2 hours ago, Ingo said:

Back to my initial options. Is the isolation transformer enough to satisfy the requirement that the generator is not directly connected to a fixed installation?

I'm not sure you'd get away with this. If we took that form of connection to its logical conclusion.

There wouldn't be a single ESKOM generator connected to the grid in SA, if just being the other-side of a transformer was all it took to say that.

And I think we can all agree there are many ESKOM generators attached to the grid.

No, I think you have to modify the generator to no longer be a V0V, or it's disqualified, full stop.

22 minutes ago, phil.g00 said:

No, I think you have to modify the generator to no longer be a V0V, or it's disqualified, full stop.

I disagree. One could argue that the transformer modifies the generator and makes it no-longer-a-0-V-0 thing. Which is what it does.

2 minutes ago, plonkster said:

One could argue that the transformer modifies the generator and makes it no-longer-a-0-V-0 thing. Which is what it does.

Yes, that is the implicit argument I am countering.

It doesn't modify the V0V nature of generator to the transformer circuit, it just embeds it and hides the effect of the different earth reference in normal operation.

That's unfortunately, only half the story, as it must also work during abnormal operation.

If there is a earth fault on the load side the transformer, earth fault current  must flow between the fault site and the source earth fault connection. Practically, that means from a live wire into the ground, through the ground, and up the earthed transformer neutral to complete the circuit.  The earth leakage detects what is outgoing current on the live being unmatched by incoming current on the neutral and switches the circuit off.  So a low impedance path is required by the circuit to enable this form of protection.

This circulating current will have to ampere-turn matched by an equivalent circulating current on the generator side of the transformer. Essentially magnetism, on one side of transformer must be matched by magnetism (translating into circulating current) on the other side of the transformer.

Earth fault current has to be able to circulate on the gen-side of the trafo, to allow it circulate on the load-side, or otherwise the earth leakage protection wont work and neutral voltages will go dangerously high.

So immediately, the quality/quantity of earthing gen-side  becomes an issue as it compromises the functionality of earth-leakage protection on the load side of the transformer.

As none of this is mentioned by the rules, and it is critical to safety, a blanket "No", I think that it means "No" , not "Maybe".

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

I'm not sure I understand your earth fault argument. The generator side is isolated from the load side. A leakage on the load side just looks like a part of the load to the other side. You can even earth the midpoint on the generator and it won't matter.

This is what boats/yachts do if they end up in a harbour where things don't match. And to avoid the galvanic corrosion thing to the hull and all that.

Or another way to think of it.... the transformer itself becomes a generator. The source of magnetic energy should not matter... should it? :-)

I can´t see where is the problem in this scheme:

 

 

12 minutes ago, plonkster said:

Or another way to think of it.... the transformer itself becomes a generator. The source of magnetic energy should not matter... should it? :-)

Putting the different earth reference arguments to one side.

Lets start backwards:

OK there are a few things at play here.

The obvious, a circuit is an unbroken loop.

Ohms' law V=IZ, or stated differently I =V/Z, Z being impedance.

We require current to flow for earth fault protection. However many mA sensitivity blah-blah, this doesn't me we only want that amount of mili-amps to flow, it means we want to be damn sure that at least that many mili-amps will flow in the case of an earth fault. So we want a low earth loop impedance, otherwise our earth fault protection is compromised.

I think is understood.

What you probably need to get your head around is where current goes in an earth fault. It goes back to the source reference earthed point.

Looking at the circuit purely from the load-side transformer winding the circuit is out the live wire of the trafo, through the one pole of the earth leakage into the fault, into the mass of earth, out of the mass of earth and back up neutral of the transformer completing the circuit.

So far so good?

Right,  now in order to maintain this low impedance loop, the transformer winding must represent a low impedance. It can only do that if the magnetism it produces is "burnt off" by the other-side winding.  (An equal ,opposite ampere-turn balance, this is the principle of transformer operation).

So a circulating earth fault current load-side has to be matched by a magnetic circuit on the other side the transformer. This in turn will be dissipated by a circulating current on the gen-side.

Any impedances in this circulating current path on the gen-side of the trafo will be imposed impedance on the load -side and could defeat the earth leakage protection.

In other words if you measured the impedance (not resistance!) of an open circuit two winding transformer as opposed to a short-circuit two winding transformer you would get two different impedances.

What I am saying because this impedance now becomes critical to safety, and no consideration is given to it, it has been considered and completely rejected.

