Grounding an inverter

[sailorchic34]

Originally Posted by transmitterdan

How exactly can there be any "ground fault current" if there is no system safety ground circuit? A person would have to somehow connect themselves between hot and neutral at two different outlets to trip the GFCI. In which case both of them wod trip. But how likely is that?

The ground fault they are talking about is where there is leakage to ground/earth in one of the two conductors. That could be to a ground wire, but could be line to ground/earth also.

The self test button WILL work without a ground as it uses a resistor to short a low amp load of 5ma to neutral after the duel current taps in the GFCI recept. This causes a imbalance in the current in/out and causes the GFCI to trip.

Now you are correct that the plug in testers used by inspectors to test gfci outlets will not work without a ground. But the test button on the device will trip without a ground and the GFCI WILL function without a ground wire connected.

To answer Ron's question: The ground wire is there to provide a low impedance path for current to return to earth. It does not protect from a ground fault, that is if you grab a hot line and are standing on wet grass for example. With a grounded case, the ground provides the return path and not your body. This works because the US power grid is ground "earth" referenced. In many cases the ground will cause enough current to flow to trip the breaker which prevents fire.

So the ground wire will offer personal protection when the ground wire impedance is very low in relation to a higher impedance between hot and ground which your in contact with, for example. However if you standing on wet grass, in the bathtub, or in a improperly grounded pool (alas it happens) then you may be at a low enough impedance where there is still current flow and your shocked.

In the case of a portable genny or a inverter with out ground installed on a boat not connected to shore power , current flows only from leg to leg. There isn't a current path to ground.

This is why every operating room in the USA is installed with an isolation transformer. In reality every transformer is an isolation transformer until the secondary side is grounded to the primary side ground. Then both sides are ground referenced.

In the USA residential power wiring includes a hot, neutral and ground. 240v residential power is 240v leg to leg. The transformer is tapped in the center of the winding which is oddly always at 0 volts or neutral voltage. This is where the name "neutral" comes from. Both Neutral and ground are at the same potential (or very very close anyway). You can touch a neutral wire (Not in circuit) safely as only the hot conducts to ground (US only) EU is hot and full phase leg to leg. Both neutral (US) and ground are at the same potential.

So standard 110v power is hot, 120v leg to 0V neutral and the hot leg cycles 60 times a second between +120 and -120. This is called split phase power as the neutral side does not alternate. This is different then 240v leg to leg where both legs are 180 degrees out of phase from one another.

Inverters generally are NOT center tapped and operate at full phase and not split phase.

[CharlieJ]

#46 is well written with one exception:

To answer Ron's question: The ground wire is there to provide a low impedance path for current to return to earth.

...the ground wire (the Holy Grail of the system on a vessel) is there to provide a redundant, low impedance path for fault current to return to the source and trip an OCPD to clear the fault.

[DotDun]

Originally Posted by sailorchic34

To answer Ron's question: The ground wire is there to provide a low impedance path for current to return to earth.

Ron's question was why is the power system grounded, not about the grounding conductor.

That's the way I read it.

[sailorchic34]

Originally Posted by CharlieJ

#46 is well written with one exception:

...the ground wire (the Holy Grail of the system on a vessel) is there to provide a redundant, low impedance path for fault current to return to the source and trip an OCPD to clear the fault.

I pretty much said that. In that the US, power grid is ground referenced. So saying a low impedence path to ground is also correct. Mind you a ground wire only works when the source is ground referenced. If the source is not ground referenced, say via an isolation transformer, the ground wire does not a thing.

It's called a ground as it's normally connected to a copper rod driven into the ground (or earth). That is why we call it a ground. it does not really connect to source. Well not what I call a source. You'll note that US residential power is 2 conductor. There isn't a ground wire on the primary side of the transformer. So ground goes to earth and not to source. Even on a boat ground should be connected to shore ground that is itself connected to one or more copper rods hammered into ground.

THE OCPD or circuit breaker will trip on a higher then it's listed current with or without a ground. A ground wire may provide, in some cases, an over current path to facilitate the breaker tripping, but not always.

Mind you on a vessel with a metal hull, a ground wire would be the holy grail. On a fiberglass boat that does a fair job of insulating from a ground path anyway, it's less important.

BTW, I'm all for grounds, but sometimes it's just a wire and will not conduct if the source is not grounded. In isolated power where both legs are isolated from ground, there is not a current path to earth. So the only shock potential is if you come in contact with both legs and complete the circuit. Touching one leg offers no shock potential. otherwise there would be a hell of a lot of dead birds in the world.

