Generator bonding

Status
Not open for further replies.
ceb58

ceb58

Senior Member
Location
Raeford, NC
Situation. 45kw gen. set non-SDS 1ph 240v at communication tower. Contractor that installed the gen. and transfer switch bonded the neutral and ground at the gen. disconnect which I know is wrong and will be changed PDQ. But what has me wondering is they did not pull an EGC with the feeders from the gen set to the disco they instead bonded the frame off of the underground grounding ring with a #2 solid. There are 2 places on the gen. to install an EGC. One is on the frame and the other is on the sheet metal enclosure where the breaker on the gen. is located. Mechanically they are together through the gen. frame. Should another EGC be pulled with the feeders and be bonded in the enclosure? We have had some lightning strikes at this site and every time something in the transfer switch and/or the gen. gets fried. I think the main reason is due to the fact they did not separate the N-G in the gen. disconnect.
 
C

Cold Fusion

Senior Member
Location
way north
Keep in mind that nearly all of my experience is with 3 phase, generally 480V. About the only single phase I do is my own house.

Sounds like the gen is outside. True?

If so, I have seen some discussion on lightning strikes having a traveling wave across the ground. Might have been an MH grounding class several years back.

The premise was the closer earthed parts went up in voltage before the farther away parts. So there would be a large potential across any metalic interconnection between the two parts.

I agree with removing the N/G bond at the gen.

Now this ground ring: Is it just around the gen pad? Or does it circle the gen pad, structure, and tower?

I suspect the ground ring encircles all. If so, I'm not sure that removing the gen N/G bond will alleviate the damage the gen/transfer switch caused by the lightning strikes.

Have you looked at NFPA 780 yet? That might be a good place to start.

cf
 
walkerj

walkerj

Senior Member
Location
Baton Rouge
If it is a non-SDS, the only N-G bond should be at the service and the N-G bond removed in the generator(if present.)

The EGC is run FROM the service and TO the genny.

Or am I way off?
 
ceb58

ceb58

Senior Member
Location
Raeford, NC
How is a generator a "non-SDS"?:confused:
Click to expand...
By keeping the utility and generator grounded conductors solidly connected. See the definition of SDS in Art. 100.
Click to expand...
Thanks Larry

I suspect the ground ring encircles all. If so, I'm not sure that removing the gen N/G bond will alleviate the damage the gen/transfer switch caused by the lightning strikes
Click to expand...

Yes, the ring encircles every thing. The fence,tower legs and gen frame and main grounding bar for the inside halo are bonded to the ring. It is also the GEC to the main service disconnect.
Removing the N-G bond in the Gen disconnect will make it code compliant and prevent what I think happened last week after a strike. One of the boards in the ATS was fried, as in hole burned through it, I think the cabinet was energized due to the N-G bond that should not be there.
 
C

Cold Fusion

Senior Member
Location
way north
ceb58 said:
... Yes, the ring encircles every thing. The fence,tower legs and gen frame and main grounding bar for the inside halo are bonded to the ring. It is also the GEC to the main service disconnect. ...
Click to expand...
That sound good.

ceb58 said:
... Removing the N-G bond in the Gen disconnect will make it code compliant ...
Click to expand...
And you are also pulling an EGC from the gen frame with the gen feeders? You should. And are you going to keep the connection from the gen frame to the ground ring? I would recommend it.

ceb58 said:
... Removing the N-G bond in the Gen disconnect will ... prevent what I think happened last week after a strike. One of the boards in the ATS was fried, as in hole burned through it, I think the cabinet was energized due to the N-G bond that should not be there.
Click to expand...
Okay, sketch this one out, show us where you think the currents are going that caused the burn up. Then show how these currents are interupted if the NG bond at the gen is removed.

I don't see it.

I don't think the N/G bond should be there either. My reason is the ground ring is a metalic path between the gen frame and the service neutral. And the bond creates a parallel path between the gen neutral and the service neutral.

Getting rid of the parallel neutral path is good, but I don't know that will fix your issues with lightning strikes. The NEC is not a design guide that will save equipment from lightning strikes.

