Separately derived source grounding

[websparky]

Re: Separately derived source grounding

Bennie,

If the grounds shown above are earthed, how can there be a ground loop if the earth is not a conductor unless, we are talking about extremely high impedance?

How close / far apart do these grounds need to be before there is high impedance?

 

[bennie]

Re: Separately derived source grounding

Ground loops are one of the most important grounding problems in the commercial and industrial environment. Slightly different potential differences in two power system grounds can cause circulating current. The circulating current causes serious noise problems.

Equipment connected to the two power sources and having a common communication cable will experience a lot of problems with noise.

One power system grounded to the building steel, and one grounded to the system ground electrode, will all have some degree of circulating current. The magnitude will be dependant on the resistance between the two ground electrodes.

Eliminating the two earth connections is of prime importance. Hence single point grounding(earthing), is part of the cure.

 

[dereckbc]

Re: Separately derived source grounding

Originally posted by websparky:
Bennie,

How close / far apart do these grounds need to be before there is high impedance?

It depends on what frequency you are talking about. Look at a 750 MCM cable; you would think there should be no problem, right? Wrong.

A 10-foot piece of 750 MCM exhibits 3.8 micro-henries of inductance. So at 60 hz you have close to 0 ohms, at 1 Mhz you would have 24 ohms of inductance, at 10 Mhz you have 240 ohms, at 100 Mhz you have 2.4Khz and so on. If you use a smaller gauge like 14 AWG there is not much difference in the high frequency range since a 10-foot 14 AWG exhibits about 4.3 micro-henries of inductance.

There are two types of ground design. One is the "power/safety" range (DC to about 1 Khz), which is the NEC arena. The other is the "performance range" (above 1 Khz), which is in the equipment designer arena.

The "single point ground" system Bennie is referring to is a popular method used by high-tech companies, but they are not in reference to earth, rather in reference to the power source which could be a battery plant or transformer. Establishing a “ground window” at the power source makes it work. Now if you are in a large facility you will have several power sources, hence several "ground windows" and several "single point isolated ground planes" systems. Now you have a design problem. How do keep impedance low between systems. Simple, you use multi-point grounding to create several loops.

There are good loops, and there are bad loops. For example in the data centers I design I use a raised floor signal grid to reference all the transformers on the technical floor space. I construct it out of # 4 AWG on 2 to 4 foot centers. The impedance between any two points on the grid is no more than 30 ohms at 45 Mhz or lower even if measured at opposite corners of 200 feet or more. It is not important if the grid is actually connected to earth or not, but it is to comply with NEC requirements. What makes it work is all the loops on 2-foot centers.

Now here is something to think about. A "single point ground" system has extremely high impedance at high frequencies. Go back to my example of a 10-foot piece of 750 MCM. Now imagine a 100-foot piece of # 12 AWG you use for an EGC or even a short section of a GEC. What is the impedance at a high frequency? It is so high you can consider it an open circuit.

Bennie think about that for a while and we will discuss multi-point grounding verses single point grounding and how the two can be used to compliment each other.

 

[pierre]

Re: Separately derived source grounding

Ed
Thank you for the drawings, they really make this a lot easier to see.
In your second drawing, the connection between the two systems (your green wire)is the connection that I wonder about (not that I am saying you think it has to be such). The Equipment Grounding Conductor that is run with the circuit conductors would seem to me to be sufficient for connection between the two systems (at the transformer case). My understanding of tne NEC is that a 'bonding' conductor is required between two systems in the same building. I see a parallel path there. From reading Dereck's post it would seem that this is not such a 'terrible' situation. Maybe Dereck can explain it a little further.
Thanks

Pierre

 

[dereckbc]

Re: Separately derived source grounding

Pierre, lets try a different approach. For the sake of a discussion lets say we have a thirty-story building constructed of steel frame and we want to install a electrical distribution system on the 20th floor. The building management supplies a 480/277 riser bus with an insulated ground from the basement. In the basement the ground electrode consist of a system of rods and properly bonded to the water pipe and building steel.

All is good so far. We determine we need a 50 KVA transformer, which we install on the 20th floor, so we bond the Xo to the insulated ground riser to make you know who happy and make a SPG. Life is good until one day during a thunderstorm the building is struck by lightning. The newspaper headlines read “SEVERAL OFFICE WORKERS KILLED BY ELECTROCUTION AND SMOKE INHALATION DURING THUNDER STORM”. What happened?