 

10 minutes ago, Javi Martínez said:

I can´t see where is the problem in this scheme:

 

 

 This possibly illustrates my point as well, if the earth -fault protection on the generator side relies on current being driven through the transformer. And that transformer is open-circuit on its secondary as you've drawn it, that transformer will represent a very high impedance that is reflected on the gen-side.

So the current would be greatly limited by the impedance and voltages would rise.

1 hour ago, phil.g00 said:

Looking at the circuit purely from the load-side transformer winding the circuit is out the live wire of the trafo, through the one pole of the earth leakage into the fault, into the mass of earth, out of the mass of earth and back up neutral of the transformer completing the circuit. 

Right.

1 hour ago, phil.g00 said:

maintain this low impedance loop

This bit I don't get. If I recall we're concerned about the impedance of the earth loop, ie how well the neutral is bonded to terra firma... and we want this to be low so that large fault currents result when there is some kind of problem. I'm not quite sure how the impedance of the transformer comes into that. Wouldn't the impedance drop as soon as you load it down anyway, even if you load it down with an earth fault? :-)

I can sort of see how earth fault detection might be impacted on the primary side, and to that there is an easy answer: Don't worry about it. You don't need an RCD on that side, because it isn't feeding any sockets or appliances. Just earth the generator and put in overcurrent protection.

Edit: Because it is earthed in the middle, I suspect you'd have to apply overcurrent protection in the American style, that is in both legs.

Edited January 6, 20197 yr by plonkster

Remember the earth on the Generator side is connected to the transformer earth, like the illustration in the beginning. The earth on the secondary side is the utility earth. Also, the generator has an internal overcurrent protection device (that's what the specs say). If it's single or double pole I don't know but it's better to be safe than sorry and put one in anyway.

1 hour ago, phil.g00 said:

 This possibly illustrates my point as well, if the earth -fault protection on the generator side relies on current being driven through the transformer. And that transformer is open-circuit on its secondary as you've drawn it, that transformer will represent a very high impedance that is reflected on the gen-side.

So the current would be greatly limited by the impedance and voltages would rise.

I would say the only issue is an overvoltage in transformer side if current stops suddenly. But luckily, windings are not ideal inductors and voltage doesn´t go into the sunset...

The same issue we have when we switch off a motor.

30 minutes ago, plonkster said:

Right.

This bit I don't get. If I recall we're concerned about the impedance of the earth loop, ie how well the neutral is bonded to terra firma... and we want this to be low so that large fault currents result when there is some kind of problem. I'm not quite sure how the impedance of the transformer comes into that. Wouldn't the impedance drop as soon as you load it down anyway, even if you load it down with an earth fault? 🙂

I can sort of see how earth fault detection might be impacted on the primary side, and to that there is an easy answer: Don't worry about it. You don't need an RCD on that side, because it isn't feeding any sockets or appliances. Just earth the generator and put in overcurrent protection.

Edit: Because it is earthed in the middle, I suspect you'd have to apply overcurrent protection in the American style, that is in both legs.

Yes, it is used too in IT schemes to prevent from "second fault" since the "first fault" has no consequence. First fault can be detected using resistors from ground to phases; I think it would be easy to implement for MR DIY.

13 hours ago, plonkster said:

You don't need an RCD on that side, because it isn't feeding any sockets or appliances.

That's an assumption. It's 220V,  it will switch things on, it will work, it will be used. Things could conceivably even be fed from the grid,  on the gen-side, when the generator is switched off.

13 hours ago, Ingo said:

The earth on the secondary side is the utility earth.

But, will all appliances have effective earth leakage protection?

It has the possibility of having 110V between the casing of this appliance and the next one.

My point is the regulations don't speak to mitigating or entertaining this situation.

My interpretation of this lack of consideration is the regulations say "no",  too messy, not in the same place, all neutrals are at earth potential here, if something faults then this standard thing saves your life, not Mr DIY's concoction.

Well... then I suppose it depends on how scared you are. Isolation transformers are frequently used for exactly this sort of thing. If you have three phase power on a farm, for example, you have three lines on the pole and an earth right at the top, and then a delta-star transformer with its star earthed on the other end. Basically the same thing.

No, a delta winding is not the same thing. A delta winding can (as a closed circuit) inherently circulate current around the delta winding and achieve ampere-turn balance. Therefore the earth fault current available (loop impedance) on the earthed star-side is not compromised.

Well... I'd love to understand what you are talking about.

Let's ignore the fact that there are three phases on that delta side, and take the typical residential case where I get only one of them. So for all practical purposes it's a single 230VAC winding with the one side earthed (as per low voltage regulations)... with the other side having its own earth in a completely different spot by different regulations (and sometimes none at all).

