Galvanic Isolator Installation, How?

[WotEver]

Just now, Markinaboat said:

Would this (galvanic corrosion) not only apply when connected to a shore line though? And if so. an isolation transformer or galvanic isolator would interrupt this process?

Nope. 
 

Alastair referred to “floating voltages on our earth (5-50v, but current too small to measure with my meter)”, which would be generated by the equipment on board. Some will be induced currents, some will be caused by bypass capacitors taking noise down to earth. Either way, the currents exist within the boat on the earth cable and as long as they’re not permitted to flow through the hull (by avoiding multiple earth bonds) they will do no harm. 

This has nothing to do with galvanic currents flowing from a shore lead. 

[Tony Brooks]

2 minutes ago, Markinaboat said:

Would this (galvanic corrosion) not only apply when connected to a shore line though? And if so. an isolation transformer or galvanic isolator would interrupt this process?

Steel is not a consistent material resistance wise throughout its volume, especially when there are welds in it so one can not be sure it won't be easier for the current to leave the hull to "jump" a higher resistance part of the steel. That will also apply to DC currents in the hell and DC would be worse because it only carries the eroded steel in one direction.

[WotEver]

One place I worked at had a measured 170V on an earth wire, but the current was so small that you couldn’t feel it if you grabbed the cable. Inductance over long distances can generate surprisingly high voltages, but they have to flow through a steel hull to cause any problems. 

[Keith M]

There is another issue which might need to be considered when installing a Galvanic Isolator.

There is alot of documentation that galvanic isolators are being activated 

By the voltages found on the earth cable within the distribution network on the shore side of the galvanic isolator

[canalboat]

OK so just as I thought I understood this earthing thing, you electrickery folk have me confused again.

What you seem to be saying is that if I run my extension lead to the boat to use for working electrical tools etc, I must earth it to the hull. AND as my original question was "do I need a G.I." (I use the initials because I am not sure how to spell the words)  to which the answer was "no",   as it MUST be earthed to the hull (and not back to the building RCD where I plug in the lead) the answer is "yes". 

Is that correct?

And by the way, to the person who said I should have a consumer unit on the boat, my extension lead ends with a 13amp socket equipped with an RCD, so no I don't.

[Alan de Enfield]

49 minutes ago, canalboat said:

OK so just as I thought I understood this earthing thing, you electrickery folk have me confused again.

What you seem to be saying is that if I run my extension lead to the boat to use for working electrical tools etc, I must earth it to the hull. AND as my original question was "do I need a G.I." (I use the initials because I am not sure how to spell the words)  to which the answer was "no",   as it MUST be earthed to the hull (and not back to the building RCD where I plug in the lead) the answer is "yes". 

Is that correct?

And by the way, to the person who said I should have a consumer unit on the boat, my extension lead ends with a 13amp socket equipped with an RCD, so no I don't.

There is a big difference to running an extension lead from the 'house' to the boat and just plugging in your drill / angle grinder etc,

 

To

 

Running your shoreline into the boat, into a 'distribution box' and powering your own '230v AC ring main' and sockets.

[ditchcrawler]

12 hours ago, TheBiscuits said:

 

Are you sure?  That sounds exactly opposite to the advice I'd give, which is have ONE mains earth stud on the hull, as close as possible to the ONE 12/24V DC ground hull connection but on a different stud.

I'm with you on this one

2 hours ago, jenevers said:

Is it possible to test the resistance of canal water?

Yes, but its not something I have done since my apprentice days at college,  I remember a Wheatstone bridge came into it somewhere and lots of wires run out over the college car park and stakes driven in

[ditchcrawler]

12 hours ago, Alan de Enfield said:

 

 

 

 

 

 

 

 

The EN spec actually says the Earth for the AC and the 12v Negative SHALL (must ?) use the same stud.

 

 

 

 

And is what I have done

[ditchcrawler]

2 hours ago, WotEver said:

One place I worked at had a measured 170V on an earth wire, but the current was so small that you couldn’t feel it if you grabbed the cable. Inductance over long distances can generate surprisingly high voltages, but they have to flow through a steel hull to cause any problems. 

At home we have a double bed electric blanket, If my side is off and Diana's side is on we get an electric tingle if we touch when laying in bed, likewise if mine is on and hers off. Its just a tingle on touching, once contact is made there is nothing.

 

I demonstrated to a chap that if you took a coil of 3 core cable, put a plug top on one end, just connecting live and neutral and powering it up with a good DVM at the other end you could read 230 volts on the earth wire

[Keeping Up]

3 hours ago, jenevers said:

Is it possible to test the resistance of canal water?