The answer is simple. All the electrical equipment was referenced to earth in the basement. When the building was stuck at the top a voltage divider was formed from the top to bottom. The 20th floor was at 10,000 volts where the employees were while the electrical equipment was at 0 volts from the earth reference in the basement. What is the solution?

Answer is simple, bond the Xo of the transformer to building steel on the 20th floor. Had that been done to code nothing would have happened, every thing would have been at the same 10,000-volt level. Like I keep trying to say the SPG is in reference with the source, not earth.

Personally I have installed designed hundreds of sites under the same circumstances. And here is how I approach it. I install a ground bus on the floor where the equipment will be installed. I bond the bus to the ground riser, building steel and any water pipes in the area. All the loops formed by the building steel framework lower the AC impedance much lower than the ground riser cable presents by itself. I then go from there everything is safe and noise free.

Bennie are your ears burning yet?

[ September 03, 2003, 06:22 PM: Message edited by: dereckbc ]

 

[karl riley]

Re: Separately derived source grounding

Dereck, your explanations are extremely clear. I could never understand the talk about referencing everything to the earth's zero potential, since the potential of earth is always relative. So your grids and the building steel are the real references. This all makes sense.

I have seen high rises where the building was concrete and the upper-floor Tformers were grounded to a piece of reinforcing rod at that floor level. I wondered about the continuity, but would this still be better than running an insulated GEC to the basement?

Karl

 

[websparky]

Re: Separately derived source grounding

dereckbc,

Thank you for your explanations above concerning SPG. Makes perfect sense to me !

It seems there is much to be learned and discussed about equalizing potential differences and their effects on systems as well as persons.

Dave

 

[dereckbc]

Re: Separately derived source grounding

Originally posted by karl riley:
I have seen high rises where the building was concrete and the upper-floor Tformers were grounded to a piece of reinforcing rod at that floor level. I wondered about the continuity, but would this still be better than running an insulated GEC to the basement?

Karl

Absolutely! I would use both and bond at the floor where it is used. And if possible I would try to get permission to bond the ground riser to the rebar at multiple places like each floor.

 

[shoffnerm]

Re: Separately derived source grounding

I want to THANK YOU ALL!!!! The replies to this question have been very helpful.

I see that Pierre has seen my concern and questioned ?My understanding is the transformer GEC is connected as close as possible, one reason is to maintain an equal potential between the new source and the surounding area (almost like a equipotential plane).

I am not sure why we need to bond the two systems together with a separate conductor when there is already a connection via the equipment grounding conductor to the transformer case (x/o).

If at the new SDS and a separate conductor is to bond the two systems together?This is what I am seeing a s a LOOP.

According to Dereckbc the loops that I am seeing are actual parallel paths to ground. Which I can agree with. But according to the logic of tripping a breaker, the current is seeking the path back to the source X/O to trip that breaker or blow that fuse.

The way I see it?at the new source after the SDS? the new distribution system protection is the new X/O connection. I hope I am seeing this correctly.

I also have a lightning protection requirement in my scheme and if I connect this to my ground system (single point ground system) ?if I get a strike and the lightning travels to the least path to ground. Which is the best method to predetermine which route it will take?

Lets say we place a ground electrode at the first source and at every new SDS and the lightning protection. Connect these Ground electrodes through the equipment grounds. This lightning would have to dissipate though the system equipment grounds.

With the same scenario? if all the grounds are connected together ( as a ground plane) with additional separate conductors? WILL the lightning find its path to ground through the ground plane and dissipate through the ground plane and not travel through the equipment grounds? These are the LOOPS that I am concerned about.

 

[shoffnerm]

Re: Separately derived source grounding

I have an established industrial building and I am adding separate outside pads of portable equipment. I will be utilizing two branch circuits from two separate distribution systems within the industrial building. I will describe only one branch circuit.

I have a 20kV to 440/254 VAC 50 Hz transformer. This has a NG Bond at the substation which is 20 ft from the industrial building distribution panel. At my 440/254 Distribution panel the NG is bonded ?one time?.

