Grounding a big existing building

[lile001]

I am always in a quandry when rewiring an existing building as to how to ground it. I believe that wimpy ground rods are fine for a small residence, but inadequate for a commercial building of any size - I'd rather see a Ufer ground than anything else, failing that a nice ground ring at the footing. We have two projects right now where we are doing major remodels on existing buildings.

Building A is a historic theatre built in the 1920's, when "grounding" meant your Model A was high centered on the road ruts. We are putting a new 1200A 208V service in. I'd love to clamp onto some rebar for a Ufer ground, but there isn't much new concrete available. It's all brick, there is almost no building steel. There is a little bit of steam piping, and some new water pipe no doubt connected to a plastic pipe out in the street. The building covers a city block, and I'd love to put in a ground ring around the whole thing, except we'd have to dig up a quarter mile of sidewalks to do it. I hear horror stories about big buildings grounded with a skinny ground rod at one corner, where the other corner is 50V to "ground" due to stray current.

Building B is worse - it is a physical plant for a hospital complex. There are a total of 6 service entrances. I go back through plans to the '60's, and find that grounding is not addressed. I do find ground buses in some of the switchgear, but nothing on old plans indicating a ground ring, Ufer ground, ground rods, and so on. I doubt there is anything there at all. I also doubt that the various service entrances are tied together - unless some conduit happens to go from frame to frame, the switchgear grounds are probably floating. So far tracing down ground conductors has resulted in no conclusions.

I run into this constantly - we are asked to put a new service into an existing building, where the existing ground system is either nonexistent or inadequate. What approaches have people used in the past? Do you really dig up all the sidewalks for a ground ring? Do you start jackhammering concrete until you can expose some rebar? Do you drive ground rods at four corners of the building and hope for the best?

 

[bphgravity]

I would drive a single rod outside at the nearest point of the service and extend several bare conductors (#6) off the rod in a radial configuration about 10-15 feet in length each, buried about a foot into the earth.

The resulting configuration should resemble a dendritic or tree-like pattern. This would likely provide the best grounding performance.

 

[don_resqcapt19]

lile001,

I believe that wimpy ground rods are fine for a small residence, but inadequate for a commercial building of any size

Why?, What function do you think that the grounding electrode system performs?

I hear horror stories about big buildings grounded with a skinny ground rod at one corner, where the other corner is 50V to "ground" due to stray current.

Additional connections to earth won't change that.

Don

 

[lile001]

What function does the grounding electrode perform?

What function does the grounding electrode perform?

lile001,

I believe that wimpy ground rods are fine for a small residence, but inadequate for a commercial building of any size

Why?, What function do you think that the grounding electrode system performs?

Don

Now this brings us to a basic question. There are many who have stated on this forum "the grounding electrode system serves no purpose/is a waste of copper/etc etc" So far I haven't heard why, just the assertion. (I've been reading these sorts of posts all afternoon)

The purpose of connecting a system to ground is to make sure that the entire building is at the same potential during abnormal conditions. The reasons it may not be at the same potential are: stray voltage, lightning, recloser surges, or ground faults that don't trip a breaker and don't energize any equipment ground conductors. These are all abnormal circumstances, under normal circumstances you are right the grounding system doesn't do anything. The emergency brakes on a car don't do anything either, under normal circumstances.

In a large building, say one covering a city block or more, there might easily be 50V between the earth at one corner and the earth at the other corner. If there is a lightning strike at one corner, there might be 10000 volts between the corners for a short time. A hapless occupant standing barefoot on the concrete holding onto a grounded pipe that was only earthed at the other corner might recieve a shock if the concrete and rebar is not bonded. The ground system, IMHO, should connect all points of the building to prevent these voltages from affecting the occupants. In most new buildings, if the rebar is bonded, this is automatically accomplished. If a ground ring is run around the building, this also accomplishes this. If there is a single ground rod at one corner of the building, and no other means of earthing, then this is not accomplished (and it violates the code too)

I would argue that a larger service entrance would justify a more complex ground system meant to achieve a lower ground resistance. A larger system is subject to larger faults, more chance of getting hit by lightning, and a larger footprint is more likely to be affected by stray voltage. Tieing these earth voltages together is the point.

