangerous (to life) voltage but a voltage sufficient to force a galvanic isolator into conduction. There are serveral causes for this voltage being present on the earth conductor but those reasons are not relevant here. The test is simply to measure the voltage between the earth conductor and the actual ground under your feet using a suitable earth rod. Any voltage present here is reducing the headroom (and thus effectiveness) of the galvanic isolator.

The only alternative is to fit an isolation transformer.

And just for those who think this all sounds very biased: We do not make, sell or install either galvanic isolators or isolation transfomers.

And finally, finally, we have been asked to point out one possible safety issue with isolation transformers.......

The ideal physical (from a safety point of view) position for an isolation transformer is on the quayside with the (now) isolated AC power feed running from the quayside to the boat. However in many cases this is simply not at all practical. This means the transformer has to be installed in the boat which leaves the problem that the incoming shorelead runs from the quayside to the boat and is not grounded to the hull.

Some argue that this defeats the whole argument that the incoming AC must be bonded to the hull. To a certain extent this is correct however the only possible point of fault is where the shorelead passes onto the boat and down to the transformer. Whereas normally the entire electrical system has to be contended with.

Without the isolation transformer there are possibly hundreds of different places around the vessel that could result in a live/hull fault. Behind plug sockets, in distribution panels etc. With an isolation transformer this is reduced to the feed from the shorepower socket to the transformer itself.

Remember that AC mains on board can never be made 100% safe. That simply isn't possible. We have to accept this and make the system as safe as is practically possible.