Marine AC - Why Tie Neutral and Earth?

[dkenny64]

Originally Posted by Wotname

They work ONLY if one of the source wires (i.e N) is tied to 'ground'. In a fully floating (2 wire) output neither wire is connected to anything anywhere except the load.

yes, until there is a fault.. then where does the current go?
drop one side the into the water.. salt water...wouldn't this current go somewhere
else besides where its supposed to go?
there by tripping a GFCI?


dkenny64

[Jammer]

Originally Posted by Wotname

Several posters keep referring to 'ground' faults but exactly what is 'ground' fault in a two wire floating output?

A ground fault is a low-impedance path for either conductor to ground. In the absence of a concerted attempt to bond conductive materials together, a ground fault is understood as a low-impedance path for either conductor to any conductive material not intended to carry current.

How can a 'ground' fault exist if the potential difference (say 120V AC) only exists between the two conductors. AFAIK, there can't be any potential difference between either conductor (A or N) and any other surface.

In the real world there's going to be measurable voltage because there will be leakage. Off isn't completely off, insulation doesn't have infinite resistance, etc. As noted upthread, in the absence of any faults you'll typically see around 60 volts between either conductor and nearby surfaces if neither conductor is bonded.

The only possible time an operator could experience a risky potential difference could occur (AFAIK) is IF a fault (ie conductive path) exists between one conductor (say A) and a conductive case of an load and at the same time another similar fault occurs in another similar load between the other conductor (i.e. N) and if one touches both the conductive cases at the same time.

Put another way, in a fully floating two wire output, the only electrical risk occurs when one's body becomes part of the circuit between the two wires. The ground or earth or engine or hull or sea water or the man on the moon can never be part of the circuit (unless two separate faults (one on each wire) occur on two different objects , and your body connects the two seperate objects.

The problem with this line of reasoning is that in the absence of bonding you don't know whether you have a ground fault unless you are somehow testing, and testing is difficult. For example, a GFCI won't trip due to a single ground fault in an isolated-ground system. As a result, over time, you end up with a system that is actually bonded at some unknown location, with unknown impedance, and to any one of the load-carrying conductors.


The result of this? You can get a potentially lethal shock from touching, say, the air conditioning vent and the stove at the same time. Neither one is bonded to the other and neither one is bonded to neutral, so all it takes is latent defects, a short to hot in one, a short to neutral in the other, and you have a hazard.

[Wotname]

Originally Posted by mitiempo

I assume you are referring to a GFCI without a ground connection - eg only hot and neutral. In this case a "ground fault" is when there is 5 millivolts or more difference between hot and neutral.

It is a ground fault because the user either becomes the grounded conductor or there is leakage to another item - moisture for example - from either the hot or neutral.

GFCIs work just fine without a ground connection.

OK, this helpful - thanks. I'll work though the equivalent circuit later today / tonight to make sure I am fully understanding it and report back

[Frankly]

All real electrical systems have leakage resistance to the environment and in an AC system capacitance (more system more leakage and more capacitance). Doesn't take much of this to cause problems for us humans.

And yes a GFI can trip even on an ungrounded (practical) electrical circuit. GFI work by comparing the current flowing in both legs. Everything OK if these currents are the same. If they differ by more than 5ma then off goes the power. How can this happen in an ungrounded system, capacitance to the environment always exists and an electrical path to this environment has developed allowing an upset in this balance.

A ground fault typically indicates a hard connection to ground (any conductive surface) but really it is just a matter of what is the number.

You can see a truck coming your way and may have the opportunity to get out of the way. Electricity does not provide that luxury.


Frankly

[Dsanduril]

Originally Posted by Wotname

Several posters keep referring to 'ground' faults but exactly what is 'ground' fault in a two wire floating output?

How can a 'ground' fault exist if the potential difference (say 120V AC) only exists between the two conductors. ...

This is why I prefer RCD, and even that is a misnomer, a “differential current detector” would be a better term. When called a GFCI it is because the presumption is that any differential in the current is flowing through a ground connection. But as you have clearly shown this isn’t true in a floating system and isn’t necessary for a differential current device to operate - all they care about is the balance of electrons in both wires of the circuit.

[Jammer]

Originally Posted by mitiempo

I assume you are referring to a GFCI without a ground connection - eg only hot and neutral. In this case a "ground fault" is when there is 5 millivolts or more difference between hot and neutral.