One horror story goes like this: a large footprint building has a proper NEC ground system at the service entrance, but a telephone system improperly connected to a ground rod at the opposite corner of the building, and not bonded to the electrical system. Operators are getting 50V shocks when they put on their headphones, if they happened to have conductive shoes or touch thier metal desks. The earth actually had 50VAC between ground rods at the two corners of the building.

Is this a bonding error? Yes. But it is also a ground system error. If the building rebar was properly bonded, there would not be 50V potential between opposite corners of the building because all points would have been tied together.

 

[don_resqcapt19]

lile001,

The purpose of connecting a system to ground is to make sure that the entire building is at the same potential during abnormal conditions.

That is the purpose of bonding and not of grounding. You cannot keep everything at the same potential by grounding.

The ground system, IMHO, should connect all points of the building to prevent these voltages from affecting the occupants. In most new buildings, if the rebar is bonded, this is automatically accomplished. If a ground ring is run around the building, this also accomplishes this. If there is a single ground rod at one corner of the building, and no other means of earthing, then this is not accomplished (and it violates the code too)

The bonding of all of the conductive parts of the building to the electrical system or to other conductive items do this. Additional connections to earth without bonding cannot do this and would be a violation. Also a single rod would not be a violation, no matter how large the building is, assuming that none of the other grounding electrodes listed in 250.52 exist and that the single ground rod has a resistance of 25 ohms or less.

I would argue that a larger service entrance would justify a more complex ground system meant to achieve a lower ground resistance. A larger system is subject to larger faults, more chance of getting hit by lightning, and a larger footprint is more likely to be affected by stray voltage. Tieing these earth voltages together is the point.

The grounding electrode system so play no real part in fault clearing...that is the job of the equipment grounding (bonding) conductor and the main bonding jumper. Additional connections to earth do nothing to get rid of stray voltage.

One horror story goes like this: a large footprint building has a proper NEC ground system at the service entrance, but a telephone system improperly connected to a ground rod at the opposite corner of the building, and not bonded to the electrical system. Operators are getting 50V shocks when they put on their headphones, if they happened to have conductive shoes or touch thier metal desks. The earth actually had 50VAC between ground rods at the two corners of the building.
Is this a bonding error? Yes. But it is also a ground system error. If the building rebar was properly bonded, there would not be 50V potential between opposite corners of the building because all points would have been tied together.

That really makes my point, bonding keeps everything at the same potential...earthing doesn't.
Don

 

[petersonra]

Re: Grounding a big existing building

What difference does the size of the building make?

The ground connection is made because the code requires it. The minimum code requirements are adequate because IMO the primary reason we ground electrical services is because the code requires it.

Bonding is another issue entirely.

 

[lile001]

More fun with grounding

More fun with grounding

lile001,
You cannot keep everything at the same potential by grounding.

Bonding the rebar of a foundation does accomplish creating a ground plane under a building, and I would argue that the entire concrete structure would tend to be at nearly the same potential, unless very large currents were flowing in it. The CEE qualifies as a grounding electrode under 250.52.

Additional connections to earth without bonding cannot do this and would be a violation.

250.52 requires multiple connections to grounding electrodes, so I guess I can't really agree with this statement. Of course all the connections to earth must be bonded together, and there should be a single point connection to the electrical system.

Are the elaborate ground mats used at substations unnecessary? I don't suppose so. There is a lot of energy available at a substation, large switching surges, and substations wear T-Shirts that say "Zeus, strike here" as far as lightning is concerned. As a building gets larger, it's electrical system tends to look more like a substation.

