Which galvanic isolator

[BartW]

As regards floating open-circuit induced voltages, in other non-boaty, faulty domestic mains electric supplies, the max I found was 27 volts. Was I lucky?

it depends how many amps that connection can deliver;
sometimes in power supply's from computers or other modern devices with 230V supply, there are capacitors between live and earth for RF suppression,
so when the that Power supply ground is floating, and you measure Voltage between a 230V live wire and the floating PSU ground, with a digital Volt meter, (high internal resistance) you could measure anything between 230V and 0V AV, but as soon as you "touch" that flouting ground, the voltage drops, / you might feel some electric, but not lethal.

many years ago, I was told that electric is lethal above 40V,
but only with a continious high enough current through your body,
not with a small leakage current from a RF suppression capacitor...

 

[Solent sailer]

when we were on a swinging mooring only visiting marinas, I didn’t bother with anything, now we have an aluminium boat and we are in a marina much greater risk and potential damage so have fitted the isolating transformer. also, the amount of electronics on the boat has increased with the potential problems this can cause. if we had a fibreglass boat on a swinging mooring i would go for a galvanic isolator they are cheap and you won’t be plugged in all the time, with a metal boat i would always go transformer.
have a look at: -
http://www.smartgauge.co.uk/galv_tran.html
there are some very good thoughts on the problems.

 

[Solent sailer]

The rating of the GI must match or exceed the power supply rating the boat uses. A GI can't sensibly be too big but can be too small. A 32 amp GI will safely cope with an earth fault current of 32 amps, whereas a 16amp won't.

Also be aware of false earth bypass to the GI along the neutral line. If you have any then the GI becomes useless.

I had this on my last boat where an LED had been used as a Earth to neutral indicator for reversed shore power polarity indication. The problem with an LED with its ballast resistor is that it provides a direct path to the neutral, which is earthed at the source point of the shorepower, for any stray galvanic currents. If you wish to have a polarity indicator then consider instead using a test plug that is plugged in when shore power is connected, then once the connection is confirmed correct the plug is removed and any earth GI bypass is removed with it.

It is also important that if you have a generator that the power connection switches both Live and Neutral lines, because the neutral should be grounded at the generator, so if left connected again this creates a false earth through the neutral line, again bypassing the GI.

Finally it is also worthwhile checking for +VE shorts to the bonding circuit, as these will also play havoc with anodes.

To do this turn all normal services off, which normally means the bilge pumps are still connected to power. Disconnect the +VE connection at the battery and measure for volts drop between the loose battery connection and the now bare battery post. Any reading other than 0 volts means there is a circuit to somewhere, and that a leakage current exists.

I now undertake this check once each year, and only when I had worked my way through several separate but equally culpable leakage / bypass defects did I finally get the anode wear under control on my last boat. This included failure of the above test which turned out to be a flapper type bilge pump switch that had let some saltwater in, providing a path for the 12 v +VE being switched by the flapper switch to leak into the salty water laying in the bilge and then into the bonding circuit. This made the bonding circuit very slightly +VE relative to the rest of the bonding circuit, in turn lighting up the anodes up stream of the leak connection point slightly more anodic than usual and causing them to fizz more vigorously. This was not a major battery leakage, but sufficient to cause accelerated anode wear, and nothing at all to do with the shore power and associated GI. Calder's Boatowner's Electrical & Mechanical Manual explains more of this.

Dont forget that you are trying to trip a 30mA RCD not the 32A MCB this will take (Type B mcb 5x 32 = 160amps or Type c mcb 10x 32 = 320amps) to operater in 0.1seconds. The galvanic isolator is only in the earth conductor and its capasity has little to do the the size of the supply.
Generator isolation this is very important as the neutral should be referenced (connected) to earth when running on generator without this your RCD and to some extent the MCB will not work, but clearly must be isolated when on shore power to prevent stray currents.

 

[maby]

Dont forget that you are trying to trip a 30mA RCD not the 32A MCB this will take (Type B mcb 5x 32 = 160amps or Type c mcb 10x 32 = 320amps) to operater in 0.1seconds. The galvanic isolator is only in the earth conductor and its capasity has little to do the the size of the supply.
...

Exactly - earth fault currents are not measured in amps or even tens of amps - if you get a genuine electrical failure on your boat, you will be pushing hundreds or even thousands of amps down the earth wire until the breaker trips! The GI on a 10m speed boat must be able to stand up to exactly the same conditions as the GI on a superyacht.

