Marine AC - Why Tie Neutral and Earth?

[Wotname]

Originally Posted by Thumbs Up

.......
We have established that dedicated GFCI outlets greatly improve the safety and would be considered best practice. The question (once again to be clear) is: Where is the best place to lead that third wire (the outlet grounds) on a fully isolated IT (floating) distribution system. Your opinion is appreciated. There doesn't seem to be a standard for this type of system on a small yacht and we are trying learn what we can because once we deviate from standard practice, any accident would implicate us.
The question is directed at Jedi but of course want to hear what the OP thinks and anyone else who is following this. (and goboatingnow and Steve DAntonio)

This diagram is what the OP (me) has distilled from this thread (and the earlier 2013 thread). It should be self explanatory).

I make a note that the downstream RCD outlets are single outlets - this should satisfy the very extremely rare occurrence of the danger of two metal cased loads plugged in a double outlet each with a opposite Line to Case fault and the cases being connected by a human...

Additionally I can't see any reason to connect the "third conductor' (blue line) to anywhere else.

EDIT: I have not included two pole overload circuit breakers as I have assumed readers would realize they are part of the AC source output.

Disclaimer - Not to used for any connection to shore power

[s/v Jedi]

Originally Posted by Wotname

This diagram is what the OP (me) has distilled from this thread (and the earlier 2013 thread). It should be self explanatory).

I make a note that the downstream RCD outlets are single outlets - this should satisfy the very extremely rare occurrence of the danger of two metal cased loads plugged in a double outlet each with a opposite Line to Case fault and the cases being connected by a human...

Additionally I can't see any reason to connect the "third conductor' (blue line) to anywhere else.

EDIT: I have not included two pole overload circuit breakers as I have assumed readers would realize they are part of the AC source output.

Disclaimer - Not to used for any connection to shore power

The ground wiring also need to connect to metal parts of appliances not connected through an outlet, like the inverter housing, the aluminum breaker panel etc.

You need not worry about the fault scenario you describe with double outlets: when this would happen, it would short L1 and L2 through the ground conductor in the appliance cables and the breaker would trip. You do need to make this gfci outlets.

[s/v Jedi]

I have talked about “gfci downstream outlets”. This is a feature used by American gfci outlets, I’m not sure if these exist elsewhere.

What these are: regular outlets, connected to the -output- terminals of a gfci outlet. This is so that one gfci outlet can protect multiple outlets. I recommend this not to be used, especially not for floating power distribution.

Even for regular polarized (which means grounded neutral) systems these are a nuisance because outlets end up without power while the breaker is good and the tripped gfci outlet can’t be found or it is not realized that another outlet has gfci which tripped and took power away from other outlets.

[s/v Jedi]

Originally Posted by Thumbs Up

I knew all of this except for the part about the EMS.
.
I admit to being out of my league on autotransformers. I thought it was more like an isolation transformer (I do understand what it is for just not how it works)



I thought that ABYC doesn't recommend isolated power supply (IT) systems period. Neutral is normally bonded to ground at the source. With an IT supply, the source is the output of the transformer. It would seem normal to connect the outlet safety ground wires here, before the output breaker. Since there is no real neutral then either output would do, say L2. If the output breaker was also an RCD then that connection would act as a fault protective wire, because current on it would create an imbalance and trip the RCD. If it was only connected to the chassis of the transformer it would not have the protective feature. This is what I was studying before I decided to try to critique your system.

It started as more of a question because I was thinking about where to tie the outlet grounds. I was looking for an example when I found the other thread. Goboatingnow described it this way:


Option B is what I am referring to.
So without getting wrapped up on how to apply this to your setup, Jedi, the question that I was looking to verify (and partially did from what goboatingnow had wrote) is: Is this a good way to do it? (and if not why, and what is the safest possible option)
My other takeaway, as an alternative was that if you happened to have a center tap, that you could lead the outlet grounds there.
We have established that dedicated GFCI outlets greatly improve the safety and would be considered best practice. The question (once again to be clear) is: Where is the best place to lead that third wire (the outlet grounds) on a fully isolated IT (floating) distribution system. Your opinion is appreciated. There doesn't seem to be a standard for this type of system on a small yacht and we are trying learn what we can because once we deviate from standard practice, any accident would implicate us.
The question is directed at Jedi but of course want to hear what the OP thinks and anyone else who is following this. (and goboatingnow and Steve DAntonio)