For the branch circuit I will run 3 phases and 1 ground conductor ? 500 Ft? to this outdoor pad. On this pad I have a disconnect. And according to Ed?s schematic I am connecting the Equipment ground to the enclosure. From the disconnect I have three phases and a ground to a MG set (Motor Generator). The MG set is 440 VAC 50Hz to 208/ 120 60Hz VAC. Since this is a SDS? I will make a new NG Bond to ground. With my distance? I am going to install a new ground electrode. Again with my distance and this all being on the same premises? DO I Connect a separate conductor to connect the ground rods??? OR are the equipment grounds sufficient.???

My Pad? does not have footers? just simply a slab on grade. It has rebar throughout. I have two ground triangles on the outside of the pad. With the pad system isolated from the source grounding system? the ground triangles? alone? measured at 105 ohms ground resistance. After connecting the slab to the triangles ?my total ground system is at 3.2 ohms. I have 4 each connecting points installed for various equipment requirements as originally requested. But now I am concerned about these points and possibly creating loops.

After the MG set I have a Distribution center with an input 4-pole Manual Transfer Switch. I have an alternate 208/120 60 Hz VAC Generator as an alternate source in case of failure of my Mains. I am going to have the generator NG bonded and connected to Ground. Since my sources are with in close proximity (10 ft.) I am going to make each NG bond for the separate sources at the same ground connection point on the slab. I do not see a problem with this as it is an either or situation.

After this distribution center it is distributed to the loads. Thank goodness something normal. It will be the typical P-N-G for each circuit.

I also have a Communication signal ground requirement. This equipment is isolated from the equipment ground requirements. I am going to connect a single isolated ground from the distribution panel ground to these separate requirements. Originally I was told to make a separate ground system? But I understand this is very dangerous from a past posting I made.

There is an additional requirement? But I need all details before posting. But for now? ALL OF YOU OUT THERE?. THANK YOU!!!!! For all your inputs.

R/ Martin

 

[pierre]

Re: Separately derived source grounding

Dereck
Thanks for the reply! When I explain something to somebody I ask them to repeat it back to me, this way I know they understand it. So I am going to repeat it back to you to see if I have grasped what you are trying to explain, no disrespect is meant.

In the 20 story steel constructed building I am going to install a transformer lets say 480/277 to 120/208. My understanding was (is) that I am to go to building steel (the inspectors usually want this connection 'above' the transformer)as close to the transformer as possible. I make the connection to x/o in the transformer and also connect x/o to the case which is essentially connected to the primary (480) equipment grounding conductor.
In the past that would usually be the extent of the grounding.
Now, do I need to (by code) run a bonding connection from the utility service to the separately derived system? 250.30(A)(1)

My understanding about the grounding electrode at a SDS in a large steel framed building is that it is not for lightning protection of the SDS. Is this correct?

I have to say, as I ask these questions more pop into my head. I know this can keep going on, but I will stop with the questions for now and see if I can understand this part first.
Thanks

Pierre

 

[don_resqcapt19]

Re: Separately derived source grounding

Pierre,

Now, do I need to (by code) run a bonding connection from the utility service to the separately derived system?

No additional connection is needed. If you are using the building steel as the grounding electrode, it must be effectively grounded. If the steel is effectively grounded, it is required to be used as a grounding electrode for the service by 250.50. If the building steel was not used as a service grounding electrode, it would still be bonded to the service grounding system because of 250.104(C).
Don

 

[dereckbc]

Re: Separately derived source grounding

Pierre, I cannot add anything to what Don said, he is right on target.

 

[dereckbc]

Re: Separately derived source grounding

Originally posted by pierre:
My understanding about the grounding electrode at a SDS in a large steel framed building is that it is not for lightning protection of the SDS. Is this correct?

Yes and No. If the secondary circuit conductors only serve equipment inside the structure then there is not much of a chance for lightning enter. The bond to building steel is in case one of the secondary conductors accidently comes in with the steel. It just provides a planned low impedance path.

If the transformer supplies circuits that go outside where exposed to the elements, then yes it would provide some lightning protection by providing a direct path to a ground electrode rather than rely on the primary EGC.

 

[pierre]

Re: Separately derived source grounding

This was great, I have a very confident feeling about my understanding of this subject. Now all I need to do is find a way to get all of you on to my jobs when the inspector is there.
Thanks

Pierre