Why the Limit Here?

[paulengr]

How do you define an effective ground fault current path though? If you mean direct contact, sure. Yet thats not ULs or the CMP's reasoning behind GFP and GFCIs in many locations.

I would define it two ways. First is resistance needs to be low enough to trip the ground fault protection in the event of a fault with a grounded metal surface. Obviously this is a moving target. We can always make ground fault detection more sensitive especially if in doing so we increase the minimum trip time. The GEMI approach actually looks at effective grounding from the phase overcurrent perspective. With ground fault detection and tripping obviously much lower limits are acceptable. The GEMI approach applies to a basic thermal magnetic breaker on say a 300 foot conduit run.

Second is to restrict transients to some reasonable maximum such as 1.0 per unit. So this implies draining the system capacitance. In high resistance grounds we want to set the resistor to no larger than 300% of the system capacitance Xc or it acts like an ungrounded system with the attendant transient issues. This is a very reasonable maximum for effective grounding. NEC requires solidly grounded systems under 250 V but that certainly doesn’t restrict “effectively grounded” at long distances from the transformer.

Again don’t forget that “effectively grounded” in terms of Earth grounding is a short distance issue. Earth resistance is inversely proportional to distance so that “remote Earth”, say a mile away or so, is under 1 ohm and certainly better than any intentional conductor. Edison was right if you get far enough apart and “stray voltage” is acceptable.

 

[GoldDigger]

Again don’t forget that “effectively grounded” in terms of Earth grounding is a short distance issue. Earth resistance is inversely proportional to distance so that “remote Earth”, say a mile away or so, is under 1 ohm and certainly better than any intentional conductor. Edison was right if you get far enough apart and “stray voltage” is acceptable.

I am not sure quite how to interpret that statement. The classic ideal theoretical formulation is that the electrode resistance to remote earth is a constant once you get beyond the direct zone of influence of an earth electrode, while the earth portion of the resistance is a constant rather than changing at all with distance. It neither increases nor decreases with distance once the effective electrode resistance at each end is subtracted.
Can you explain the derivation of the result you describe? The situation for a multiply grounded system may be different since the longer distance adds more intermediate earth electrodes with metallic paths between them.

 

[mbrooke]

Equipment falls into the broad definition of property and the purposes of the code is "protect persons and property" from the hazards of electricity.

Going by Black's Law dictionary, perhaps. A fire from building wiring would certainly destroy a lot of consumer goods. However, little if anything in the code dictates what goes on inside an appliance or what gets plugged in.

Its clear the zone of electrical protection doesn't extended into appliances and consumer goods...

However I'd say your thinking is in parellel to the CMPs. They'd love to mandate fuses and receptacle devices that can respond to anomalies within consumer goods when any such requirement should originate from UL, not the NEC.

 

[mbrooke]

If you go back in time, the GFCI protection was all for cord and plug connected equipment. One of the driving factors was the fact that it is much more common for the EGC to fail on cord and plug connected equipment than on hard wired equipment. That is still the case. GFCI provides people protection for cord and plug connected equipment with a ground fault and a failed EGC. No testing of the fixed wiring will prevent damage to the EGC in the cord.

Of course! Most cord and plug equipment didn't even have an EGC at one point in time.

So when I have real world scenarios where an EGC is unlikely to break, should I still be required to install a GFCI because I might run said feeder or branch circuit a really long distance without taking things like voltage drop into consideration?

 

[mbrooke]

Not only that but even in a situation where it’s all an effective path...e.g. a totally nude body on a metal deck touching a grounded metal surface, it takes SIGNIFICANT current considering that one very common mistake with the IEC data is that human resistance drops to 600 ohms minimum depending on applied voltage (its nonlinear) where at 150 V it’s around 1200 ohms or more.

Well I wouldn't call that effective because it won't trip a standard thermal mag breaker. But I get what you're saying.

As compared to the more conservative constant value of 1,000 ohms used in IEEE standard 80 for substations or in MSHA mining standards

.

The IEEE standards for substations basically force gravel over a substation's ground grid, where as the IEC takes a worker's boots and gloves into account and as such many EU substations have grass instead of gravel over their ground grid.