Okay, I get it now... was a bit slow. What they mean is to create a chassis ground like I do, then connect either L1 or L2 to that. The reasoning behind this is that absolutely nothing happens when an isolation fault to ground occurs, so they do it on purpose (let’s assume L1 is grounded). Now that this is done, when L2 has an isolation fault, the breaker trips.
Even though this is exactly the same as my setup, this is done to prevent the situation where you have an isolation fault but it remains unnoticed.

Imo this is flawed. Imagine having this and an appliance develops an isolation fault between L1 and ground. It is unnoticed as well because that connection was made on purpose. The reasoning that’s it’s better to know you have this at all times does not stack up to the loss of safety by taking that step. It would be much better to detect the first fault, which is easy to do.

[TeddyDiver]

Originally Posted by s/v Jedi

Okay, I get it now... was a bit slow. What they mean is to create a chassis ground like I do, then connect either L1 or L2 to that. The reasoning behind this is that absolutely nothing happens when an isolation fault to ground occurs, so they do it on purpose (let’s assume L1 is grounded). Now that this is done, when L2 has an isolation fault, the breaker trips.
Even though this is exactly the same as my setup, this is done to prevent the situation where you have an isolation fault but it remains unnoticed.

Imo this is flawed. Imagine having this and an appliance develops an isolation fault between L1 and ground. It is unnoticed as well because that connection was made on purpose. The reasoning that’s it’s better to know you have this at all times does not stack up to the loss of safety by taking that step. It would be much better to detect the first fault, which is easy to do.

Concur with that. It's basicly like a deliberate fault with zero gain..

[s/v Jedi]

Addendum to previous post:

So that discussed L1 and L2, but what about the center tap from the auto transformer used to create 120V. I make the mistake of calling it a neutral because not being grounded, it isn’t neutral.

The auto transformer is a single winding so the center tap is galvanically connected to both L1 and L2. When you have multiple auto transformers, all these center taps are galvanically connected to L1 and L2 but they are all a little different to each other, which is why you can’t connect them together, nor to a center tap at the IT secondary winding or a genset output winding etc.

I’m just gonna call this auto transformer center tap N. N is only used for 120V distribution, where it is paired with either L1 or L2 through a double pole breaker. You alternate L1 and L2 so as to balance load between these. Any unbalance that is left is taken care of by the auto transformer which has a maximum current rating, which is incentive to balance as good as possible.
Whenever N shorts to either L1 or L2, a full power short circuit current happens which trips the distribution breaker. In my case, those are in the 15/16A range while the auto transformer is 32A. If a smaller auto transformer is used, the breaker on that can trigger before the distribution breaker does.
My inverters have an OUT2 which is only 16A. OUT2 is only energized when the inverter is passing incoming power and OUT2 is de-energized when it starts inverter mode. This means that a short on a distribution group that uses OUT2 may trigger one of those breakers instead of the distribution breaker. I have two inverters in parallel, so this would only happen when one inverter is offline.

Now I write this, I wonder how those 120-240 moving transformers work... the ones that allow a 240V appliance to run off a 120V outlet or the other way around. I think that ground wires only connect to it’s metal parts, meaning that the output has no grounded neutral. Same thing with a portable generator: it does not ground neutral.

[s/v Jedi]

Here’s a good page on ground fault detection. Note the listed advantages of floating power distribution at the end!
https://www.benderinc.com/know-how/t...ult-monitoring

[Jammer]

Originally Posted by Wotname

This diagram is what the OP (me) has distilled from this thread (and the earlier 2013 thread). It should be self explanatory).
[...]
Additionally I can't see any reason to connect the "third conductor' (blue line) to anywhere else.

Well, lots of problems actually.


You have to realize that you can't make the ground actually float unless you're in some sort of medical/industrial environment where all the connected loads are designed with that in mind.