Have another toke

[Jammer]

Originally Posted by Dsanduril

This is why I prefer RCD, and even that is a misnomer, a “differential current detector” would be a better term. When called a GFCI it is because the presumption is that any differential in the current is flowing through a ground connection. But as you have clearly shown this isn’t true in a floating system and isn’t necessary for a differential current device to operate - all they care about is the balance of electrons in both wires of the circuit.

I think you will find that it depends where the RCD is placed. In the absence of grounding, a single RCD placed at the power source (generator/inverter) achieves nothing except coverage of the corner case where there is a ground fault in the generator/inverter as well as a ground fault elsewhere. The common hazard in an ungrounded system -- that is, a ground fault from hot on one piece of equipment and a ground fault from neutral on another -- will not be detected unless there is an RCD protecting only one of the two pieces of equipment. This is in contrast to a neutral bonded system where the RCD provides meaningful protection as long as the bond is properly placed, that is, on the supply side of the RCD.

[s/v Jedi]

99% of what is written in this thread is wrong. A true floating system is at least as safe, if not safer than a polarized system.

But that doesn’t mean you don’t need ground wiring to outlets, because you do.

First, there can be no ground fault because ground is not a return path. If you throw L1 overboard, nothing happens, no current will flow. Only L2 is a return path.

So the case for ground wiring: appliances think they are connected to a live/hot/phase conductor, a neutral conductor and a ground conductor. Any metal parts of the housing are connected to that ground conductor.

So lets assume it develops a short between the hot and the metal housing. Now that housing is hot and no breaker flips. But that isn’t a problem because when you touch that nothing bad happens. Even if you touch that with onehand and stick the other in the sea, nothing happens because ground is not a return path.

But it can get worse: another appliance develops a short between neutral and the metal housing. Now if you would touch that with one hand while touching the other housing with the other hand.... nothing happens because the breaker flipped as the ground wiring completed that circuit.

Polarized systems don’t even detect a short between neutral and the metal housing. It is a fault that lives on unnoticed.

Here is a problem with a floating system according to many: when you push the test button on a gfci outlet, it fails the test. So let’s see what happens with this button: it is a switch that connects a resistor between the hot and the ground conductors. The current flows back through ground instead of neutral, which is detected and trips the outlet. But with a floating system ground is not a return path so no current flows so it doesn’t trip. You simply don’t need gfci protection.

So finally a real “problem” : you need double pole breakers. That’s it. In return you get a safer system but beware the moment you connect shore power, you must use an isolation transformer. Don’t forget to remove the neutral to ground jumper from it

[Thumbs Up]

Pretty sure that the European standard and the newest AYBC standards allow floating grounds with proper residual current devices (RCD's). A lot of what has been said already is wrong. It is a contentious subject because the holy grail of boat safety is supposed to be about that green wire being a return path to ground that would trip a breaker in case of a ground fault. There are two definitions to the phrase ground fault though. The first is basically a short circuit that occurs when an energized conductor touches anything that is grounded and hopefully causes a breaker to trip. The second definition is a type of fault in which the unintentional pathway of the straying electrical current flows. It is this second definition that we refer to when talking about RCDs and GFCIs (ground fault circuit interrupters). These devices need no reference to earth ground (as some have alleged). They do not compare voltage (as someone said), they compare current between the two conductors. If there is a difference , the device trips. With a floating system (no ground reference), outlets do not need their ground wire connected. The scary thing about grounded systems is that they are tied to the bonding system, which is tied to the water (through the propshaft/thruhull bonding) If that earth wire inadvertently becomes live in a way that does not trip a breaker, the water around the boat as well as the boat itself become energized. The beauty of the floating system is that this can not happen. The problem with return path to ground logic is that a partial or intermittent short might not trip a breaker but can still transfer enough juice to cause an electrocution or electric shock drowning. The problem with RCD devices is nuisance tripping, and people who bypass them (instead of correcting the fault).

[Thumbs Up]

Originally Posted by Jammer

.....The problem with this line of reasoning is that in the absence of bonding you don't know whether you have a ground fault unless you are somehow testing, and testing is difficult. For example, a GFCI won't trip due to a single ground fault in an isolated-ground system. As a result, over time, you end up with a system that is actually bonded at some unknown location, with unknown impedance, and to any one of the load-carrying conductors.

The result of this? You can get a potentially lethal shock from touching, say, the air conditioning vent and the stove at the same time. Neither one is bonded to the other and neither one is bonded to neutral, so all it takes is latent defects, a short to hot in one, a short to neutral in the other, and you have a hazard.