Way back when all electrical appliances were metal with non polarized plugs, depending on how the plug was inserted in the outlet you could have a 120v potential on the case. The ground was there to pop the circuit breaker (OCPD) in that case or in case of other internal faults.

[transmitterdan]

Some GFCI outlets can self test without a ground. I made a mistake in implying they are all that way.

The OP has an inverter with no ground reference. It has 2 outputs that are most likely both referenced to the case and thus the 12V negative of the battery supply. That is probably where the hum originates. He should definitely not follow the advice to "ground" the neutral. That might start a fire.

[CharlieJ]

SailorChic34:
I agree with you about the genesis of "ground". It is badly misunderstood by the layman and by a lot of professionals.

Mind you a ground wire only works when the source is ground referenced.

it does not really connect to source.

These statements are not true. On recreational vessels the N and G are bonded at every derived source of AC power including gensets, isolation transformers,inverters and inverter/chargers. When on shore power, the N/G bond is at the shore power supply, usually a sub-panel or a transformer.

Regarding the copper rod pounded into the ground for a terrestrial system. It can have a resistance of as much as 25 ohms per the NEC. Not really "low impedance". That is why it is important to understand that the "redundant, low impedance path" provided by the safety ground wire goes back to the source. In a well designed, installed and maintained system on a boat and or shore power system, this value is typically found to be < 1 ohm.

I really recommend that you watch the following two videos by Mike Holt, the preeminent NEC instructor:
http://tinyurl.com/yc727zn2
http://tinyurl.com/yc727zn2

Mind you on a vessel with a metal hull, a ground wire would be the holy grail. On a fiberglass boat that does a fair job of insulating from a ground path anyway, it's less important.

This can be dangerous advice. Here is the setup;
- AC safety ground and DC B- (engine block or its buss) are bonded as required by ABYC E-11.
- There is no safety ground to neutral bond (or no onboard safety ground wire/connection) at the source of AC power.
- There is a fault to ground on a metal case; e.g., battery charger, water heater, etc.
- Danger #1: AC is injected into the water column from the propeller/shaft and any underwater metal components that are bonded for cathodic protection. This becomes the set up for electric shock drowning (ESD). Note that the potential for ESD is much more prevalent in fresh water.
- Danger #2: The component with the fault to the case is at supply potential. The underwater metal is at a much lower potential. If a person touches the case and an underwater metal object or the block, they will experience an electrical shock that may or may not be lethal.

Fundamentally the old adage that "Electricity seeks the path of least resistance." is flawed. Electrical current returns to the source by all paths. If there is no N/G bond then the current will use the water path and the driven ground rod to return to the source. The results of this will be a function of water salinity, ground resistance and ground rod/ground resistance. For sure, the OCPD will not trip. For sure a whole vessel RCD (ELCI in the USA) will trip.

This is important stuff. Watch the videos.

[DotDun]

Corrections/Clarifications:

Originally Posted by sailorchic34

...... In reality every transformer is an isolation transformer until the secondary side is grounded to the primary side ground. Then both sides are ground referenced.

Not true. Auto transformers are metallic coupled, hence no isolation provided. If the primary side of an auto transformer is grounded, then so is the secondary (through that same connection).

Originally Posted by sailorchic34

......You'll note that US residential power is 2 conductor. There isn't a ground wire on the primary side of the transformer. So ground goes to earth and not to source.

We need to use a common set of terms. The neutral (white in the US/blue in the EU) is a grounded conductor. The safety wire (green in the US/yellow-green in the EU) is a grounding conductor. They have different purposes and we need to be able to distinguish between them to have a meaningful conversation.

And, yes, the primary side of the transformer in front of my house has a grounded conductor but no grounding conductor running in parallel with the neutral and hot.

[hellosailor]

Not sailing related, but of interest to anyone who doesn't have an EE degree yet sometimes plays Sparkie:
There's a huge YouTube channel by a mildly deranged Scotsman called "bigclivedotcom". That's the name of the channel. And Clive, who apparently has a great deal of professional experience with electrons, usually is tearing down and analyzing various bits of electrical and electronic gear.
A large number of no-name Chinese devices apparently are being sold with USB ports that are poorly isolated from the mains power, enough to that they can be lethal. These small differences in "grounding" and isolation are among his many points. He's torn down a number of whole-house devices and adapters as well as the portable stuff.
Arguably entertaining AND educational, a good way to learn a bit more about sparky things without any danger of electrocution.