In simple terms the IEEE worker is significantly more protected during a fault at a substation than an IEC worker.

So, if the IEEE (for the sake of the argument) miscalculated the human body impedance, would it ever be called into question considering all North American substation have an added insulators on top of their ground grid?

Just some food for thought...

After 5 seconds if fibrillation has not occurred it won’t. A more realistic maximum is 100 mA if you plug in the numbers at even a couple seconds. The low 20 mA numbers are at 5 seconds. IEC screwed this one up royally. They mixed things together that don’t go together. Under 0.1 seconds it can’t cause fibrillation because the pulse is to short to interrupt the heart rhythm so at that point the limit is on organ damage (1 A+, based on ohmic heating) NOT fibrillation.

Blame UL...

Even if this is true (it might be, I'm open to all data that challenges established standards) it doesn't change the fact that under the NEC someone can legally run a circuit such that it never trips a breaker.

As an unfortunate example we used 25 A resistors on a 7200 V medium voltage system so the resistor is 166 ohms and the line to ground voltage is 4160. Tripping was set to 1-2 seconds. Mining application. With the resistor in series it will not cause fibrillation. Due to confusion over which system he was on and some serious LOTO issues a neighbor took a hand to knee hit on this system long to ground and lived to tell about it.

I'd need more details. Clothes, weight, source impedance, bonding of the environment around him, ect all play a role and may not reflect worse case.
Keep in mind that the IEC typically allows for 5 second disconnections times on circuits over 32 amps, some countries over 100amps. The idea is that with larger wire more current will flow due to lower Z, pulling down the voltage at the spades of the transformer.

As such X0-X1 might dip down to 100 volts instead of holding 277 on a 200 amp feeder fault vs a 20 amp branch circuit fault.

So instead of 139 volts to remote earth at point of contact, the person only "sees" 50 volts.[/QUOTE]

 

[mbrooke]

I would define it two ways. First is resistance needs to be low enough to trip the ground fault protection in the event of a fault with a grounded metal surface. Obviously this is a moving target. We can always make ground fault detection more sensitive especially if in doing so we increase the minimum trip time. The GEMI approach actually looks at effective grounding from the phase overcurrent perspective. With ground fault detection and tripping obviously much lower limits are acceptable. The GEMI approach applies to a basic thermal magnetic breaker on say a 300 foot conduit run.

Phase over current is the cheapest option, and is already present in practically 100% of all LV installations.

Do you really want a commercial panel filled with 80 GFCI/GFP breakers? Or a sub main that takes out a chunk of a building every time someone turns off a light? Or a GFP device mounted near every piece of equipment?

Second is to restrict transients to some reasonable maximum such as 1.0 per unit. So this implies draining the system capacitance. In high resistance grounds we want to set the resistor to no larger than 300% of the system capacitance Xc or it acts like an ungrounded system with the attendant transient issues. This is a very reasonable maximum for effective grounding. NEC requires solidly grounded systems under 250 V but that certainly doesn’t restrict “effectively grounded” at long distances from the transformer.

Of course. But you can't eliminate that capacitance to ground. Xc=1/2pifc only plays stable during steady state conditions... meaning a GFCI/GFP that holds while the machine is running will trip when a relay sparks open...

Again don’t forget that “effectively grounded” in terms of Earth grounding is a short distance issue. Earth resistance is inversely proportional to distance so that “remote Earth”, say a mile away or so, is under 1 ohm and certainly better than any intentional conductor. Edison was right if you get far enough apart and “stray voltage” is acceptable.

So then why do 12kv uni grounded distribution systems behave like multi grounded systems near the stations, low impedance 1 mile away, medium impedance 6 miles, high impedance 15 miles out and totally ungrounded 25 miles out?

 

[GoldDigger]

So then why do 12kv uni grounded distribution systems behave like multi grounded systems near the stations, low impedance 1 mile away, medium impedance 6 miles, high impedance 15 miles out and totally ungrounded 25 miles out?