When another Forum member and I did some testing a few years ago on my boat (steel, blacked, in dirty canal water) we found that the impedance through the water to shore was about 3 ohms. At that time there was an externally-generated voltage of about 0.7v rms present on the shore line at an impedance of less than 10 ohms.

 

If you would like to read the full report, read on, otherwise skip from here:

---

Chris and I made a number of measurements on Keeping Up yesterday: most of them were very useful, and one of them possibly turns conventional thinking on its head by suggesting that other people’s mains-powered equipment presents you with a greater corrosion hazard than does your own equipment.

Keeping up has a fairly conventional setup; the shore-line earth connects via a GI to the ring-main earth which is also connected to the hull. The hull is floating in the muddy waters of the GrandUnionCanal but the piling next to the boat is made of plastic. The readings we took are typical, but naturally wide variations are possible on other boats in other waters.

First we separated the three earths (ring main, hull, and shore); there was no equipment connected to the ring main and we even powered the scope from the inverter to ensure that it did not affect our readings. The only unexpected reading was a potential of 700mV rms (50Hz ac) between the shore line earth and the hull, with a source impedance of less than 10 ohms. Since there was nothing on the boat that could be generating this voltage, we concluded that its source was the shore line. This triggered the thought that a voltage coming from the shore line earth could drive a GI into conduction and make it ineffective.

Next we connected a 100K resistor from ring main live to ring main earth. This of course raised the potential on the ring main earth to 240v ac (don’t touch anything !!!) until we reconnected the shore line earth to the GI. We now had 2.4mA ac flowing through the GI and into the shore-line earth. As expected the GI was driven into conduction and we observed a square wave of about 2.5v peak to peak across it, from ring-main earth to shore-line earth. Nothing unexpected here, it demonstrates that a GI cannot work unless the water provides a parallel path around it; in the absence of any parallel path the GI is forced into conduction by small leakage currents. If your boat is out of the water the GI will not have this parallel path and will be ineffective, but luckily because the boat is out of the water there is no risk of electrolytic corrosion either!

Switching off our leakage resistor, we next disconnected the shore line from the GI and reconnected the ring main earth to the hull. We observed an ac voltage between hull/ring-main earth and the shore earth, reaching approx 2v peak-to-peak; this was the shore-generated voltage (700mV rms) as mentioned above. To our surprise the scope showed no difference in this whether we generated a leakage current of 2, 4, or 20 mA; I repeat, even 20mA leakage did not noticeably raise the potential of the hull relative to shore. Re-connecting the GI also had no effect, and there was no noticeable clipping; the GI was demonstrably not conducting, even with 20mA of leakage, because of the parallel path provided by the water.

The ac voltage that was present on the shore earth line was seen to vary considerably with time, jumping from one value to another within the range 50 to 700 mV rms several times per minute. This made it difficult to see the precise effect of various leakage currents, but by watching the digital voltmeter carefully while repeatedly switching the leakage on and off we were able to determine reliably that 2.4 mA leakage actually REDUCED the voltage across the GI by 6.5 mV, 4.8 mA reduced it by 13 mV, and 7.2 mA reduced it by 20 mV. This result was consistent over several dozen measurements, and was not dependent on the absolute value of the voltage at the time.

Several things can be inferred from these measurements:

-     The externally applied voltage is almost certainly coming from other equipment sharing the shore line, which switches on and off to cause step changes in the voltage, as well as from induction within the shore line itself. If this voltage were large enough (whether at 50Hz or at RF) it would drive the GI into conduction.

-     The externally applied 50Hz voltage is in phase with the applied leakage (it would be illogical for it to be otherwise at such a low frequency), which is why increasing the leakage reduces the voltage across the GI.

-     The dynamic impedance of the path from boat to shore via the water is less than 3 ohms at 50Hz (ohms law, 6.5mV per 2.4mA). It would need hundreds of simple devices to cause enough 50Hz leakage to drive the GI into conduction, and the RCD would trip long before you reached this condition.

-     We need to measure the dynamic impedance at higher frequencies, and ideally plot a graph of impedance versus frequency, so we can determine whether or not switched-mode devices can drive the GI into conduction. However I personally would be very surprised if the watery path were to appear so very inductive as to present a high impedance path at any normal frequency.

My take from all this is that the devices on your own boat are unlikely to make the GI conduct, and that you are at greater risk of corrosion from voltages induced in your shore-line earth by devices on shore (and on other boats if their hull is not bonded to earth!). Other reasoned viewpoints will be eagerly considered.