This is a conundrum, although this still could happen with a ground wire and no RCD (through a partial short). It could still happen with a ground wire and both appliances connected to the same RCD. It could not happen if they were connected through different RCDs nor if the appliances were double insulated (no metal case). As to the original question about the logic of tying neutral/ ground with an onboard AC supply you guys explained it now I understand. It is still a floating system in a way as it is not bonded to the hull or the shore nor the water, but you need that ground wire to protect you still when using non double insulated appliances that have the case grounded. On a boat, the chances of having two or more single insulated devices in close proximity that both have faults are really slim but I can now see why there would be a rule requiring the ground wire to be connected to the outlets.

[coopec43]

I'm trying to follow.

I realize I have a lot of reading to do. I'll start with the real basic stuff (most of it I think I know)

How To Wire (or re-wire) a boat

https://newwiremarine.com/how-to/wiring-a-boat/

But this is where I get lost


Marine Grounding Systems
This article apparently first appeared in Practical Sailor (Oct 1996) (which I now have a subscription to )

https://www.westmarine.com/WestAdvis...unding-Systems

[Wotname]

Originally Posted by Wotname

OK, this helpful - thanks. I'll work though the equivalent circuit later today / tonight to make sure I am fully understanding it and report back

Reporting back - I can't find any circuit condition where a CFCI / RCD / GFI will trip when using a fully floating AC source (assuming the AC source is not faulty internally).

So I remain very doubtful about the claims by some who state otherwise. But I'm happy to be corrected if you can show me the current flows that will trip a CFI / RCD / GFI.

The equivalent circuit should be self explanatory - AC floating source, two loads, leakage faults and a 'ground'.

[Wotname]

Originally Posted by s/v Jedi

99% of what is written in this thread is wrong. A true floating system is at least as safe, if not safer than a polarized system.

I am fast coming around to the same conclusion so I am left wondering why others can't see it. Maybe they are smarter than us though (??). If so, I wish they would demonstrate their understanding.

But that doesn’t mean you don’t need ground wiring to outlets, because you do.

Can you explain why the ground wiring is needed?

First, there can be no ground fault because ground is not a return path. If you throw L1 overboard, nothing happens, no current will flow. Only L2 is a return path.

So the case for ground wiring: appliances think they are connected to a live/hot/phase conductor, a neutral conductor and a ground conductor. Any metal parts of the housing are connected to that ground conductor.

So lets assume it develops a short between the hot and the metal housing. Now that housing is hot and no breaker flips. But that isn’t a problem because when you touch that nothing bad happens. Even if you touch that with onehand and stick the other in the sea, nothing happens because ground is not a return path.

But it can get worse: another appliance develops a short between neutral and the metal housing. Now if you would touch that with one hand while touching the other housing with the other hand.... nothing happens because the breaker flipped as the ground wiring completed that circuit.

Polarized systems don’t even detect a short between neutral and the metal housing. It is a fault that lives on unnoticed.

Here is a problem with a floating system according to many: when you push the test button on a gfci outlet, it fails the test. So let’s see what happens with this button: it is a switch that connects a resistor between the hot and the ground conductors. The current flows back through ground instead of neutral, which is detected and trips the outlet. But with a floating system ground is not a return path so no current flows so it doesn’t trip. You simply don’t need gfci protection.

So finally a real “problem” : you need double pole breakers. That’s it. In return you get a safer system but beware the moment you connect shore power, you must use an isolation transformer. Don’t forget to remove the neutral to ground jumper from it

And please explain why you need double pole breakers as I am not seeing the reasons (yet)?

[Wotname]

Originally Posted by dkenny64

yes, until there is a fault.. then where does the current go?
drop one side the into the water.. salt water...wouldn't this current go somewhere
else besides where its supposed to go?
there by tripping a GFCI?


dkenny64

Hmm... I don't think you fully understand electrical principles.

Current only flows between points of potential difference via a conductive (resistive/reactive) path.

[Wotname]

Originally Posted by dkenny64

hum
no mention for using GFCI for the AC.. of you want safe...use them..
gnd or not..they only check for the delta between hot and neutral...NOT GND
so they work on 2 wire systems just as well as 3 wire..
yes the cheap wall tester will fail..


-dkenny64

Can you show me how / where a delta (current) can occur between hot and neutral with arrangement of fault conditions if the source output is floating.