So for example your laptop power supply, and your microwave, and most other electronics, and your surge-protector power strip if you have one, all have MOVs connected from ground to neutral and from ground to hot. MOVs by design conduct once a threshold voltage is exceeded. So when you have an overvoltage event -- static electricity, maybe a fault in your microwave oven, maybe a kinky guest waving a violet wand around -- it's going to be the MOVs in your electronics that short neutral to ground. You get some nuisance GFCI trips and maybe damage the MOV.


While lighting protection and the effects of lightning discharge are controversial topics on CF, I believe it is widely accepted elsewhere that isolated electrical systems are more susceptible to lightning damage than those that have a bonded neutral. At utility scale, the practice of grounding the neutral return conductor at each pole was originally instituted (and continues to be followed in most places) as a strategy to mitigate lightning damage.


You can run into problems with DC, particularly high voltage DC. Your GFCIs won't trip on a DC imbalance. So if your microwave has a fault that dumps 2000 volts between ground and hot/neutral someone could get a lethal shock that would have been prevented by bonding ground and neutral together. Or you could get equipment damage if you exceed the breakdown voltage on a motor winding or something.


I think you're overstating the safety benefits because there will always be leakage anyway. In practice you're creating a system where each conductor is 60 volts to ground (in the absence of a fault) because of the cumulative leakage and capacitive coupling. If that's what you want you're better off using a center-tapped 120v transformer and grounding the center tap as is the practice on some construction sites in England. The result would be similar to the use of zigzag grounding transformers in some industrial environments to create a midpoint ground for a higher voltage (usually 460 volt) delta system in that you keep the ground-to-phase voltage as low as possible while still having a ground reference that avoids the risks inherent in floating systems.

[s/v Jedi]

Oh, that site is a good source of info: https://www.benderinc.com/fileadmin/..._en_iso685.pdf

Itturns out that floating systems are much more common than I even thought!

[Wotname]

Originally Posted by s/v Jedi

You need not worry about the fault scenario you describe with double outlets: when this would happen, it would short L1 and L2 through the ground conductor in the appliance cables and the breaker would trip. ......

Only if the fault conditions were low enough in resistance to allow the current to exceed the breaker threshold. Higher resistance leakage type faults could allow enough of a PD to be dangerous. Single RCD outlets minimise such risks.

However I agree such events would be rarer than the moon rock you previously quoted.

[Wotname]

Originally Posted by Jammer

Well, lots of problems actually.


You have to realize that you can't make the ground actually float unless you're in some sort of medical/industrial environment where all the connected loads are designed with that in mind.


So for example your laptop power supply, and your microwave, and most other electronics, and your surge-protector power strip if you have one, all have MOVs connected from ground to neutral and from ground to hot. MOVs by design conduct once a threshold voltage is exceeded. So when you have an overvoltage event -- static electricity, maybe a fault in your microwave oven, maybe a kinky guest waving a violet wand around -- it's going to be the MOVs in your electronics that short neutral to ground. You get some nuisance GFCI trips and maybe damage the MOV.


While lighting protection and the effects of lightning discharge are controversial topics on CF, I believe it is widely accepted elsewhere that isolated electrical systems are more susceptible to lightning damage than those that have a bonded neutral. At utility scale, the practice of grounding the neutral return conductor at each pole was originally instituted (and continues to be followed in most places) as a strategy to mitigate lightning damage.


You can run into problems with DC, particularly high voltage DC. Your GFCIs won't trip on a DC imbalance. So if your microwave has a fault that dumps 2000 volts between ground and hot/neutral someone could get a lethal shock that would have been prevented by bonding ground and neutral together. Or you could get equipment damage if you exceed the breakdown voltage on a motor winding or something.


I think you're overstating the safety benefits because there will always be leakage anyway. In practice you're creating a system where each conductor is 60 volts to ground (in the absence of a fault) because of the cumulative leakage and capacitive coupling. If that's what you want you're better off using a center-tapped 120v transformer and grounding the center tap as is the practice on some construction sites in England. The result would be similar to the use of zigzag grounding transformers in some industrial environments to create a midpoint ground for a higher voltage (usually 460 volt) delta system in that you keep the ground-to-phase voltage as low as possible while still having a ground reference that avoids the risks inherent in floating systems.

Appreciate you posting these POVs - thanks.

