Why Ground Split Phase Residential?

[don_resqcapt19]

... That said, these effects would probably result in quite a small overall leakage current (a few hundreds of uA maybe?)...

Greg

I know that with 480 ungrounded systems fed from transformers in industrial applications will often pull in my solenoid voltage tester when I place it between one of the circuit conductors and earth. It takes ~20mA to pull in the solenoid on the voltage tester. The amount of wire and size of the system does make a difference, but a typical utility residential transformer feeds a number of houses and there would be a lot connected. I expect the current to earth from the conductors of the ungrounded system could very well be high enough to be a serious hazard.

 

[iwire]

Yes, maintenance adds to expense. So does quality control at the factory. So I think you are just repeating part of what I said in a roundabout way.

You are absolutely correct, if you spend enough money and do enough maintance it could work.

But that is total fantasy, homeowners will not maintain it. GFCI and smokes prove that.

So the reality is it would not work.

As for OCPDs, there have been some threads here about many of them not being reliable.

Yes a very small percentage of the millions of breakers fail and they make news on the forum. On the other hand that GFCI study indicated a much much higher failure rate.

And I've seen my share of installations, including my own apartment, where the fuses clearly are not blowing at their rating or they would have done so long ago.

That is not possible, it sounds like you don't understand fuses or overcurrent device trip curves.

And of course, the whole reason that GFCIs were mandated is because OCPDs don't afford very complete personnel protection.

I agree.

To repeat, it's a tradeoff.

I agree, and the most cost effective way to go is grounded systems. Can you point to any countries that use ungrounded systems for dwelling units?

It occurs to me at this point that perhaps the simplest answer to the OP's question could have been "to guard against an open neutral".

What?:?

 

[don_resqcapt19]

... And I've seen my share of installations, including my own apartment, where the fuses clearly are not blowing at their rating or they would have done so long ago. ...

The UL standards for both fuses and breakers permit them to carry 134% of their rating forever. They must open the circuit at 135% of their rating in 60 minutes or less. Most will open the circuit before that amount of current, but they are within the requirements if they don't.

 

[big john]

Never been a fan of ungrounded systems. Generally, they aren't a safety measure, they are a continuity-of service measure, and that's not an important consideration in residential.

For the reasons others have already stated, I would agree that from a reliability standpoint, there is more likelihood than an ungrounded system in a house would allow a shock hazard to exist without clearing the fault, than the existing grounded system will: Why make a system that relies on many fancy electronics when you can rely on a few passive components instead?

 

[jaggedben]

You are absolutely correct, if you spend enough money and do enough maintance it could work.

But that is total fantasy, homeowners will not maintain it. GFCI and smokes prove that.

So the reality is it would not work.

It does work, but countries that do it don't rely on homeowners. See below.

I agree, and the most cost effective way to go is grounded systems. Can you point to any countries that use ungrounded systems for dwelling units?

I think the point applies to any countries that wire dwelling units with TT systems. So apparently Japan, France, Argentina, and I'm sure others. Strictly speaking the neutral might be grounded but that is not relied upon for fault protection. They use GFPs on the utility side (which BTW the rest of the English speaking world calls Residual Current Devices, or RCDs).

What?:?

I misspoke in terms of suggesting that it's the only reason, but guarding against an open neutral definitely required in our systems and not in the countries mentioned above.

Split phase 120/240 for residential service is pretty much a North American thing.

 

[mbrooke]

Grounding a leg allows for a single fault to clear an OCP. It reduces over voltages caused by capacitive potential between phase and hot during an arcing ground fault. It prevents one leg from rising to 240 during a hot to ground fault.


Of course we could add a sufficient resistor or reactance to the neutral before grounding it to reduce the risk of over voltages from arcing ground faults, but you would need ground detectors at each residence. Ground detectors arent something people will pay attention to, and even if they did, how would you be able to tell go about telling one of your 20 neighbors they have a ground fault somewhere? One less would also rise to 240, which means any shock will be far more dangerous.


FWIW, Norway has used ungrounded power for residential extensively and most of these systems have standing ground faults on them, which defeats the purpose of having the system ungrounded to begin with. Although we aren't to innocent either, due to the lack of branch circuit RCD protection most of our power systems have standing neutral to ground faults on them.

 

[mbrooke]

Another thing to remember is that with a grounded neutral you know what voltage the insulation in cords, motors, lamp sockets, breakers and other equipment needs to withstand. With an ungrounded system that voltage can get pumped up by non linear leakage to many times the nominal voltage.
Most important though is what iwire stated, namely that the first fault on the isolated system will cause a reference and current path to ground. That fault will NOT trip a GFCI since no current will flow. Now you have an "ungrounded" system that is no longer as safe as people assume.