 

 

[Keeping Up]

13 hours ago, Alan de Enfield said:

The EN spec actually says the Earth for the AC and the 12v Negative SHALL (must ?) use the same stud.

 

I know that goes against much of the advice nornally given on the forum, but, that is what the spec says :

 

 

Many years ago, for about 5 years I was employed by BT as an electrical safety assessor for connected private equipment. One of the requirements which I was absolutely required to implement, was that NEVER should 2 different types of safety earth be allowed to share the same stud. At the time most American kit that was imported had a single stud for both the incoming and local earths, and our standards required that the UK vendor must always modify the equipment by fitting a seond stud, before we would allow its connection to our services.

[nicknorman]

1 hour ago, canalboat said:

OK so just as I thought I understood this earthing thing, you electrickery folk have me confused again.

What you seem to be saying is that if I run my extension lead to the boat to use for working electrical tools etc, I must earth it to the hull. AND as my original question was "do I need a G.I." (I use the initials because I am not sure how to spell the words)  to which the answer was "no",   as it MUST be earthed to the hull (and not back to the building RCD where I plug in the lead) the answer is "yes". 

Is that correct?

And by the way, to the person who said I should have a consumer unit on the boat, my extension lead ends with a 13amp socket equipped with an RCD, so no I don't.

Bear in mind that anything upstream of an RCD gets no protection from it. So if you extension lead chafes and the live contacts the hull, the RCD at the end won’t trip even if there is plenty of conductivity to earth.

[WotEver]

1 hour ago, Keeping Up said:

One of the requirements which I was absolutely required to implement, was that NEVER should 2 different types of safety earth be allowed to share the same stud

As you’ll read from an earlier post I believe that Alan has misunderstood that Reg. 

[blackrose]

4 hours ago, Keith M said:

There is another issue which might need to be considered when installing a Galvanic Isolator.

There is alot of documentation that galvanic isolators are being activated 

By the voltages found on the earth cable within the distribution network on the shore side of the galvanic isolator

 

It depends which model of GI you have. RFI can put the diodes into conduction but it takes a lot to do that on the higher spec models. Mine starts to conduct at 2.4v and I tested it with a frend using his ossiloscope, 2 laptops, 2 battery chargers, a couple of mobile phone chargers and a few other AC appliances. We couldn't force it into conduction. 

[blackrose]

3 hours ago, canalboat said:

OK so just as I thought I understood this earthing thing, you electrickery folk have me confused again.

What you seem to be saying is that if I run my extension lead to the boat to use for working electrical tools etc, I must earth it to the hull. AND as my original question was "do I need a G.I." (I use the initials because I am not sure how to spell the words)  to which the answer was "no",   as it MUST be earthed to the hull (and not back to the building RCD where I plug in the lead) the answer is "yes". 

Is that correct?

And by the way, to the person who said I should have a consumer unit on the boat, my extension lead ends with a 13amp socket equipped with an RCD, so no I don't.

 

I think the advice is that you can certainly run an extension lead to your boat to use power tools and for that you don't really need to earth it to the boat's hull as long as the earthing arrangements on the shore are ok. But as soon as you make that extension lead a more permanent fixture, bring it inside and you're using for living then it should run through a consumer unit and be properly earthed to the hull.

 

As we all said earlier, if your boat isn't hull-earth bonded then there's no point fitting a galvanic isolator because the boat is already isolated - it just isn't very safe!  

[David Mack]

3 hours ago, canalboat said:

 

And by the way, to the person who said I should have a consumer unit on the boat, my extension lead ends with a 13amp socket equipped with an RCD, so no I don't.

 

2 hours ago, nicknorman said:

Bear in mind that anything upstream of an RCD gets no protection from it. So if you extension lead chafes and the live contacts the hull, the RCD at the end won’t trip even if there is plenty of conductivity to earth.

 

So the RCD should be at the shore socket end of the extension cable, not the boat end. But if you are taking power from a marina bollard, that should already include an RCD for just that reason.

[Alastair]

So the reason for not having multiple earth points is that this can potentially (sic) induce current flows in the hull.

 

This sounds like an argument of safety (we were getting 50V off computer equipment) vs boat preservation.

[nicknorman]

3 minutes ago, Alastair said:

So the reason for not having multiple earth points is that this can potentially (sic) induce current flows in the hull.

 

This sounds like an argument of safety (we were getting 50V off computer equipment) vs boat preservation.

No, there is no argument for safety. You may get 50v into near-infinite impedance, but so what?