 

[superheat6k]

Dont forget that you are trying to trip a 30mA RCD not the 32A MCB this will take (Type B mcb 5x 32 = 160amps or Type c mcb 10x 32 = 320amps) to operater in 0.1seconds. The galvanic isolator is only in the earth conductor and its capasity has little to do the the size of the supply.
Generator isolation this is very important as the neutral should be referenced (connected) to earth when running on generator without this your RCD and to some extent the MCB will not work, but clearly must be isolated when on shore power to prevent stray currents.

It is important to understand the GI is a completely independent device.

So whereas your opening statement is correct, to follow your logic would mean the Protective Earth conductor itself would not have to match the rating of the supply, nor also a device introduced such as a GI, only the RCD. However the PE conductor should match the size and capacity of the supply, on the assumption that the RCD could itself become faulty, in which case the PE, now with its GI in series, must be able to handle the fault earth current until a safety device operates. This also means the GI diodes must be able to handle the initial surge current until a fuse or MCB operates to shut down the circuit.

Also what is to say that the fault current must be the full MCB rating of 32 amps, or even a sub-circuit rating of say 16 amps. If for instance a short occurred between the Neutral and Earth of say a 2 kW appliance (e.g. Some salty water drips in and forms a corrosive bridge) then the fault current could be up to ~9 amps (this could vary according to severity of the specific fault), which would be below the fuse rating of the appliance (13 a), leaving only the RCD to effect an isolation. If the RCD failed to operate, then the PE conductor would have to sustain the fault current and keep the entire boat's bonded system at least close to 0 volts. For this reason the GI must present minimal resistance and must be capable of handling the full supply current, and including an initial surge as the fault first establishes.

Therefore in the case of the OP he has stated this is a 32a supply, so IMHO requires a suitably rated (32a) GI, and not the more commonly found 16 amp type most boats will use. He could use two 16a in parallel, but I would prefer to see a single 32a rated GI.

 

[BartW]

have a look at: -
http://www.smartgauge.co.uk/galv_tran.html
there are some very good thoughts on the problems.

+1, this is indeed a very good description of the issues

 

[MapisM]

+1, this is indeed a very good description of the issues

If that description is correct, it's also rather worrying!
I mean, having just a fraction of the IT cost/weight/size/complications by fitting a GI instead is a very attractive proposition, of course.
But if the protection offered by a GI can be close to nothing (for all the reasons described there), and there's also no way to understand it ('cause I'm neither willing nor capable to fiddle with an oscilloscope to check that), well, that brings us back to the option of doing nothing at all, I reckon...

 

[maby]

If that description is correct, it's also rather worrying!
I mean, having just a fraction of the IT cost/weight/size/complications by fitting a GI instead is a very attractive proposition, of course.
But if the protection offered by a GI can be close to nothing (for all the reasons described there), and there's also no way to understand it ('cause I'm neither willing nor capable to fiddle with an oscilloscope to check that), well, that brings us back to the option of doing nothing at all, I reckon...

It is certainly true that a GI is extremely difficult to test and you really should assume that it needs to be replaced after any event on the boat that resulted in the circuit breaker tripping. No GI on sale on the leisure market can carry the full earth current corresponding to a live to earth short circuit for more than a short fraction of a second and the junctions in the diodes will be stressed far beyond their specified rating during that period. They may still test OK on a simple multimeter, but they could be so compromised that they cannot last long enough on the next event to trip the breaker - leaving you with mains voltages on the boat earth rail.

 

[kashurst]

The thing to remember is that the boats 240v earth to shore power is also connected to the boats anodes. The idea being in the absence of a shore power earth AND the boat is in the water any 240v fault to earth goes into the sea. So if the GI is compromised if the boat is in the water you still have an earth connection. Also the RCD in the 240v switch panel will switch off the power in 30 milliseconds - so enough to make you swear but very unlikely to kill anyone.
Also marina supplies are fused/have circuit breakers too.
Also how much current can an Isolation transformer stand before it is compromised.

 

[MapisM]

It is certainly true that a GI is extremely difficult to test and you really should assume that it needs to be replaced after any event on the boat that resulted in the circuit breaker tripping.

Do you mean that the GIs suggested by rafiki, of which there are versions incorporating a status monitor (see this webpage), are unreliable and/or not worth the additional cost - even if not much?