I'm not an electrician. I work merely from the fact that isolation transformers (usually including 115V->230V and vice versa options) are frequently used to adapt the power system on your boat to the shore power of the particular place you're moored at :-)

Edited January 7, 20197 yr by plonkster

8 minutes ago, plonkster said:

Let's ignore the fact that there are three phases on that delta side, and take the typical residential case where I get only one of them.

I know symmetrical components are head wrecking.

I'll try.

The magnitude of available fault current includes an earth fault component which is governed by quality of the earth fault loop impedance. The entire fault current also includes current contributions that have nothing to do with the connection to earth.

1.Picture this, a star/star isolation transformer with both neutrals earthed. (Could be a two winding - both neutrals earthed if you want)

This allows earth fault current to circulate from the fault site up the earth neutral and back again, this in turn is matched by a current on the other side of the transformer flows between the earth and neutral and circulates through in turn through its own source.

2.Then the far sides earthed neutral is removed so that current can no longer circulate, this high impedance is then reflected back on your side of the transformer and consequently even though your system is earthed the earth loop impedance is high. So touching a hot wire, the neutral is still the preferred current path and a person is effectively in series with an open circuit.

The possibility of creating a short circuit is obviously still available, but that is a low impedance between the phase and neutral and exists whether the earth is there or not.

 

 

 

Still don't see why the two earths influence each other. I modified the above pic a bit. I have a load of 1A and a leakage current of another 1A. Since the transformer is 1:1, I expect 2A to flow on the other side (more or less). There should be no current in the earth conductor on that side. There should be no reason for anything on the left to affect anything on the right. There is no direct path for current.

If I measure the loop impedance between live and earth on the left from this source is it low enough?

I am saying that that impedance is a function of the loop impedance on the right, by virtue of the ampere-turns ratio, not because of an electrical connection.

 

In other i am not saying that what is happens on one side doesn't happen on the the other,  rather I am questioning why is the fault current is 1A instead of 100A?

(I know you picked an arbitrary number, but earth fault current actually can be limited like this).

Okay, I think I sort of see your argument. Though I sort-of think that's why you use an RCD... because you can't rely purely on the magnitude of the fault current. In other words, the same sort of things that would apply with a TT earthing system.

Yeah, its very un-intuitive, it would be a different fault current if you also earthed the centre-tap of the transformer on the gen -side.

Then you would be able to circulate the zero sequence component  between neutral and earth on that side, and would result in a lower loop impedance on the other -side.

I didn't want to get into that, its bad enough realizing that the impedance of a circuit is affected by things that only magnetically, and not electrically in the circuit.

That said I got my R Pi today, I have never even seen one up close and personal before.

I've got it working on the TV, and I am about to look see how to install your Venus  stuff on it.

 

On 2019/01/05 at 3:22 PM, plonkster said:

There is a kind of earthing system known as TT (Terra-terra) where you don't have bonding. The RCD is still required and it still works: If you touch either side (neutral and live are at the same 115V potential) and enough current flows to ground, it still causes an imbalance and the RCD still detects it and trips. In such a system. I sorta-kinda feel like it might be okay to leave the bonding off. But then you need to earth the center point on the generator so that your two "hots" are tied to a 0V reference somewhere (in the middle in this case).

Probably best to ask an electrician to be sure, but right now I can't see why it would be a problem to temporarily convert the house to a US standard.

Problem with this type of system is that you need an earthing kit for Solar inverters and on variable speed drives you need to take the varistor screw or connection out. I`m not sure what it will do to other equipment not meant for this type of system as it is normally found in the States.

I said i'd report back with Honda's response, so for those of you that speak Chinglish:

Please refer to the chapter in OM for generator use. According to our information it is not possible, due to the fact Honda have never test the generator in such electrical system. In some country are such kind of electrical systems have strong regulations. Also synchronizing of this system according to the customer information, will be not possible.

There you go, I suppose it's fair, they didn't understand my question, and I don't understand their answer.

19 hours ago, phil.g00 said:

I said i'd report back with Honda's response, so for those of you that speak Chinglish:

Please refer to the chapter in OM for generator use. According to our information it is not possible, due to the fact Honda have never test the generator in such electrical system. In some country are such kind of electrical systems have strong regulations. Also synchronizing of this system according to the customer information, will be not possible.

There you go, I suppose it's fair, they didn't understand my question, and I don't understand their answer.

"See no evil, hear no evil" ...something like that...