[cuthound]

On 08/04/2020 at 22:59, nicknorman said:

Our boat doesn’t have a consumer unit. A consumer unit is just a box to hold several breakers. I have never seen the point in having multiple breakers on a boat (unless you have a separate circuit for eg an immersion heater or battery charger, which we don’t). We just have the one RCBO which is mounted on a little plastic plate, job done. Our hull earth wire is connected to the inboard stud of the GI.

 

Likewise on our boat.

[cuthound]

On 09/04/2020 at 09:22, WotEver said:

One place I worked at had a measured 170V on an earth wire, but the current was so small that you couldn’t feel it if you grabbed the cable. Inductance over long distances can generate surprisingly high voltages, but they have to flow through a steel hull to cause any problems. 

 

 Can across this in a teleport in London.  The steel wiring wound around the conductors to provide physical protection was causing an induced voltage.

 

I solved it by cutting back the steel wire armouring at the glands to insulate it from earth and running new PVC earth cables in conduit to provide electrical protection. Then boxed in the cable tray carrying the armoured cables.

Edited April 10, 2020 by cuthound
Clarification

[cuthound]

23 hours ago, Keeping Up said:

When another Forum member and I did some testing a few years ago on my boat (steel, blacked, in dirty canal water) we found that the impedance through the water to shore was about 3 ohms. At that time there was an externally-generated voltage of about 0.7v rms present on the shore line at an impedance of less than 10 ohms.

 

If you would like to read the full report, read on, otherwise skip from here:

---

Chris and I made a number of measurements on Keeping Up yesterday: most of them were very useful, and one of them possibly turns conventional thinking on its head by suggesting that other people’s mains-powered equipment presents you with a greater corrosion hazard than does your own equipment.

Keeping up has a fairly conventional setup; the shore-line earth connects via a GI to the ring-main earth which is also connected to the hull. The hull is floating in the muddy waters of the GrandUnionCanal but the piling next to the boat is made of plastic. The readings we took are typical, but naturally wide variations are possible on other boats in other waters.

First we separated the three earths (ring main, hull, and shore); there was no equipment connected to the ring main and we even powered the scope from the inverter to ensure that it did not affect our readings. The only unexpected reading was a potential of 700mV rms (50Hz ac) between the shore line earth and the hull, with a source impedance of less than 10 ohms. Since there was nothing on the boat that could be generating this voltage, we concluded that its source was the shore line. This triggered the thought that a voltage coming from the shore line earth could drive a GI into conduction and make it ineffective.

Next we connected a 100K resistor from ring main live to ring main earth. This of course raised the potential on the ring main earth to 240v ac (don’t touch anything !!!) until we reconnected the shore line earth to the GI. We now had 2.4mA ac flowing through the GI and into the shore-line earth. As expected the GI was driven into conduction and we observed a square wave of about 2.5v peak to peak across it, from ring-main earth to shore-line earth. Nothing unexpected here, it demonstrates that a GI cannot work unless the water provides a parallel path around it; in the absence of any parallel path the GI is forced into conduction by small leakage currents. If your boat is out of the water the GI will not have this parallel path and will be ineffective, but luckily because the boat is out of the water there is no risk of electrolytic corrosion either!

Switching off our leakage resistor, we next disconnected the shore line from the GI and reconnected the ring main earth to the hull. We observed an ac voltage between hull/ring-main earth and the shore earth, reaching approx 2v peak-to-peak; this was the shore-generated voltage (700mV rms) as mentioned above. To our surprise the scope showed no difference in this whether we generated a leakage current of 2, 4, or 20 mA; I repeat, even 20mA leakage did not noticeably raise the potential of the hull relative to shore. Re-connecting the GI also had no effect, and there was no noticeable clipping; the GI was demonstrably not conducting, even with 20mA of leakage, because of the parallel path provided by the water.

The ac voltage that was present on the shore earth line was seen to vary considerably with time, jumping from one value to another within the range 50 to 700 mV rms several times per minute. This made it difficult to see the precise effect of various leakage currents, but by watching the digital voltmeter carefully while repeatedly switching the leakage on and off we were able to determine reliably that 2.4 mA leakage actually REDUCED the voltage across the GI by 6.5 mV, 4.8 mA reduced it by 13 mV, and 7.2 mA reduced it by 20 mV. This result was consistent over several dozen measurements, and was not dependent on the absolute value of the voltage at the time.

Several things can be inferred from these measurements:

-     The externally applied voltage is almost certainly coming from other equipment sharing the shore line, which switches on and off to cause step changes in the voltage, as well as from induction within the shore line itself. If this voltage were large enough (whether at 50Hz or at RF) it would drive the GI into conduction.