[ottow]

Originally Posted by transmitterdan

Some GFCI outlets can self test without a ground. I made a mistake in implying they are all that way.

The OP has an inverter with no ground reference. It has 2 outputs that are most likely both referenced to the case and thus the 12V negative of the battery supply. That is probably where the hum originates. He should definitely not follow the advice to "ground" the neutral. That might start a fire.

The outputs are not measurably referenced to the case, or the 12V negative supply. I've determined that with a multimeter.
The outlet has a ground pin, but it is not connected to anything internally.
As mentioned earlier, this unit could possibly be 2-phase without a neutral, so connecting what appears to be the neutral wire inside it to something like the 12V negative or the sea may very well kill it. Or have dangerous consequences.

Some pictures below for clarification
I'll be replacing this with a proper Victron unit, something designed for permanent installation in a marine environment.
But such a unit may very well still produce a noticeable hum when used with PA systems as I am doing, so I will investigate 230V filter options.

[Thumbs Up]

Every time this subject comes up I have to wrap my head around it again. I think that there is a problem with the aybc recommended approach of tying all of the grounds together (a/c, d/c, through hulls, etc.). What is are the mostly likely failure modes and could they be the cause of some of the most dangerous scenarios? The fact that any power cable has the potential anywhere on it to connect a hot wire to the safety ground (which is connected to the water through the bonding system) is the problem. If the cable gets pinched or overheats enough to connect the safety ground and the hot wire (but not enough to throw the breaker) then you have a major electrocution hazard in the water. On marine inverters and generators that have a switching relay to connect the safety ground at the source when activated; what about when these relays fail? Electric shock drowning is very real, mostly happens on fresh water, but can be caused by only milliamps of leakage (a hundred times less current than would be required to trip a breaker) The only proven solution is the use of gfci (or similar but better if the threshold is very sensitive like 5 milliamps) The reason that most marinas don't use gfi breakers in their supplies is "nuisance tripping" caused by marine a/c systems (not sure why). Wouldn't marinas be safer if the boats all had isolated grounding (not connected to the water)? If you look up recommendations on this subject you will see that (in the US) is that they are warning to never swim in or near a marina. Apparently, the new aybc code is more aligned with the euro standard and requires gfi protection but the "green" wire seems to be the "holy grail" of what most marine electricians in the US consider to be a "safe" system. But how often is that bonded (to the water) "safety" ground actually the path that the current is taking to the water that is causing an unsafe situation. The "low impedance path to ground" argument in my opinion is faulty because it doesn't take a full short to cause a life and death situation but it does require a full short to trip a breaker. The 15 or 20 amps needed to trip a breaker could kill a person many times over. You could easily start a fire on board with only a few amps of leakage. I think that all boats and marinas should require ground fault protection with a very low (5 milliamp) threshold and if your boat causes nuisance tripping you should not be allowed to connect. I think that boats and marinas would be safer if that ac/dc neutral connection was eliminated as well as through hull bonding. There are many threads here on this forum that dance around this subject. A lot of people will disagree. But lets talk about this. Bring it on! This is the forum, we have a lot of experts here. It is high time we had this debate.

[hellosailor]

"Hi, I'm from the government and I'm here to help you."

https://www.osha.gov/OshDoc/data_Hur..._generator.pdf

One of many opinions about grounding portable stuff. And from another professional view:

Do Honda Portable Generators have to be grounded? - Lighting - Cinematography.com

[CharlieJ]

Thumbs up:

I think that all boats and marinas should require ground fault protection with a very low (5 milliamp) threshold and if your boat causes nuisance tripping you should not be allowed to connect.

NFPA 70-2017 (NEC) Article 555.3 now requires a 30 mAAC RCD on the supply to a marinaI rarely find a recreational vessel with < 6 to 8 mAAC flowing so this is an unattainable requirement.

ABYC E-11 now requires a whole vessel RCD (ELCI) with a 30mAAC/100mS trip sensitivity. That is reasonable. 30mAAC for the entire dock/marina is not reasonable.

Everybody who is truly interested in understanding this swamp, take some time and watch the Mike Holt videos I referenced in #51. It is a worthwhile endeavor.

