grounded and equipment ground together at sub-panel

[electro7]

Hi,
I have ran into a few older homes that have sub-panels that are fed without an equipment grounding conductor. So the grounded/neutral conductor and equipment grounds from the branch circuits are landed together. I have always learned that the nuetral/grounded conductor and equipment ground should only be bonded at the main service disconnect. I was wondering if anybody could explain what would happen if there was a ground fault in one of the branch circuits? Would the current travel back to the main panel through the neutral, thus violating code that the nuetral should never carry fault current? So, I am wondering if I should run an equipment ground back from the sub-panel to the main panel??

I have also run into systems that are similar to this except there is a grounding electrode at the sub-panel which all the equipment grounds are tied to. Is this okay?

Thanks for your help.

 

[hurk27]

Hi,
I have ran into a few older homes that have sub-panels that are fed without an equipment grounding conductor. So the grounded/neutral conductor and equipment grounds from the branch circuits are landed together. I have always learned that the neutral/grounded conductor and equipment ground should only be bonded at the main service disconnect. I was wondering if anybody could explain what would happen if there was a ground fault in one of the branch circuits? Would the current travel back to the main panel through the neutral, thus violating code that the neutral should never carry fault current? So, I am wondering if I should run an equipment ground back from the sub-panel to the main panel??

You have the reason a little backwards, one of the main reasons for not using the grounded conductor as a equipment ground after the main service disconnect is any voltage drop or loss of this neutral conductor will be applied to all grounding that is using this neutral as a return point, also a fault between a hot and neutral on the load side of this neutral will also produce a high voltage drop on this neutral also imposed on the EGC's using it as a return path so the voltage to Earth or to another appliance bonded via an EGC to the main panel will have this high potential between the two appliances.

I have also run into systems that are similar to this except there is a grounding electrode at the sub-panel which all the equipment grounds are tied to. Is this okay? Thanks for your help.

Prior to the 2008 250.32(B) allowed us to run a 3-wire circuit to an out building and as long as we bonded the grounding to the neutral and installed a GE it was ok if no other metallic paths between the two building existed, this included TV, phone, and water that might be bonded to the main service that would give you the same difference of potential in a fault as stated above, but they removed this allowance since many home owners will run water, phone or cable after the fact and create the same dangerous problem the code was trying to avoid, but there is a way it can still be done and code compliant, by using 230.40 exception3 where you would run from a double lug meter base back into the ground and hit the out building as if it was a service.

 

[electrofelon]

You are correct that in the situation you describe, a fault would travel back to the source through the neutral (grounded conductor). That is not the problem though. In fact the neutral is usually sized larger than the equipment grounding conductor so would clear a fault faster. The short answer for what is wrong with this type of installation is that the NEC is geared toward having a separate conductor for equipment bonding/fault clearing, and they require this basically as soon as they have the authority too - after the main disconnect. They would like to have them be separate at the service as well, however utilities do not run four wire services so neutrals and grounds have to be bonded at the service by the main bonding jumper to provide a path for fault current to get back to the source. The main reason grounded and grounding conductors ("neutrals and grounds") are separated after the main disconnect is because if the grounded conductor in, say, a "sub-panel" feeder were to be compromised or broken, then every metallic piece of equipment connected to the sub would rise to line voltage. Do you understand why? Let us know if that doesn't make sense. Hint: at the extreme low current levels that will shock, injure, or kill a human, there will be essentially no voltage drop through a resistor. Another reason to separate grounded and grounding conductors is that without doing so, you are more likely to get neutral current flowing through metallic parts that are not supposed to carry current which can result in a shock hazard. To give an example, say you have a fridge and a microwave both with the metallic frames grounded by the neutral. Also lets say these appliances are adjacent to each other. Also lets say the microwave is on a 14 gauge circuit that is a long run from the panel, and the fridge is on a 12 gauge circuit that is not a long home run. During use, the microwave's case will be a fair amount above ground reference because there is a lot of voltage drop due to the high current and high resistance of the long grounded conductor wire run. The fridge only draws a few amps and has a relatively low impedance grounded conductor path so its case is much closer to ground reference. Because of this we have a difference of potential between the two and one could possibly get a shock by touching both in extreme circumstances. Further, if they were touching, current would travel between them causing arcing and a fire hazard. All of this could be much worse during a fault or surge situation where the potential difference could be many times higher. During a fault you could have hundreds of amps flow, so now to an ohms law calculation with a few hundred amps and you will get significant potential differences in very short distances on the same appliance.

Dirt is not a suitable path back to the source for systems under 600 volts. A hot wire connected directly to a ground rod will only pass a few amps - clearly not enough to trigger an OCPD. This would certainly be a dangerous situation, thus there needs to be a low impedance fault path which would be a conductor sized by 250.122 or an approved raceway. Note that a "sub-panel" in a remote building requires a ground rod (or two) but fault clearing is not its purpose at all.

Hope that helps

 

[electro7]

Yeah, thanks a lot for the feedback. I have a lot more to learn here, however I think I am getting most of what you are saying.

So I see the reason of, in case the neutral is lost from the main panel to the sub-panel then any fault current has no where to go except on the metal enclosures, etc., and would not open the OCPD, making the conducting materials hot (a different potential than ground), and would cause an electrical shock, right? So, am I understanding right that we are to always run an equipment ground to sub-panels (and everything for that matter) when they are fed from the load side of the main disconnect? How about existing residential systems in which this isn't the case? Is there provision in the code for that?