OK lets define what I mean by "large". Lets say we have a building with six separate 5000KVA service entrances at 4160V (I happen to be working on one now). Perhaps a football stadium, or a 22 story high rise. Maybe a chiller complex for a University, with a dozen 2000 ton chillers. How elaborate a grounding electrode system do people generally use in such circumstances? I can't imagine a single ground rod with a #6 wire would really be adequate in these situations. Of course new construction always has rebar available, which would accomplish my goals stated earlier. But what about existing construction?

Let's say you are working on an existing 22 story high rise, or a basketball stadium, and you know that the original service installed in the 1930's was not properly grounded, or the connections have all corroded away. Would a small grounding electrode at one corner of such a building be adequate?

 

[don_resqcapt19]

lile001

don_resqcapt19 wrote:
lile001, You cannot keep everything at the same potential by grounding.

Bonding the rebar of a foundation does accomplish creating a ground plane under a building, and I would argue that the entire concrete structure would tend to be at nearly the same potential, unless very large currents were flowing in it. The CEE qualifies as a grounding electrode under 250.52.

It is not the connection to ground that is keeping everything at the same potential, it is the bonding of the conductive metal parts. This would work even if the concrete slab and rebar was insulated from the earth.

don_resqcapt19 wrote:
Additional connections to earth without bonding cannot do this and would be a violation.


250.52 requires multiple connections to grounding electrodes, so I guess I can't really agree with this statement. Of course all the connections to earth must be bonded together, and there should be a single point connection to the electrical system.

That is what I said, isolated grounding electrodes are not permitted.

Are the elaborate ground mats used at substations unnecessary? I don't suppose so. There is a lot of energy available at a substation, large switching surges, and substations wear T-Shirts that say "Zeus, strike here" as far as lightning is concerned.

Yes, they use elaborate equipotential grids at the substations to keep all of the bonded items at the same potential, however this potential will be many thousands of volts above earth under fault conditions. The installation of the grid is made in a manner that limits the flow of current to earth via the grid. If there is current flow in the grid you lose the safety of the equal potential. The main purpose is to keep every thing in the substation as low step or touch potential. The person working there is like the bird on a wire as long as everything is at the same voltage.

I can't imagine a single ground rod with a #6 wire would really be adequate in these situations. Of course new construction always has rebar available, which would accomplish my goals stated earlier. But what about existing construction?

As long as everything is bonded together as required by the code I don't really see a problem.
Don

 

[Pierre C Belarge]

lile001
"or a 22 story high rise"

There are hundreds of buildings in NYC that are grounded to the water pipe and that is it, no other connections made. Many of these are more than 22 stories tall.

 

[lile001]

lile001
"or a 22 story high rise"

There are hundreds of buildings in NYC that are grounded to the water pipe and that is it, no other connections made. Many of these are more than 22 stories tall.

I hope they are not visited by any inspectors! Or are on a street with plastic water pipe!

I also know of hundreds of buildings with asbestos, ungrounded separately derived systems, lead water pipes and undersized wires, that also don't meet the code, either.

I am trying to determine what people use as standard practice in these situations, and so far, many seem to think nothing more than a water pipe connection and a ground rod with a #6 wire to it is needed for a football stadium. Somehow that doesn't add up for me.

 

[bphgravity]

I don't understand what your concern is? What does it matter if the service is supplying a football stadium or hotdog stand? The purpose for grounding a service is rather insignificant and provides a very minor function of the entire electrical system. The minimum grounding system that is permitted by the code is all that is needed for just about any electrical system regardless of what type of structure it serves.

If there are concerns such as lightning or excessive transient voltages, other more substantial protective systems should be considered such as TVSS and lightning protection systems.

 

[George Stolz]

Re: More fun with grounding

Re: More fun with grounding

Bonding the rebar of a foundation does accomplish creating a ground plane under a building, and I would argue that the entire concrete structure would tend to be at nearly the same potential, unless very large currents were flowing in it. The CEE qualifies as a grounding electrode under 250.52.

Do not take offense from what I'm about to write, Lawrence, this isn't intended as a jab to you, but simply a related story (almost a parable).

