Generator earthing for a caravan

[amnlicnoln]

Hi,

I am planning on using a portable generator (Honda eu20i if it matters) to power an old caravan. The caravan doesn't have a permanent electrical installation so I was planning on using one of the tent hookup cables with 3 plugs and and MCB and RCD built in (http://www.gooutdoors.co.uk/mobile-mains-kit-p155317)

The question is do I need to earth the generator with this setup or will it work as is? Also if I do need to earth I assume I would have to put in a Neutral-Earth link?

Thanks

Adam

 

[ericmark]

As long as a generator or isolation transformer supplies one item only there is no need for an earth in the same way as a shaver socket in a bathroom. However once you start to split the supply then earth is required so I would say it will need a common earth point likely bonded to neutral inside the generator but no real need for an earth rod.

SECTION 717 MOBILE OR TRANSPORTABLE UNITS does detail what is required Fig 717.1 See Here page 49

Some common sense Class II items don't need an earth so 4 items 3 being class II does not really need an earth but 3 class 1 and 1 class II then really it does.

Personally I would fit one anyway.

As a P.S. I have had problems with Honda generators where the internal link is not earth to neutral but earth to a 55v point so would not bond neutral to earth in any box without first ensuring not already done in the generator.

 

[viewer]

This is a question.
On the basis of my limited electrical knowledge and (?) logic, I would have thought that (given all other protection is good and appropriate, the earth wire is sound caravan to generator and there is a RCD on the generator) :

With neither generator or caravan earthed it would be safe.
If the caravan is earthed but the generator insulated from earth it would be safe.
If both ends are properly earthed it would be safe.
If the generator was earthed, but the caravan not earthed then safety could rely on the RCD working so there is a slight potential hazard.
Have I missed something?

 

[ericmark]

Let us consider why we earth.
If one item has a fault and live 1 is connected to exposed metal work then you will not get a shock this is what happens with a shaver socket. However if a second items has a fault and live 2 is connected to exposed metal work then touching both bits of metal work you will get a shock. But if the two bits of metal work are connected together then excess current will flow and the fuse will blow making it safe this is why we bond.

The three way socket bar you linked to will effectively bond exposed metal work together and so with two independent faults the fuse should blow.

With just three items in use the chances of two faults one connecting live 1 to exposed metalwork and the other live 2 to exposed metalwork is very slim.

But for the RCD to work there must be a path for any leakage to return without passing through the RCD and without all exposed metalwork being bonded there is a possibility the RCD will fail to operate due to a fault.

There is also a chance using a small generator that there will not be enough current to blow a fuse.

However if you do a risk assessment the risk is still low but never the less there is a risk.

So your question is without visiting the site can we do a risk assessment and decide if you need to bond and sink in earth rods and also you have no idea as to the existing internal bonding within the generator.

The answer has to be this can't be done. All we can do is make you aware of the risks that may be involved and the end you have to do your own risk assessment and decide for your self. As I have said it is so easy to bond I would bond I would check if a live is already linked to earth forming a neutral likely it is. Personally I would never connect a neutral to earth outside of the generator as so often it's not connected that way in the generator and it would be too easy to get a short circuit as a result.

If I found neutral was not bonded in a new generator I would simply return it as not fit for the job I would not attempt to correct in any way myself.

So bond and earth rod nothing else and even that only after assessing the risks.

 

[amnlicnoln]

Thank you for the useful replies. I'm going to have a look at the appliances I plan to run from the generator and weigh up the risks. I may be back for further advice!

 

[JohnW2]

Let us consider why we earth. ....

Eric, although you started your post with those words, everything you went on to say was really about the merit of bonding together exposed-conductive-parts when a generator (or isolation transformer) supplies more than one item of Class I equipment. I agree that such bonding will minimise the risk of shock in the (extremely unlikely) event of two simultaneous faults - one resulting in connection of one side of the supply to exposed parts of one load, and the other connecting the other side of the supply to exposed parts of another item of equipment (and then both items being touched simultaneously).

However, in terms of true earthing (in this context, not only connecting the exposed-c-ps together, but also connecting them to an earth rod), I am much more confused/uncertain. At first sight, I can think of only disadvantages (increased risk of shock) resulting from such earthing. With a normal grid supply, the fact that it is already earth-referenced (at transformer) is unavoidable fact, so the option of having a 'floating' supply does not exist. However, if (as with a local generator) one often does have the theoretical option of a floating supply, I find it hard to see why that is not preferable. What am I missing?

Kind Regards, John

 

[bernardgreen]

In theory a generator that does not have it's output referenced to ground would be a floating supply. ( referenced to ground meaning one side or the mid point connected to an earth rod ).

However in reality there will be some weak reference to ground via the capacitive coupling of windings to the frame of the generator which may be in contact with the ground. So a mild low current shock could be felt when touching the supply from the generator. But not enough earth leakage current to operate an RCD. Does one earthing one side of the generator make it safer when there is a fault by increasing the earth leakage current to a level high enough to trip the RCD.

 

[JohnW2]

In theory a generator that does not have it's output referenced to ground would be a floating supply. ( referenced to ground meaning one side or the mid point connected to an earth rod ).

Quite so - that's obviously what I was talking about.