How about the impedance of the circuit conductors? The resistance between two well constructed earth electrodes 25 miles apart will not be significantly higher than the resistance between the same two electrodes 1/4 mile apart. But the impedance to earth of a metallic line 25 miles long, grounded at the other end, can be quite significant.

If you earth both ends of one conductor of a pair you should have a lower end-to-end impedance than if only one end is earthed. And you will not see a big difference if you break the long conductor.

However, looking at transmission line theory, a balanced center grounded line pair or balanced three phase line group will behave differently than a single isolated line with an earth return.

 

[mbrooke]

How about the impedance of the circuit conductors? The resistance between two well constructed earth electrodes 25 miles apart will not be significantly higher than the resistance between the same two electrodes 1/4 mile apart. But the impedance to earth of a metallic line 25 miles long, grounded at the other end, can be quite significant.

If you earth both ends of one conductor of a pair you should have a lower end-to-end impedance than if only one end is earthed. And you will not see a big difference if you break the long conductor.

However, looking at transmission line theory, a balanced center grounded line pair or balanced three phase line group will behave differently than a single isolated line with an earth return.

Well, reality says other wise it seems. Distance between ground rods does play a difference.

 

[paulengr]

I am not sure quite how to interpret that statement. The classic ideal theoretical formulation is that the electrode resistance to remote earth is a constant once you get beyond the direct zone of influence of an earth electrode, while the earth portion of the resistance is a constant rather than changing at all with distance. It neither increases nor decreases with distance once the effective electrode resistance at each end is subtracted.
Can you explain the derivation of the result you describe? The situation for a multiply grounded system may be different since the longer distance adds more intermediate earth electrodes with metallic paths between them.

It is straight out of the IEEE Green book (grounding) among others. The Earth path between two electrodes is a 2D system. As the distance increases the number of paths increases with the square of distance and the resistance between two grounds is the sum of the conductances of those paths proportional to 1/(d*d). Simultaneously the resistance is increasing linearly along a given path proportional to the distance d. Thus it is proportional to d/(d*d) or just 1/d.

This is not just theoretical. The principle behind the fall of potential ground measurement method includes a 1/d factor. The derivation is usually given in the fall of potential manuals as well.

At long (remote Earth) distances it is only “constant” because the limiting factor becomes how well the ground electrode is coupled to Earth.

 

[paulengr]

Well I wouldn't call that effective because it won't trip a standard thermal mag breaker. But I get what you're saying.


.

The IEEE standards for substations basically force gravel over a substation's ground grid, where as the IEC takes a worker's boots and gloves into account and as such many EU substations have grass instead of gravel over their ground grid.

In simple terms the IEEE worker is significantly more protected during a fault at a substation than an IEC worker.

So, if the IEEE (for the sake of the argument) miscalculated the human body impedance, would it ever be called into question considering all North American substation have an added insulators on top of their ground grid?

Just some food for thought...




Blame UL...

Even if this is true (it might be, I'm open to all data that challenges established standards) it doesn't change the fact that under the NEC someone can legally run a circuit such that it never trips a breaker.




I'd need more details. Clothes, weight, source impedance, bonding of the environment around him, ect all play a role and may not reflect worse case.
Keep in mind that the IEC typically allows for 5 second disconnections times on circuits over 32 amps, some countries over 100amps. The idea is that with larger wire more current will flow due to lower Z, pulling down the voltage at the spades of the transformer.

As such X0-X1 might dip down to 100 volts instead of holding 277 on a 200 amp feeder fault vs a 20 amp branch circuit fault.

So instead of 139 volts to remote earth at point of contact, the person only "sees" 50 volts.

[/QUOTE]

IEEE standard 80 uses the 1,000 ohm constant because they use the research by Charles Dalzeil. The variable resistance with voltage is a refinement that was experimentally derived some time later.

 

[paulengr]

Well, reality says other wise it seems. Distance between ground rods does play a difference.

Earth isn’t a factor here. It’s the transmission line parameters (Xc and Xl). R is also increasing linearly but the inductive term especially dominates.

 

[mbrooke]

Earth isn’t a factor here. It’s the transmission line parameters (Xc and Xl). R is also increasing linearly but the inductive term especially dominates.

That can't explain a uni grounded line's behaviors.