[Thumbs Up]

4-6 maac is the trip rating on gfci outlets. 30 maac is too high of a threshold to prevent electric shock drowning! I will look at the Mike Holt videos later when I have time. My question is where does the current leakage come from and how to remedy it so that we can install more sensitive protection?

[sailorchic34]

Originally Posted by DotDun

Not true. Auto transformers are metallic coupled, hence no isolation provided. If the primary side of an auto transformer is grounded, then so is the secondary (through that same connection).

And, yes, the primary side of the transformer in front of my house has a grounded conductor but no grounding conductor running in parallel with the neutral and hot.

AH but the inverters we are talking about don't generally have autotransformers. Nor are most residential 240v pole mount either. Lets not muddy the waters. For this discussion we're talking about magnetic coupling but not electrically coupled.

While it is possible to have a primary side leg being tied to ground, most of what I see is two (out of three) primary legs, that is two phases of three on the pole, feeding the primary side of the transformer at a land side address. (Northern California PGE). Really depends on location I guess. On the secondary side, the center tap of a single phase 240v transformer being neutral with a connection/bond to the ground rod located at the base of the pole.

For shore power, most marina's have 208v/3ph (or 480v 3ph at larger marina's). with three separate single phase primary's connected, one to each one leg of the primary side of the transformer. The secondary side generally being a Y/delta with center neutral and 120v leg to center neutral and 208v leg to leg. 480v has 277v leg to neutral There is a ground rod (or three) that the center tap is grounded to.

Larger marinas will extend 480v/3ph power to dock mounted transformers (208v/3 or three 240v/1 transformers) . This gives three single phase 120v feeders to the power pedestals. The ground being taken back to land and all center tap neutrals also bonded to ground.

Now back to the OP's inverter which is a type that is quite common on boats. I have one very similar to it. There isn't a case ground and most likely the secondary side ground may not be connected at all or is connected to one hot leg which at least gives a redundent conductor. But the leg is not at neutral but has a voltage, generally 60v ish.

If this inverter has one or more transformers (generally 2-1/2" square) then the secondary side is isolated from the primary side.

Should a secondary leg touch say the case, is there a shock danger? I say no as there is not a circuit path back to the other leg. Please tell me how a single wire on secondary side touching the case creates a shock hazard?

On the newer switching inverters that do it all with transistors, cap's and filters, there is some potential for hazard shock. but only from the "hot" leg to case. These types mimic split phase 110v, unlike the transformer based inverters that give full phase 110v.

The disadvantage of switching inverters are several, More noise and potential for equipment damage and potential shock hazard. They are made as it saves the manufacturer a few dollars over a transformer inverter.

You can check if your inverter is a transformer, auto transformer or switching type, by measuring resistance between a secondary leg and each primary leg (DC+/-)with power disconnected. A transformer type will show zero (open circuit) reading, which means the secondary side is decoupled. Any resistance measured, means its either switching or auto transformer type. This is for inverters only.

You can do the same measuring resistance between the secondary ground plug and case. If it's open/ no measurement, it isn't connected.

[CharlieJ]

For a 120VAC system there are three return paths to the source; neutral, safety ground and the water. We protect people on the boat with the ubiquitous 4-6mA trip GFCI. In the water, we have to also consider the voltage drop caused by the water column. By the way, a 2 VAC /foot voltage gradient is considered lethal. It takes a significant amount of leakage current to produce this gradient.

For this worst case scenario: 2VAC/foot over a 6 foot human reach (hand to foot) = 12 VAC. A person has a resistance of about 1000 ohms from hand to foot. Ohm's Law for current: I = E/R; I = 12VAC/1000 ohms = 12 mAAC. At approximately 15 mAAC, a victim will suffer paralysis and will be unable to help themselves.

As working limits of leakage we use 100 mAAC in freshwater and 500 mAAC in salt water. A whole boat 30 mAAC/100mSec RCD is a good compromise to allow the normal leakage found on boats and still afford protection.

Note that the NEC now requires a "No Swimming" sign at marinas.

A seminal work on ESD came out a few months ago: "Electric Shock Drowning:Causes and Prevention" by G.S. Cargill.

If you want to see how our marinas are performing; take a look at http://www.boatus.com/seaworthy/assets/pdf/in-water-shock-hazard-mitigation-strategies.pdf

And the resources available at the Electric Shock Drowning Prevention Association website:http://www.electricshockdrowning.org/esd-resources.html

This is a big, complex topic and there is way too much well meaning, but wrong, info out there.