However in reality there will be some weak reference to ground via the capacitive coupling of windings to the frame of the generator which may be in contact with the ground. So a mild low current shock could be felt when touching the supply from the generator. But not enough earth leakage current to operate an RCD.

Indeed - nor (other than in very exceptional circumstances) enough current to do any harm.

Does one earthing one side of the generator make it safer when there is a fault by increasing the earth leakage current to a level high enough to trip the RCD.

Very debatable, I would have thought! Deliberately increasing the potential shock current (in certain circumstances), from a 'harmless' level to a potentially lethal level in order that an RCD can limit the duration of that (deliberately increased) shock current seems to me to be to be an exceptionally iffy concept, isn't it?

Can you think of any other advantages of earthing a supply which is otherwise floating (other than through capacitive coupling to earth)?

Kind Regards, John

 

[bernardgreen]

I was thinking about the generator developing a fault that grounded the output and thus changing the supply from un-referenced to ground referenced. By intentionally earthing the generator output that fault would not be as serious.

As you say open to discussion and very much affected by the circumstances of each situation.

 

[JohnW2]

I was thinking about the generator developing a fault that grounded the output and thus changing the supply from un-referenced to ground referenced.

Fair enough, but that 'fault' would merely be creating the very same connection to earth which I suspect many people would advocate deliberately installing in the first place! ..."

...By intentionally earthing the generator output that fault would not be as serious.

As above, I'm not sure in what sense your proposed happening would ever represent a 'serious fault', since it would merely be reproducing what many people would probably do deliberately. ... particularly since a fault at the generator which resulted in one side of its output being earthed would presumably normally be upstream of, and therefore not detected by, any RCD.

As you say open to discussion and very much affected by the circumstances of each situation.

I don't doubt that some people would feel the need for discussion but, as I've said, I am personally so far struggling to think of anything other than 'downsides' (potentially increased hazards) which would result from 'unnecessarily' earth-referencing a supply which is essentially floating, no matter what the circumstances of the particular situation.

Kind Regards, John

 

[maltaron]

Perhaps a relevent point is that many small generators have, in their instructions, the statement that the neutral must not be connected to earth, only the generator frame should be earthed.

 

[JohnW2]

Perhaps a relevent point is that many small generators have, in their instructions, the statement that the neutral must not be connected to earth, only the generator frame should be earthed.

I think what you say is true - and is essentially underlining my point (that I don't see the point in 'unnecessarily' earthing eitehr side (or midpoint) of the genny's output

As for the totality of the instructions you mention, if the generator's output is floating relative to earth (neither side of output, nor its midpoint, earthed), then I would find it very hard to understand what would be gained by earthing its frame.

Kind Regards, John

 

[maltaron]

To throw another situation into this topic, I have in my bungalow, an oil fired cooker/boiler. In the case of a power outage I run this off an invertor. I have just purchased a small generator to replace the invertor. Now this is connected by unplugging the 13 amp plug from the mains and connecting to the invertor/generator thus isolated the earths. But, the cooker boiler is also connected to incoming PME supply earth via the water pipe and oil feed bonding. In this case should I disconnect the invertor/ generator earth connection in the plugtop?

 

[JohnW2]

To throw another situation into this topic, I have in my bungalow, an oil fired cooker/boiler. In the case of a power outage I run this off an invertor. I have just purchased a small generator to replace the invertor. Now this is connected by unplugging the 13 amp plug from the mains and connecting to the invertor/generator thus isolated the earths. But, the cooker boiler is also connected to incoming PME supply earth via the water pipe and oil feed bonding. In this case should I disconnect the invertor/ generator earth connection in the plugtop?

That's obviously a somewhat different question, and one on which I suspect opinions may well vary. In the situation you now describe, the question is not really whether or not one should earth the generator supply but, rather, whether it is 'safe' to have the frame of a generator connected to a PME earth during a power cut.

I can certainly see an argument for doing as you suggest. Very rarely, a 'power cut' may be due to (or associated with) a supply-side fault that results in the PME 'earth' rising to a potential much higher than true earth. If that were the situation, and if that PME earth were connected to the genny frame, someone touching that frame at the same time as being in contact with 'truth earth' could theoretically be at some risk. I will be interested to hear what others have to say.

Kind Regards, John

 

[ericmark]

I have like I am sure many others made my fair share of mistakes. On a Farm on the Falklands I wired up a generator using a centre tap as earth. It seemed a good idea at the time only a 120 volt shock with a fault but problem is there are no fuses on the neutral so I had no protection on line 2 as using a standard consumer unit not a split phase. Realised my error and corrected it.

However after that I found I was not the only one to have made that error. Found it on 110 volt control circuits and even Honda Generators.

Today I always check but to tell some one how to do it rather than doing it yourself it is all too easy to miss something. Glad it has started a debate as only by having statements reviewed by ones peers can any errors be highlighted.

Be it a caravan or electric car on charge with TT earth close to premises with TN-C-S or the situation detailed here the question of what to bond or earth is complex.

Maybe the question should be "Why do you ask?" most people would not think twice generator has a socket you can buy a lead with changes this from a single 16A to three 13A with a RCD included why should anyone consider this as may not be safe? Often it is because of something they have failed to relate. Maybe the fan belt was catching the guard and they got a build up of static and a shock? However I would hope between the answers given the question is now answered.