 

[mbrooke]

IEEE standard 80 uses the 1,000 ohm constant because they use the research by Charles Dalzeil. The variable resistance with voltage is a refinement that was experimentally derived some time later.
[/QUOTE]


Alright, but why do you think the IEC is wrong in assuming a lower resistance at higher voltages?

 

[yuhong]

What was the CSA's reasoning behind adding this table? What purpose does it serve?


View attachment 2554170

There was a voltage drop requirement in the CEC, but they made an exception if this table is followed.

 

[yuhong]

See for example https://www.technicalsafetybc.ca/sites/default/files/csa-fia3660-voltagedropcalc.pdf

 

[mbrooke]

See for example https://www.technicalsafetybc.ca/sites/default/files/csa-fia3660-voltagedropcalc.pdf

This is what I was looking for, Thank you!

 

[mikeames]

Blame UL...

Even if this is true (it might be, I'm open to all data that challenges established standards) it doesn't change the fact that under the NEC someone can legally run a circuit such that it never trips a breaker.

I don't disagree but its an easy problem to solve for qualified electricians that understand. I prefer not to see more restrictions for a minority of circuits that fall into this category. We have cars that do 100mph but no roads where you can safely go that fast. Why is that permitted? Perhaps because sane individuals understand the risk. Just because you can does not mean you should. So why not have state laws that prohibit cars having a max speed above the state limit. Some speed an create problems but that law is unnecessary.

Perhaps I am being over simplistic, but even understanding the issue, I am not passionate about a code change addressing it. I put that on the shoulders of the engineers and the EC. I understand this situation may come up with a long run in a residence, but again its limited. Where does it stop. What about a 4/0 SER in a residence? If it faults will it clear a 200a main if the utility transformer can't supply 200 amps? I have seen small old pole mounted transformers that may not.

 

[mbrooke]

I don't disagree but its an easy problem to solve for qualified electricians that understand. I prefer not to see more restrictions for a minority of circuits that fall into this category. We have cars that do 100mph but no roads where you can safely go that fast. Why is that permitted? Perhaps because sane individuals understand the risk. Just because you can does not mean you should. So why not have state laws that prohibit cars having a max speed above the state limit. Some speed an create problems but that law is unnecessary.

Perhaps I am being over simplistic, but even understanding the issue, I am not passionate about a code change addressing it. I put that on the shoulders of the engineers and the EC. I understand this situation may come up with a long run in a residence, but again its limited. Where does it stop. What about a 4/0 SER in a residence? If it faults will it clear a 200a main if the utility transformer can't supply 200 amps? I have seen small old pole mounted transformers that may not.

Right, but remember that a lot of newer code mandates come from trying to cover for the minority of circuits that fall into this category.

 

[AdrianWint]

I just bought a tester that determines available fault current at the point it is connected, don’t know how accurate it is, but I was surprised at how low the available fault current was at my house. Haven’t used it yet in a commercial setting.

So isn't this essential test gear for you guys? It is for a professional electrician here in the UK. Typical values can range from high hundreds of amps to low thousands in a domestic setting.

We use it in several modes - to determine the PSCC (perspective short circuit current) ie. Will a particular breaker trip on the instantaneous part of the curve if used on this circuit. Also to determine if that breaker has a suitable breaking capacity in order to safely clear such a short circuit fault. A PSCC test can also help identify bad/loose connections.

To determine EFLI (Earth fault loop impedance) - to determine that fault current that would flow if a bolted phase to earth fault should occur. Helps determine the type of earthing (grounding) in use and whether or not additional measures (like an RCD (GFCI) must be used). An EFLI test is done at every point of utilization to determine if the earth connection is present & effective.

 

[mbrooke]

To determine EFLI (Earth fault loop impedance) - to determine that fault current that would flow if a bolted phase to earth fault should occur. Helps determine the type of earthing (grounding) in use and whether or not additional measures (like an RCD (GFCI) must be used). An EFLI test is done at every point of utilization to determine if the earth connection is present & effective.

My understanding is that an RCD is not allowed to "correct" high EFLI in TN-S and TN-C-S systems.