This has traces of a recent argument I had with a person who believed (deep down, without outright admitting it) that he could out-bond an open neutral. As in, he thought he could control with utmost precision all alternate grounding paths, so that a bonded pipe could not shock a person standing on the second floor of a house.

I argued that alternate grounding paths are fairly difficult to predict with certainty, despite our efforts to do so (if we were so inclined).

The discussion culminated with the fellow stating that with a Ufer, there can be no path to ground inside the structure that doesn't pass through the Ufer. Thinking that he was merely referring to a NEC-defined CEE (as any sane person would, in discussing a house), I questioned how an electrode, in the footer under the basement, asserted control over the potential of the bedroom tile on the second floor?

He then said, "No, I mean a real Ufer..." expanding our hypothetical house's CEE to include every stitch of rebar in the footer, basement wall and basement floor. He then said that no path could exist that wasn't headed off by the Ufer.

My rebuttal: Our hypothetical house just received a metal window frame in the second-floor bathroom, a rock facade with metal screen mesh behind it, with an unbonded gas pipe (with no likelyhood of energization) penetrating it, into the ground. I now had a path that didn't care about the CEE's potential one bit.

As it turns out, the fellow was a brick wall, so we eventually dropped it.

My point? After all that, you need a point?

Well, you might be able to see it. Change the problem from an open neutral to a lightning strike, and imagine your person touching the window-frame while sitting in the tub, and they will feel a shock.

The grounding electrode was great; it connected the service to the earth in a very efficient manner. But that didn't save the person touching one unbonded piece of metal. Bonding is the magic; the connection to earth doesn't do much for anything.

The Ufer could not alter the potential of the gas pipe that wasn't touching it.

Picture the electrodes as energized, and perhaps your mental picture will change as you try to envision the electrode trying to raise the potential of things in contact with it. It's power is limited.

My 2?

 

[lile001]

Grounding big buildings.

Grounding big buildings.

OK, point taken, we all agree that bonding is good.

So far I don't feel the original question has been addressed, how elaborate an earthing system do we need as buildings get larger in scope? Most people have simply dismissed the idea, and their opinion is that earth connecitons are not really needed, and a ground rod and a water pipe are good enough.

IEEE Green book maintains that the earthing connection is needed for several reasons. One of them is clearing ground faults, or faults that dont connect the phase wires to a bonded conductor, like conduit, but actually connect the phase to earth. Think runaway forklift (this happens). There are also various conditions that can produce overvoltages on a system. If the connection to earth is not stiff enough, these overvoltages can cause plenty of trouble.

I agree, bonding is great stuff, and will be that major safety system clearing the majority of faults. Lightning, surges through the power system, and faults that actually go to earth will generally not be addressed by bonding however.

I finally found my moldy copy of the IEEE Green book (Been looking for that thing since I moved) which has the following to say about the subject:

4.1.2 "Smaller installations with lower available levels of ground fault current do not require as low a value of grounding resistance as do larger systems with thier higher levels of ground fault current. System ground resistances of 1 ohm ... may only be required for large substations or generating stations. Resistances in the 2-5 ohm range are generally found suitable for industrial plant substations and ... large commercial installations."

4.1.5 " One factor that should not be overlooked in designing a grounding system is the current loading capacity of a connection to earth. .... Since approximately 25% of the grounding resistance of each rod .. occurs within a 0.1 ft radius of the rod surface, serious heating and vaporization of the moisture adjacent to the rods may occur on heavy faults. When the moisture is boiled away, the effectiveness of the rod in the dried out earth is reduced."

"The 25 ohm value noted in [NEC] applies to the maximum ground resistance for a single electrode. .... There is no implication that 25 ohms per se is a satisfactory level for a grounding system"

Paraphrasing 4.1, parallel ground rods do not add like parallel resistors - two 25 ohm ground rods would not make a 12.5 ohm total ground resistance - the total would be somewhat higher.

