Separately derived system connection to GES

[jaggedben]

...
I believe that was because it was being used as both, though a solar install is hardly separately derived but rather parallel source. Similar to a generator being wired as non separately derived. Though I am pretty sure you still need a electrode to bond to the case for voltage stability and gradients.
...

In some cases (early Enphase micros) we had to run a real GEC from the roof because we were actually grounding DC conductors there, and in other cases it was just AHJ stupidity. But also, earlier string inverters were actually separately derived systems with transformers in them, although code-wise the grounding rules in 690 pretty much overrode 250. Those inverters came with marked terminals for a GEC. And, due to capacitive effects in solar modules, the GECs for those inverters actually did carry some current under normal operation. I was a bit shocked - well, not literally - when someone told me that and I measured it the first time.

 

[wwhitney]

You don't want to be making series connections to earth on the SDS by relying on an EGC to bond the secondary side to the GES.

Why not?

If a secondary conductor faults to the actual earth, the effect of the higher impedance connection to the GES will be indistinguishable from the GES itself simply having a higher impedance to earth, and the GES impedance to earth is uncontrolled. And if a secondary conductor faults to anything else, the presence of a low impedance fault path does not depend on the presence of the secondary GEC.

Cheers, Wayne

 

[electrofelon]

SBJ bonds the case to the grounded conductor to create a equipotential and an inherent path back to the source. It doesn't mean the secondary side's grounded conductor is earthed by the EGC.

Of course the EGC will be connected to Earth.

You can have a gradient on the EGC from the GES to the SDS due the the wire's impedance that can result in a sustained fault and potentially risk shock, fire, or destruction of material used in the electrical installation.

So why is a GEC not subject to this impedance? And where is all this fault current flowing on the GEC coming from? The only time I see any fault current flowing on a grounding electrode conductor is from contact with medium voltage lines. That is extremely rare, and besides we were originally talking about a separately derived system which is unlikely to have contact with higher voltage lines.

The impedance on an EGC sized to the OCPD on the primary side would not be enough for the fault currents and most likely would even allow for a current divider situation that limits the fault return. Some current through the the earth, some through the piping, some returning to the utility transformer and potentially a voltage gradient between the SDS "grounded conductor" and the service's GES since some of the paths will have an impedance. Thus imbalance and poor earthing could create a difference in potential between the system's GES at the service and the neutral point at the separately derived system. Not to mention, the restriking insulation failure issue and voltage rise in ungrounded systems. Voltage stability on the secondary is important and one of the reasons we ground services.

Again there is not fault current flowing on the gec, that is not its purpose. Also I was not talking about ungrounded systems.

The SBJ would be the most important but the GES being connected would be second.

About the only thing a GES does is help clear a MV fault, like if a MV line falls on a service drop. it does reference the system grounded conductor to Earth, but I guarantee you that's going to happen anyway even without a grounding electrode system when you consider water lines, wells, things fastened to masonry, etc. (assuming all bonding is done correctly).

 

[Elect117]

the GES impedance to earth is uncontrolled. And if a secondary conductor faults to anything else, the presence of a low impedance fault path does not depend on the presence of the secondary GEC.

I don't know if I agree with that. While it shouldn't be used as the fault path, it can be part of the path. And as the resistance grows, so too would the voltage seen across that resistance. Since the fault current would be "constant". Though with a proper EGC that current would be minimal or none.

From the Green Book, IEEE 142, chapter 4.1.3, "The development of a low-resistance ground electrode is of paramount importance to satisfy the requirements for system feeders from utilities, where the fault return path is via the ground. Also, secondarily, as a backup to the equipment grounds, which may occasionally be open-circuited unintentionally. Logically, the lower the resistance of the grounding system, the more adequately these requirements are met."

Their point is that the GES/GEC being of low resistance does serve a purpose. Their noted reasons do not include being the desired ground fault return path, nonetheless, it is one. Their main reasons were as Electron Sam said.

There are other mentions in chapter 1 about the required resistances needed to ensure it is effectively grounded but I don't think those would apply. The resistance of the EGC would probably still satisfy those. Unless a really small primary breaker was used for a large transformer but then you wouldn't be able to use it due to inrush. So, that would kinda defeat that premise.

So why is a GEC not subject to this impedance? And where is all this fault current flowing on the GEC coming from? The only time I see any fault current flowing on a grounding electrode conductor is from contact with medium voltage lines. That is extremely rare, and besides we were originally talking about a separately derived system which is unlikely to have contact with higher voltage lines.

It is.

But the NEC does not cover the calculations of ground resistance. They only specify that an additional electrode can be required if more than 25 ohms. Frankly, nobody does those calculations for circuits of less than 1000V any more. I think they used to in Canada. It would be abnormal to have fault current through earth but it can / does happen. It would not be nearly as much as a utility but it is still relevant since the NEC covers more than just dry types for an office. There are numerous industrial locations that would have their own MV separately derived systems that utilize a ground grid or similar ground studies just for this purpose.

The Soares Grounding and Bonding book goes into more detail about the grounding electrode conductor ampacity, long runs of emt used as a EGC, etc. Things that aren't addressed by the NEC but could be relevant.

Again there is not fault current flowing on the gec, that is not its purpose. Also I was not talking about ungrounded systems.

While it is not supposed to, there can be current flow through connected electrodes or bonded metal piping if they are a parallel path back. Similar to how a EGC and the metal conduit can act as parallel paths back.

 

[electrofelon]

I don't know if I agree with that. While it shouldn't be used as the fault path, it can be part of the path. And as the resistance grows, so too would the voltage seen across that resistance. Since the fault current would be "constant". Though with a proper EGC that current would be minimal or none.

From the Green Book, IEEE 142, chapter 4.1.3, "The development of a low-resistance ground electrode is of paramount importance to satisfy the requirements for system feeders from utilities, where the fault return path is via the ground. Also, secondarily, as a backup to the equipment grounds, which may occasionally be open-circuited unintentionally. Logically, the lower the resistance of the grounding system, the more adequately these requirements are met."

Their point is that the GES/GEC being of low resistance does serve a purpose. Their noted reasons do not include being the desired ground fault return path, nonetheless, it is one. Their main reasons were as Electron Sam said.

There are other mentions in chapter 1 about the required resistances needed to ensure it is effectively grounded but I don't think those would apply. The resistance of the EGC would probably still satisfy those. Unless a really small primary breaker was used for a large transformer but then you wouldn't be able to use it due to inrush. So, that would kinda defeat that premise.



It is.

But the NEC does not cover the calculations of ground resistance. They only specify that an additional electrode can be required if more than 25 ohms. Frankly, nobody does those calculations for circuits of less than 1000V any more. I think they used to in Canada. It would be abnormal to have fault current through earth but it can / does happen. It would not be nearly as much as a utility but it is still relevant since the NEC covers more than just dry types for an office. There are numerous industrial locations that would have their own MV separately derived systems that utilize a ground grid or similar ground studies just for this purpose.

The Soares Grounding and Bonding book goes into more detail about the grounding electrode conductor ampacity, long runs of emt used as a EGC, etc. Things that aren't addressed by the NEC but could be relevant.



While it is not supposed to, there can be current flow through connected electrodes or bonded metal piping if they are a parallel path back. Similar to how a EGC and the metal conduit can act as parallel paths back.

I think the problem is you are not differentiating between utility transmission and distribution systems and NEC covered systems. Yes the NEC can cover medium voltage, but it is a very very small percentage of NEC covered systems, and you still cannot use Earth as a fault path. For less than a thousand volts, grounding electrode systems have nothing at all to do with fault clearing.