-     The externally applied 50Hz voltage is in phase with the applied leakage (it would be illogical for it to be otherwise at such a low frequency), which is why increasing the leakage reduces the voltage across the GI.

-     The dynamic impedance of the path from boat to shore via the water is less than 3 ohms at 50Hz (ohms law, 6.5mV per 2.4mA). It would need hundreds of simple devices to cause enough 50Hz leakage to drive the GI into conduction, and the RCD would trip long before you reached this condition.

-     We need to measure the dynamic impedance at higher frequencies, and ideally plot a graph of impedance versus frequency, so we can determine whether or not switched-mode devices can drive the GI into conduction. However I personally would be very surprised if the watery path were to appear so very inductive as to present a high impedance path at any normal frequency.

My take from all this is that the devices on your own boat are unlikely to make the GI conduct, and that you are at greater risk of corrosion from voltages induced in your shore-line earth by devices on shore (and on other boats if their hull is not bonded to earth!). Other reasoned viewpoints will be eagerly considered.

 

 

 

 

Was there a telephone exchange, hospital, data centre or factory nearby?

 

These use equipment with large switch mode power supplies to provide resilient power or speed control of machinery, which creates high  harmonic currents in the neutral. These are limited to 5% of load current by statute, but not many electrical  Distribution Network Operators ask for evidence of this.

 

 

[Keeping Up]

17 minutes ago, cuthound said:

 

Was there a telephone exchange, hospital, data centre or factory nearby?

 

These use equipment with large switch mode power supplies to provide resilient power or speed control of machinery, which creates high  harmonic currents in the neutral. These are limited to 5% of load current by statute, but not many electrical  Distribution Network Operators ask for evidence of this.

 

 

No there's nothing like that anywhere nearby. It's a fairly isolated location, by Talbot's Lock near Stoke Hammond, and there's nothing within 1/4 mile except 2 cottages (the lock house and the pump house; the pump wasn't pumping at the time) plus the other 3 boats; after that there is just the village. Even the old Sewage Works was decommissioned many, many years ago.

[canalboat]

21 hours ago, blackrose said:

 

I think the advice is that you can certainly run an extension lead to your boat to use power tools and for that you don't really need to earth it to the boat's hull as long as the earthing arrangements on the shore are ok. But as soon as you make that extension lead a more permanent fixture, bring it inside and you're using for living then it should run through a consumer unit and be properly earthed to the hull.

 

As we all said earlier, if your boat isn't hull-earth bonded then there's no point fitting a galvanic isolator because the boat is already isolated - it just isn't very safe!  

If I wasn't confused enough already - your answer#10

["No, but you have much bigger problems (potentially life-threatening)"]

seems to say that my extension lead must be earthed or I risk killing someone but contradicts your latest answer that suggests using such a lead on a boat is OK.

But then I seem to have misunderstood all the rest so I may well be wrong again. 

Edited April 10, 2020 by canalboat

[WotEver]

16 minutes ago, canalboat said:

If I wasn't confused enough already - your answer#10

["No, but you have much bigger problems (potentially life-threatening)"]

seems to say that my extension lead must be earthed or I risk killing someone but contradicts your latest answer that suggests using such a lead on a boat is OK.

But then I seem to have misunderstood all the rest so I may well be wrong again. 

It’s no different to if I was to run an extension lead down my garden to feed a power tool. It’s temporary, and as soon as I’ve finished then I’ll wind it up and put it away. However, if I wanted permanent power to the bottom of the garden then it should be run in steel wire armoured cable, correctly installed, and terminating in a consumer unit of some kind. Just the same with a boat - a temporary extension lead is acceptable whereas a permanent lead entering the boat which could conceivably chafe against the hull over time is unacceptable, and potentially life-threatening.

[Alan de Enfield]

5 minutes ago, WotEver said:

It’s no different to if I was to run an extension lead down my garden to feed a power tool. It’s temporary, and as soon as I’ve finished then I’ll wind it up and put it away. However, if I wanted permanent power to the bottom of the garden then it should be run in steel wire armoured cable, correctly installed, and terminating in a consumer unit of some kind. Just the same with a boat - a temporary extension lead is acceptable whereas a permanent lead entering the boat which could conceivably chafe against the hull over time is unacceptable, and potentially life-threatening.

Quite rightly, the BSS are not impressed with 'permanent' extension leads running thru a hatch or window being used in lieu of properly wired 'shore-lines'.