[s/v Jedi]

Originally Posted by Wotname

Only if the fault conditions were low enough in resistance to allow the current to exceed the breaker threshold. Higher resistance leakage type faults could allow enough of a PD to be dangerous. Single RCD outlets minimise such risks.

However I agree such events would be rarer than the moon rock you previously quoted.

Agreed.

[s/v Jedi]

Originally Posted by Jammer

Well, lots of problems actually.


You have to realize that you can't make the ground actually float unless you're in some sort of medical/industrial environment where all the connected loads are designed with that in mind.


So for example your laptop power supply, and your microwave, and most other electronics, and your surge-protector power strip if you have one, all have MOVs connected from ground to neutral and from ground to hot. MOVs by design conduct once a threshold voltage is exceeded. So when you have an overvoltage event -- static electricity, maybe a fault in your microwave oven, maybe a kinky guest waving a violet wand around -- it's going to be the MOVs in your electronics that short neutral to ground. You get some nuisance GFCI trips and maybe damage the MOV.


While lighting protection and the effects of lightning discharge are controversial topics on CF, I believe it is widely accepted elsewhere that isolated electrical systems are more susceptible to lightning damage than those that have a bonded neutral. At utility scale, the practice of grounding the neutral return conductor at each pole was originally instituted (and continues to be followed in most places) as a strategy to mitigate lightning damage.


You can run into problems with DC, particularly high voltage DC. Your GFCIs won't trip on a DC imbalance. So if your microwave has a fault that dumps 2000 volts between ground and hot/neutral someone could get a lethal shock that would have been prevented by bonding ground and neutral together. Or you could get equipment damage if you exceed the breakdown voltage on a motor winding or something.


I think you're overstating the safety benefits because there will always be leakage anyway. In practice you're creating a system where each conductor is 60 volts to ground (in the absence of a fault) because of the cumulative leakage and capacitive coupling. If that's what you want you're better off using a center-tapped 120v transformer and grounding the center tap as is the practice on some construction sites in England. The result would be similar to the use of zigzag grounding transformers in some industrial environments to create a midpoint ground for a higher voltage (usually 460 volt) delta system in that you keep the ground-to-phase voltage as low as possible while still having a ground reference that avoids the risks inherent in floating systems.

Transient protection is primarily done with MOV’s between neutral and hot, which become L1 and L2. Even with the ground included, I don’t see the problem with any of these situations?

About lightning and grounded neutral: I don’t agree. Most damage from lightning strikes comes from near hits where the transient spike travels through earth. Also, where damage is limited because of a grounded neutral, this is because of grounding rods driven deep into the earth. So again, we are comparing shore based power to maritime mobile power which is apples to oranges. For biats we all know that the best cause of action during a lightning storm is isolation: remove shore power cord, turn the double pole breakers off, remove the coax cables.

[alop37]

In shore based systems the National Electrical Code requires that the neutral be grounded (earthed) ONLY at the service entrance point. No where else in the system should it be grounded.
The shore power supply neutral to your boat should NOT be connected to the boat's ground. This is true whether your boat has a galvanic isolator, a polarizing transformer, or an isolating transformer as defined in ABYC E11. If the boat is provided with an isolating transformer, one conductor of the secondary winding should be grounded to the boats ground but NOT to the incoming neutral or ground conductor in the shore power cable.

[Jammer]

Originally Posted by s/v Jedi

Transient protection is primarily done with MOV’s between neutral and hot, which become L1 and L2. Even with the ground included, I don’t see the problem with any of these situations?

This is not my experience in the many designs I have reviewed where a ground connection is present and MOV surge protection is used. There is always a neutral-to-ground MOV and a hot-to-ground MOV. On smaller electronics that don't have a ground connection to the mains plug then yes, the MOVs will be placed across the two lines.

About lightning and grounded neutral: I don’t agree.

No surprise there, and that's fine if it's your boat. I don't believe it's prudent to advocate that others follow your lead, though.



Where we disagree is that I don't believe it's wise to second-guess decades of field practice and replace it with personal theory or the piecemeal borrowing of highly specialized practices from other, very different applications.



I'm not smart enough to figure out lightning behavior from first principles. Maybe you are.