Then again maybe you see something I don't and your sketches will show it - I'd like that. I always like learning something new.

cf
 
ceb58

ceb58

Senior Member
Location
Raeford, NC
Cold Fusion said:
That sound good.


And you are also pulling an EGC from the gen frame with the gen feeders? You should. And are you going to keep the connection from the gen frame to the ground ring? I would recommend it.


Okay, sketch this one out, show us where you think the currents are going that caused the burn up. Then show how these currents are interupted if the NG bond at the gen is removed.

I don't see it.

I don't think the N/G bond should be there either. My reason is the ground ring is a metalic path between the gen frame and the service neutral. And the bond creates a parallel path between the gen neutral and the service neutral.

Getting rid of the parallel neutral path is good, but I don't know that will fix your issues with lightning strikes. The NEC is not a design guide that will save equipment from lightning strikes.

Then again maybe you see something I don't and your sketches will show it - I'd like that. I always like learning something new.

cf
Click to expand...

Maybe I misworded it, Just thinking out loud.

The main service disconnect has the N-G bond as it should. It then comes into the ATS where the 3 neutrials ( utility,gen and E side) are isolated and the 3 grounds are bonded to the can. The gen load side N and a G go out to the gen. disconnect. This is where the G , load side N and line side N from gen are bonded again. From the disconnect to the gen they did not pull an EGC with the phase conductors but relied on the bond to the ring.

As said I know the N-G bond must be changed. And I think we need to install the EGC from the disco. to the gen also.

This will get rid of the parallel path. As I said I was thinking out loud but if nothing else getting rid of the N-G bond may be it will at least bring every thing to the same potential. This may not stop damage from a strike but at least the install will be as it should and thats all I can do;)
 
augie47

augie47

Moderator
Staff member
Location
Tennessee
Occupation
State Electrical Inspector (Retired)
Short answer would be yes and EGC should be pulled with the generator feeder. The EGC needs to be there and not pulling it with the feeder would be a violation of 300.3(B), IMO.
That said, having seen how the significant grounding grid is installed at most cell sites I would not loose sleep worrying over the ground path.
 
H

hurk27

Senior Member
Location
Portage, Indiana NEC: 2008
Two things come to mind on this install, PVC, and induction of very high currents from near field lightning strikes.

From your description of the install it sounds like both control conductors and power conductors are run separately in PVC raceway methods? this by itself wouldn't seem to be a problem but add in a lightning strike or two in very close proximity to these runs and you have a recipe for very high voltage to be induced into the control wiring between the generator and transfer switch.

Lightning is a high frequency event and can induce current into near by conductive pathways, in this case you have two, the control conductors and the power conductors, each pathway will act just like a single pole of a transformer winding, the power conductors will in most cases withstand the induced current, but the control conductors (or should I say the pathways in the control board) can't, and are destroyed.
Any time I have done tower installs or cell sites I run RMC between towers or pieces of equipment when there is multiple raceways between the two points, I then make sure the raceways are electrically bonded together to form only one pathway, sizing the bonding conductors as I would any GEC

The other point of failure I see is that you have the ring electrode bonded to the generator frame, this and all electrodes should form a single point bonding at the main service disconnect where the main bonding jumper is located, this will help in not allowing the conductors to the generator becoming a pathway for a direct lightning strike.

Most of the equipment buildings I done have a copper bonding plate right where all the cables enter or exit the building, this is the point I make my single point bonding to with all my electrodes and any other bonds.
 
Last edited:
ceb58

ceb58

Senior Member
Location
Raeford, NC
hurk27 said:
Two things come to mind on this install, PVC, and induction of very high currents from near field lightning strikes.

From your description of the install it sounds like both control conductors and power conductors are run separately in PVC raceway methods? this by itself wouldn't seem to be a problem but add in a lightning strike or two in very close proximity to these runs and you have a recipe for very high voltage to be induced into the control wiring between the generator and transfer switch.

