Service Entrance Switch and SubPanels

Hello Everyone, First Thanks for all the information I have received, I continually look to learn and find better ways to do things and this site has definatly aided in that quest. I know there are many threads on subpanels but I dont seem to find any that have the following scenario which I am going to run into so here it goes:

Looking at a house where there is a Safety Switch/Service Disconnect Connected by conduit to the meter. The service entrance cable goes from this switch directly to the main panel, In turn the main panels Netural and Ground Bus Bars are seperated. From this main panel they have a 60amp circuit that feeds Two Hots and a Neutral wire through plastic conduit to the detached garages subpanel. From the subpanel in the garage they have it directly grounded to two grounding rods directly outside. My question is since the house main panel has seperate Neutral/Ground Bus Bars, should the garage also be seperated or is it fine they have both bonded/connected to each other?

Please let me know and excuse the terminology if its incorrect, if anymore info is needed I can take pictures when I go inspect the home in detail.


A three wire feeder to a detached structure from a service was permitted under the 2002 and earlier versions of the NEC. But under later versions of the NEC you would need a 4 wire feeder to the garage.

Welcome to the Forum.

Thanks Robert, to followup, since it appears this was not done recently, and running a 4th wire from the house to the garage would be at a considerable cost as there is concrete going from house to garage and no place to run wires, would the current setup be unsafe? If so should I recommend to remove the bonding strap at the subpanel to seperate the ground and neutral bars as to mimick the main panel? Let me know…

The only real solution is that you’ll need to pull a fourth conductor in the conduit and then provide separate neutral and EGC buses in the garage panel and move the respective conductors to each bus. The GEC(s) from the two rods must terminate on the EGC bus not the neutral bus. Removing the bonding strap will accomplish nothing and is likely to make it more unsafe than it already is.

Gotcha so I can explain this properly and to refer them to an electrician, why would it matter if at the garage sub panel they are seperated neutral conductors on one side and ground on another, with the grounding rods directly off the ground bus, which mimics the house main. I would think this would be safer as the ground does not have to travel back to the house…

The problem is that once you go beyond the service disconnecting means (where the main bonding jumper is installed and in this case adjacent to the meter enclosure), the EGC’s and neutral conductors must be separate. If they are not then the neutral current will be traveling on the EGC, metal parts and the neutral back to the service disconnect which is very dangerous. Here’s a graphic from Mike Holt:

Robert thanks for the diagram, I guess I had the right thought on recommending the removal of the bond between the neutral and the ground. However sorry :slight_smile: to keep pushing this and maybe im just missing something, but what is the difference of just having the grounding rods at the subpanel, without pulling a grounding conductor from the main panel? Wouldnt that serve the same purpose? From a freshman perspective it sounds like as long as the bond is removed in the subpanel and there are grounding rods at the garage leading to the ground bus it should be fine, whats going over my head?

So where would the fault current go if you had a ground fault at the garage? Into the ground? That certainly would not generate enough current to trip an OCPD. The ground fault current must go back to the neutral (grounded conductor) where it is connected to the EGC by the main bonding jumper, in this case that’s back at the service disconnect next to the meter. Some more graphics see if they help:

The key to understanding grounding and bonding is to understand how electricity behaves. Most of what you see and hear about residential grounding and bonding has to do with providing an effective path for fault currents and eliminating objectionable currents. Phrases such as “objectionable currents” don’t mean much to most people.

The illustrations at are about as good as they get. Mike Holt explains things in ways that are relevant, practical and easy to understand. His explanations and illustrations are sufficient for most purposes. However, the key to really understanding grounding and bonding is to understand how electricity behaves. Mike’s main audience is electricians. He generally doesn’t go too deep into explaining electricity.

I won’t get too technical but it is worth knowing that electricity behaves differently with different waveforms, frequencies, Voltages and currents. Electrical systems have to be designed like cars. They need to survive a crash. Comparing the crash worthiness of electrical systems to cars, electrical systems are way ahead of cars. Imagine running a car into a solid concrete wall at 100 MPH. It proabably wouldn’t do so well. A modern electrical system can - literally - handle more rogue energy than car hitting the wall but without there being any visible evidence that there was even a crash.

In a perfect world, an electrical system would not have any ground loops and there would never be a potential (aka Voltage) between any grounded conductors (of any type). Of course, we don’t live in a perfect world so there will always be the potential for a potential (no pun intended) between gounded conductors and there will always be ground loops.

Imagine an installation in which you have two earth grounds that have a slightly different resistance, say only one or two Ohms, and a fairly close together. If there is a lightning strike (more energy than 1,000 of our cars from above hitting walls) a few hundred feet away from the two grounds, the very small difference in resistence between the grounds could be enough for there to be thousands of Volts potential across them. In other words, if you had a Voltmeter connected between the grounds - in a perfect world - the meter would never, under any circumstances, read anything other than zero. We don’t live in a perfect world so, we do not want it to be possible for anything to be in direct contact with the two grounds at the same time. We accomplish that by making them into one ground connected together at a common point.

Connecting the grounds together at a common point also solves the problem of a ground loop. Going back to my example of an signal induced by a lightning strike, in a ground loop, we could have thousands of Amperes flowing.

It is certainly true that the ground system facilitates the operation of overcurrent protective devices during system faults but it can do that
in lots of different arrangements. The rules related to the arrangement of grounding systems have as much, or more, to do with external EMF sources such as lightning.

Electricity behaves the same in a residential installation as in any other type of installation but if you really want to understand grounding and bonding, you will find more information in books that deal more with non-residential electrical systems. The National Electrical Code is not the best way to learn about electrical system design. The NEC is merely a set of minimum standards. There are hundreds of interrelated standards that go into more detail than the NEC.

The NFPA publishes the NEC but most of the standards upon which the NEC is based are developed by IEEE and NEMA. The core of the IEEE standards are collectively called the “Color Books”. The two books in the Color book Series that will answer your questions in detail are the Green Book and the Emerald Book. The Emerald Book, the proper title of which is “IEEE Recommended Practice for Powering and Grounding Electronic Equipment”, IEEE Std 1100-1999, (ISBN 0-7381-1661-0 SS94741) is, in effect, a rewriting and modernization of a standard titled “Guideline on Electrical Power for ADP Installations” and that was commonly known as FIPS-Pub-94.

If you can get your hands on a copy of the Green Book, I’d suggest reading it first. It is the Emerald Book, however, that really starts to bring things into focus. FIPS-Pub-94 is in the public domain. You can download it from the Education Center on my website. It gets a little technical but the heart of what you want to know is in chapters two and three, a total of only 30 pages or so. It is not as clearly written as a Mike Holt book (it is, after all a government publication) but it does have some illustrations and explains things in plain language. In that respect, it is actually a little better the Emerald Book.

3 wires can go to the Garage if it is a separate building. At the garage you would then bond the neutrals and grounds together with the panel and install an approved grounding electrode system-which the 2 ground rods supply. Any water lines in the garage would also be bonded.

An alternate would be to have four wires one being grounding and then treat the garage panel as if it where just another sub.

Hope I understood your question and was able to help.

Yes, that’s correct if your installation is based on the 2002 or earlier versions of the NEC. One comment, if that water line that you mentioned originates in the separate structure where the feeder originates then you cannot use a 3-wire feeder.

Steve -

If you think something is wrong REFER it to a competent and licensed electrician. Not to be rude BUT if you’re NOT you got no business trying to tell a buyer, seller or electrician what to do on something like this.

You’re a trial attorneys wet dream come true if something went wrong.