Not true. Current will flow depending on the phase capacitance to ground. Its the notion that ungrounded systems dont produce fault current that gives the people they are safe to work live. On a big enough system it is enough to kill.

 

[cadpoint]

The Host has a ton of graphics here! While the exact OP seems to have changed and
is not covered with exact graphics or the/a matching Code article,
just send your friend there and see what falls out...

 

[mbrooke]

Thanks for the replies, folks!

Sparky, to clarify, I am referring to the entire 120/240V circuit from the low side of the distribution transformer, including any connected residences.

I can buy the non-linear leakage current and maybe the capacitive/inductive coupling current arguments. That said, these effects would probably result in quite a small overall leakage current (a few hundreds of uA maybe?)?would it be practical to keep the neutral reference to ground through a high-impedance ground? This might be high impedance enough to prevent fatal shocks through a grounded person who might touch an energized part of the circuit, but low enough to keep the neutral from floating above some set voltage (say <50 V) given a large-margin assumption of total expected leakage and coupling currents.

Greg

The average residential system assuming 5 to 8 houses on a transformer with average equipment and average cable insulation will produce roughly 50 to 80ma (guesstimated calculation), maybe over 450ma of current leakage in larger systems. Above a safe 5 milli amps. The resistive leakage would of course be around micro amps, but the capacitive reactance will be very well above micro amps. In any case a resistor of a greater value would have to be inserted in between the XO to ground to prevent an arcing fault from causing over voltages. It is however possible to insert a tuned reactor between the XO and ground instead of a fixed resistor to cancel out most of the capacitance to ground. An automatic tuning reactor (Peterson coil) (automatic since the network will become over/under compensated as circuits are plug/unplugged switched on and off), could when done right bring the value down to about 5 to 10ma. Difficult but doable. Cost will be a factor over anything else since both the device to adjust the reactor and the reactor itself are capitol investment. A resistor in parallel with the reactor may help add a unity power factor current flow during a fault to help standard RCD (GFCI) clear the fault. Its highly recommended the GFI/RCD clear a fault then just alarm (indicate it) since chances are the average home owner might not fix it. Having the ground fault stand eliminates any benefit of having such a system.

Also keep one thing in mind, US house hold screw shell sockets need to have the outer shell kept at ground potential since when the bulb is screwed/unscrewed its possible to inadvertently contact the shell. These screw shells should be outlawed, but, as with the US and Canada old habits die hard. A ground fault on such a system could place 240 volts on the shell, which of, course why the RCD/GFI is even more important.


However, you have all the right ideas and all the right thinking :happyyes: :thumbsup: Such a system could be pulled off, and it would indeed be far safer. And even this aside there is so much room for improvement in other areas of modern power systems (dont even get me started:rant::lol::happysad But cost, and the need to change new installation practices as well as upgrading old ones to current standards are where people will grumble. But then again remember few people have the gift to change something for the better:happysad:.

 

[mbrooke]

What is it with this fascination of always thinking what we do not have is better?

We use a solidly grounded system because after near one hundred years of using electricity it has proved to be the best among all possible systems.

Okay..... sure, there are one or two cases where something different can be called out; but for most of the world most of the time, actually almost all of the time, grounded wins.

Because sometimes we dont always have it better. Assuming what you have presented in front of you is the best it can be is the end of human progress. Thinking, questioning, studying and learning from failure are what lead to new technology or at least the betterment of something existing. Had it not been for a few guys with some pretty radical ideas we would be without I phones still useing type writers.

And lets face it, (grounding aside) when one really sits down and looks at North Americas power systems compared to IEC world there is a lot of room for improvement. In the IEC world for example electricians are required to megger all new circuits. Americans argued we are careful and dont need to, our installation methods are fine. Then came along AFCIs and began exposing wiring errors on every other new home that wouldve been caught with meggering in the first place. Point is there is always room for improvement be it a practice or new technology.

 

[GoldDigger]

Not true. Current will flow depending on the phase capacitance to ground. Its the notion that ungrounded systems dont produce fault current that gives the people they are safe to work live. On a big enough system it is enough to kill.

There can be enough capacitive current to trip 20ma or 6ma GFCI, but you cannot count on it. We may be looking at a GD rather than a GFCI. And if the fault persists, the available current when someone becomes the second fault by touching what they think is a safe conductor is higher than just the capacitive current. At that point you would trip a GFCI for sure.