AIA Masterspec has this to say: In Section 16060 3.6 D they recommend:

500KVA or less - 10 ohms ground resistance
500 - 1000 KVA 5 ohms
over 1000 KVA 3 Ohms
Power distribution units serving electronic equipment: 1 to 3 ohms

IEEE Green Book lists resistances of single rods in Table 5:

Soil : ---- Avg Resistance of rod
Ashes, cinders, brine waste ---- 8 ohms
Clay, Shale, Gumbo Loam ---- 13
Same, With added sand and gravel --- 52
Gravel, Sand, Stones ---- 311

If you are in a brine swamp, a single ground rod will achieve the recommended minimum 10 ohms, however the rod probably won't carry much fault current without overheating. Other than that, a single ground rod does not meet commonly acceptable earthing requirements and needs to be backed up by more connections to earth. If you are in the Laurentian Shield, with granite instead of soil, you probably need a chemical ground.

IEEE and Masterspec don't put these numbers out just to fill up pages. They have researtched the physics behind these statements.

 

[petersonra]

I think the point is that the ground rod in a typical installation is not intended to carry any fault current at all.

I can't think of many cases (short of a large substation) where the ground rod is part of the clearing path. If it is, than you have a totally different story.

 

[dereckbc]

Lawrence, I am getting into this thread a little late, but might be able to shine some light on a few things. Over the years I have worked for a few phone companies as what they call a Power Protection Engineer. A big fancy title for an electrical engineer who specializes in protective grounding, emergency power, DC power, lightning protection, TVSS, and AC distribution. My point is I have designed a lot of low impedance grounding systems, and have written a lot of standards for telephone companies and a few for IEEE.

Unless you have a purpose like secured communications, explosives contained, medium/high voltage, or something like that, a low impedance ground system is not of much use. Let me point out a couple of examples if you have time to follow along.

You mentioned ground faults so lets start off their. Let?s assume you installed a ring, or grid achieving 5-ohm impedance. Well 5-ohm impedance is not low enough to clear even the smallest breaker of 20-amps on a 120 volt system in any reasonable amount of time. Keep in mind the NEC forbids earth to be any part of a fault clearing path for good reason, it is not possible on low voltage systems. Earth can only be used as a fault clearing path on medium and high voltage systems which is covered in NESC.

Now let?s take a look at lightning with that same 5-ohm impedance ground electrode system (GES). So we really try to be safe and design the electrical system to be very close to the point of connection of the GES. So in this scenario we have a 10-foot 750 MCM ground electrode conductor (GEC), and we think life is great. Well here is the problem with that thinking. Let?s do a simple impedance sanity check. The problem is simple, we have a 5-ohm GES and a unknown impedance of 10-foot of 750 MCM of
GEC connected to it to form a simple series circuit, right? So we try to solve the impedance of the GEC and go to resources like IEEE, ANSI and others and find the impedance of 10-foot of 750 MCM is anywhere from 350 to 2.5 K ohms at lightning frequencies. We now have a simple addition problem of 5 + 350 = 355 ohms. So what did the 5-ohm GES buy us? IMO nothing, just fattened the copper industry and contractors pockets. Now the kicker here is the 5-ohm GES is not 5-ohms to lightning, that is a measurement at power frequencies, the HF impedance is many multiples higher.

Now here is my favorite scenario, a high rise building, one in which I face all the time with SDS installed on the floors. Let?s assume we are installing equipment on the 10th floor of a 20 story building and we have a 5-ohm GES in the basement. We install a riser ground cable to facilitate connection of SDS systems like transformers and DC battery plants. We want everything to be single point grounded. We bond all our Xo?s to the riser cable and build out our system. Then one day lightning strikes the building, some one is hurt, lots of equipment is destroyed, and find ourselves asking why. After some investigation we find out that when the building was struck, the building steel provided the discharge path for the lightning like it was designed to do. During the discharge as current flowed, the building steel acted as a voltage divider along its length like any conductor. At the top of the building where lightning struck the potential voltage to earth was 100 KV, down half way on the 10th floor where our equipment was installed the building steel was at roughly 50 KV with respect to earth. However since our equipment was bonded to earth via the riser cable, and all our equipment was at 0 volt earth potential. This left us with 50KV on the building steel, and 0 V on all our electrical equipment. The result is we get flash-over from equipment to concrete/building steel, and anyone unfortunate enough to be in the area or touching the electrical equipment was injured.