Another question I work for a solar company and just recently installed a ground mount solar array about 370ft away from the house. We were unable to trench because of rock beds under the earth so we installed poles and ran overhead wire to back feed the house panel. We put the inverter (grid-tie, not stand-alone or hybrid) at the array and ran triplex wire back to the house. I drove a ground rod at the array and bonded everything to that, and used the bare conductor in the triplex wire as the neutral. There is no equipment ground running from the array and inverter back to the house. I was wondering if this is okay? Is there danger in this, not having an equipment ground from the array and inverter to the house? I was wondering if there was a fault at the array if it would clear the breaker at the house panel?

Thanks.

 

[electrofelon]

Electro7:

So I see the reason of, in case the neutral is lost from the main panel to the sub-panel then any fault current has no where to go except on the metal enclosures, etc., and would not open the OCPD, making the conducting materials hot (a different potential than ground), and would cause an electrical shock, right?

Not just fault current, but anytime there is a load turned on, the equipment will rise to full potential. To see it, Imagine a three wire range (but of course it could be a "sub panel" too) which means the neutral is connected to the frame. Say it has a 120v light bulb between one of the hots and the neutral and say the switch controlling that bulb is on. Now break the supply neutral/ground. Now what you basically have is a hot supply wire connected to one end of the light bulb filament and the other end of the filament connected to the frame. So really you have a hot wire connected to the frame - yes there is a light bulb filament between the hot wire and the frame, but remember there is a voltage drop across a resistor only when current flows. If a person get in series with the frame and a path back to the source some current will flow, but a very very small amount and not enough to drop the voltage any significant amount. I think at first some people find this hard to understand because they think that after the load, whether its a light bulb filament or whatever, the conductor automatically turns into the neutral conductor, but without that neutral conductor connected back to the source, you wont get that current to flow and drop that voltage right down to almost 0 (it would be 0 on the neutral side of the filament if the neutral conductor was a perfect conductor).

So, am I understanding right that we are to always run an equipment ground to sub-panels (and everything for that matter) when they are fed from the load side of the main disconnect?

yes, except separately derived systems.

How about existing residential systems in which this isn't the case? Is there provision in the code for that?

there is a provision for existing three wire range circuits - that they can be reused. For other situations, the code doesnt say that existing things that are incorrect have to be fixed. Perhaps an AHJ can make such rules for glaring violations, but in general no one can make you fix an existing violation.

Regarding you solar question, I am not sure of that answer, perhaps someone else will chime in. I would depend on whether the inverter is considered a separately derived system and if it is internally bonded.

 

[electrofelon]

Ok after some further thought, I realize I was thinking about it wrong. I was starting at the wrong end of the system. You imply that your inverter needed a neutral connection, is that correct? If so than I believe that the feeder to it would need to be four wire - separate neutral and ground. The Dc side is most likely a SDS depending on your inverter type. There are others who would be better on this topic than me, I mostly do off grid stuff.

 

[electro7]

Well I can configure the inverter to 240v without a neutral, but I didnt in this case. I set it at 240v with a neutral. I was thinking about it after the fact and thought maybe I needed to go back to change it. My concern is if there is a ground fault, will it trip the breaker? Or will the frame and metallic parts of the array become hot? Actually, I think it will blow a fuse in the inverter before it would trip any breaker.

Thanks a lot for your guys help. I appreciate it!

 

[kwired]

Separation of EGC and Neutral beyond the service is not done as much for the open neutral condition as it is done because of voltage drop that will develop on the neutral, as well as if the grounded conductor is bonded in multiple places to otherwise non current carrying metal components then all those components effectively become current carrying conductors themselves. The open neutral conditions are just as much a problem if the service neutral conductor becomes open as they are if a feeder neutral were to become open. If there is current there is voltage drop, if there is voltage drop on exposed metallic components, there are potential touch points that will be shock hazards.

 

[haskindm]

The problem is not the possibility of having "fault current" on the grounded conductor, this is expected in the case of a line to neutral fault. The problem is the certainty that the equipment grounding conductors will be carrying neutral current during normal operation.

 

[hurk27]

The problem is not the possibility of having "fault current" on the grounded conductor, this is expected in the case of a line to neutral fault. The problem is the certainty that the equipment grounding conductors will be carrying neutral current during normal operation.

How will the equipment grounding conductors be caring neutral current? the EGC's will only be from the circuits in the sub panel and these EGC's should not have any current on them unless a fault happens, since they are bonded to the feeder neutral at the subpanel the only path for neutral current will be via the neutral, how ever it is true that the few appliances that might also be connected to a water line or gas line will share the path with the neutral so in these cases there will be some current on the EGC's to these few circuits but it would be very little.

Kwire has it correct, it is the problem of the voltage drop in the neutral, with the EGC's also bonded to this neutral at the sub panel voltage drop in this neutral will cause a rise of voltage potential difference between anything grounded to it, and Earth or other grounded metallic paths that are at the service neutral potential.

 

[mlnk]

Electricity does not follow the path of least resistance-it follows all paths that have anything other than infinite resistance. All we need to remember is to do our job so that the least amount of current flows on the grounding wire under normal circumstances, but there is a low resistance path under ground fault or short circuit conditions.