 

[mbrooke]

There can be enough capacitive current to trip 20ma or 6ma GFCI, but you cannot count on it. We may be looking at a GD rather than a GFCI. And if the fault persists, the available current when someone becomes the second fault by touching what they think is a safe conductor is higher than just the capacitive current. At that point you would trip a GFCI for sure.

You would at that point. But Id still want the GFI to clear the first fault.


One thing I just want to note, more like ask, a regular GFI might trip on a neighboring circuit having a fault since since they really cant establish direction? (Fault current comes from the capitive reactance from other neighboring circuits, so a fault on one pulls current from others)

 

[GoldDigger]

The GFCI defects imbalance in the current flowing through it. Any capacitive or other fault current on the line side of the GFCI cannot affect the current through the GFCI. If you have a loop circuit, as can be found in the UK, a GFCI receptacle in the middle of the loop might trip on leakage to ground (EGC) on either side, the only way to both provide GF detection and avoid nuisance trips is to use a GFCI (RCD) breaker or to use a GFCI receptacle to protect only what is plugged into it.

 

[mbrooke]

The GFCI defects imbalance in the current flowing through it. Any capacitive or other fault current on the line side of the GFCI cannot affect the current through the GFCI. If you have a loop circuit, as can be found in the UK, a GFCI receptacle in the middle of the loop might trip on leakage to ground (EGC) on either side, the only way to both provide GF detection and avoid nuisance trips is to use a GFCI (RCD) breaker or to use a GFCI receptacle to protect only what is plugged into it.

True, but in a pure ungrounded system all GFCIs will see an imbalance from current flowing from the other feeders. Each circuit after the GFCI has its own capacitance, when one circuit faults the other GFCIs see a current flow as well, so they all must be a directional variety. I guess Im thinking of medium voltage distribution feeders where substation breakers must take directional power flows during a ground fault:

 

[kwired]

Grounding a leg allows for a single fault to clear an OCP. It reduces over voltages caused by capacitive potential between phase and hot during an arcing ground fault. It prevents one leg from rising to 240 during a hot to ground fault.


Of course we could add a sufficient resistor or reactance to the neutral before grounding it to reduce the risk of over voltages from arcing ground faults, but you would need ground detectors at each residence. Ground detectors arent something people will pay attention to, and even if they did, how would you be able to tell go about telling one of your 20 neighbors they have a ground fault somewhere? One less would also rise to 240, which means any shock will be far more dangerous.


FWIW, Norway has used ungrounded power for residential extensively and most of these systems have standing ground faults on them, which defeats the purpose of having the system ungrounded to begin with. Although we aren't to innocent either, due to the lack of branch circuit RCD protection most of our power systems have standing neutral to ground faults on them.

I totally agree with the bold part above. What we have is somewhat of a compromise to make it the most practical for the most users, plus the fact that it has been the way it is for so long it would be rather expensive to just suddenly require everyone to adapt to something different.

It would be somewhat easier to adapt if each individual service were supplied by an individual transformer, you could at least leave old systems as is but any new ones need converted, but not many POCO are that willing to spend the $$ for separate transformers for multiple small capacity services in close proximity to one another

 

[mbrooke]

I totally agree with the bold part above. What we have is somewhat of a compromise to make it the most practical for the most users, plus the fact that it has been the way it is for so long it would be rather expensive to just suddenly require everyone to adapt to something different.

It would be somewhat easier to adapt if each individual service were supplied by an individual transformer, you could at least leave old systems as is but any new ones need converted, but not many POCO are that willing to spend the $$ for separate transformers for multiple small capacity services in close proximity to one another

Well spoken Should've checked for grammatical errors before hitting send:lol::ashamed:


Individual transformers would make it easier to employ, but it would be more material (money) and energy since many smaller transformers are less efficient then a larger one. As is 120 volts is doing us a disservice since the average pole transformer only serves one average 5 homes. While that doesn't seem to bad, some countries that employ 240 volts set a large 3 phase 415Y transformer and feed over 150 homes. If any businesses are close by to the transformer those are connected to it as well. This actually turns out to be both very economical and energy efficient. While everyone is a away from home working the vast majority of the power is consumed by the businesses, as people close up shop and head home the load simply does a 'polar shift' so to say going from work places over to homes. That way poco transformers always see steady loading over a 24 hour period usually not going below 40%, rarely above 125%. As appose the American setups where resi transformers will see less than 1% loading when no one is home to a 200 to 300% overload as everyone is coming home for dinner. Commercial typically sees around 100% loading while open for business (with short 125 to 150% overloads during peak hours) to only 2% loading when closed at night.