So what is the solution to the above scenario? Simple we have two possible solutions. We can either bond the riser cable to the building steel on the floor it is used on (reality we bond it to every floor it passes by), or we just bond the SDS systems to the building steel without any riser cable. Now when the building is struck all the equipment floats to the same 50 KV at the same time. Now take it one step further and let?s say we leave the riser cable installed in the same manner as above without bonding to the steel and we have a simple ground fault in our electrical system to building steel. The fault current has to travel all the way to the basement and back up the riser cable to return XO source. Chances are the impedance would not be low enough to operate the breaker, and earth has nothing to do with it in the first place. Our simple fix takes care of that.

Hope that helps.

Dereck Campbell, PE

 

[bphgravity]

No offense, but most of what I have read in the IEEE green book appears to be inaccurate if not flat-out wrong.

Again, you are putting way too much value in the function of grounding and assuming it is capable of much more than what countless studies and even basic Ohm's Law proves to be the real case.

The fact of the matter is, no matter how large the building is and no matter what size service is installed, the minimum required grounding system as required by the NEC is all that is essential for good service and operation and basic life and property safety.

Additional risk and hazard analysis needs to be made if other design considerations need to be made. However in most cases, additional or redundant grounding will not be a solution.

 

[dereckbc]

Bryan lets not beat up on IEEE too much here, I am biased as I am part of the org.

What I think is being overlooked by all is the operating voltage and is causing some disconnect and confusion.

In high and medium voltage applications the amount of rods, copper, and impedance is very important. As Bryan and I have eluded to ohm's law applies, At higher voltages a 5-ohm or lower system will pass enough fault current to operate devices. However at low voltages like 480 and 208 earth cannot be used as a fault or load path, it is forbidden period. On the low voltage side bonding is the only effective solution.

You are certainly free to bond the EGC busses and bonding conductors to earth as often as you like, but it adds nothing except unwanted noise into ground circuits.

Lawrence where I think everyone is butting heads is the demark of high voltage end point and the low voltage begin point. In a large commercial or industrial installation the service voltage will be high or medium, and as stated the GES is very important. But once you transform to low voltage, earth is just a reference point and has no other use from there on. From there it is EGC and bonding conductors.

Now if you want to talk about a communication facility, that is a topic for another day.

 

[lile001]

Unless you have a purpose like secured communications, explosives contained, medium/high voltage, or something like that, a low impedance ground system is not of much use. Let me point out a couple of examples if you have time to follow along.

Thanks, Dereck!

So in cases like medium voltage, what kind of recommendeed practices would you consider? Say a building with 6 - 4160V 5KVA service entrances? How about if you add a lightning protection system?

 

[dereckbc]

I suspected you had 4160, that is why I danced around it.

As for the lightning protection system, sub contract it to a certified UL-96 contractor. Have then draw up the prints, install, and have the system inspected by UL for a MASTER LABEL. Review UL-96A and NFPA-780, then examine/review the prints and stamp it. Your selling point to the client and your firm is insurance discount for the Master Label certification. Once you gain familiarity with the documents, in the future you can design, contract the labor, and arrange for UL inspection. For the first time sub it out.

As for the GES I have more questions than answers. Without seeing and being familiar it is hard to answer. Is the new service part of new construction? If so I would probable use a "Star Point" counterpoise under the the transformer pad. From there try to figure out how to bond to the other GES systems already in service either via sub-terrain or above ground.

Sorry not much of an answer on the GES part, need more info, lot more info.

 

[dereckbc]

Another thought is space is really limited, a chem rod or well backfilled with betonite clay.