Lightning is a high frequency event and can induce current into near by conductive pathways, in this case you have two, the control conductors and the power conductors, each pathway will act just like a single pole of a transformer winding, the power conductors will in most cases withstand the induced current, but the control conductors (or should I say the pathways in the control board) can't, and are destroyed.
Any time I have done tower installs or cell sites I run RMC between towers or pieces of equipment when there is multiple raceways between the two points, I then make sure the raceways are electrically bonded together to form only one pathway, sizing the bonding conductors as I would any GEC

The other point of failure I see is that you have the ring electrode bonded to the generator frame, this and all electrodes should form a single point bonding at the main service disconnect where the main bonding jumper is located, this will help in not allowing the conductors to the generator becoming a pathway for a direct lightning strike.

Most of the equipment buildings I done have a copper bonding plate right where all the cables enter or exit the building, this is the point I make my single point bonding to with all my electrodes and any other bonds.
Click to expand...

In reality there are three paths. Feeders from gen, 120v circ. for charger/heater and control wireing ( 22ga 6pr id. shielded) All are in individual pvc runs. There is a main bonding plate on the out side tied into a bonding plate inside. The bonds from the out side are tied onto it.
In a nut shell are you saying remove the bond from the ring that is on the frame, then pull an EGC with the phase conductors ( which should have been done) to bond the gen? Very little we can do about the pvc. When we go back we are going to do a FOP test to see what ohms to ground we are getting. We shoot for less than 5ohm
 
H

hurk27

Senior Member
Location
Portage, Indiana NEC: 2008
ceb58 said:
In reality there are three paths. Feeders from gen, 120v circ. for charger/heater and control wireing ( 22ga 6pr id. shielded) All are in individual pvc runs. There is a main bonding plate on the out side tied into a bonding plate inside. The bonds from the out side are tied onto it.
In a nut shell are you saying remove the bond from the ring that is on the frame, then pull an EGC with the phase conductors ( which should have been done) to bond the gen? Very little we can do about the pvc. When we go back we are going to do a FOP test to see what ohms to ground we are getting. We shoot for less than 5ohm
Click to expand...

I understand about the PVC when we inherit a job like this, are the PVC runs together in the same trench?, if so what about laying a copper grid mesh just above them and bonding this grid to the main bonding point, if not then theres not much you can do, unless they want to fork out to re-run in a metal raceway, removing the bond to the ring from the generator frame will help, but as I said above having separated runs in PVC will be subject to induced current from near field strikes.

When trying to visualize single pathways between equipment I use a image of a spider, the body is the single bonding point, and each leg is the single pathway to each piece of equipment, each leg can have multiple raceways, but if they are bonded together at both ends they will act like one pathway, and we all know current cant flow on just one pathway, kind of like having a transformer with only one conductor coming out of it.
The ground ring is just providing another pathway for the lightning current to flow to the generator and out its frame, the low voltage in most generators also bond the DC neg to the frame, so this can also create a pathway through the control conductors and through the control board, so yes removing it and installing the EGC should help.
 
ceb58

ceb58

Senior Member
Location
Raeford, NC
hurk27 said:
I understand about the PVC when we inherit a job like this, are the PVC runs together in the same trench?, if so what about laying a copper grid mesh just above them and bonding this grid to the main bonding point, if not then theres not much you can do, unless they want to fork out to re-run in a metal raceway, removing the bond to the ring from the generator frame will help, but as I said above having separated runs in PVC will be subject to induced current from near field strikes.

When trying to visualize single pathways between equipment I use a image of a spider, the body is the single bonding point, and each leg is the single pathway to each piece of equipment, each leg can have multiple raceways, but if they are bonded together at both ends they will act like one pathway, and we all know current cant flow on just one pathway, kind of like having a transformer with only one conductor coming out of it.
The ground ring is just providing another pathway for the lightning current to flow to the generator and out its frame, the low voltage in most generators also bond the DC neg to the frame, so this can also create a pathway through the control conductors and through the control board, so yes removing it and installing the EGC should help.
Click to expand...

No way they will fork out money to rerun, this is the State we are talking about. Mesh grid I doubt it. And the way it looks there is a 99.9% chance they are in the same ditch. But I would like to run this by you. We have at this time around 150 of these sites up and running, this one I am working on is on top of a mountain, but I went and looked at a site today that I personally witnessed a strike last year. When this tower took the hit the lightning struck the top traveled down about 300 ft and arked over to a transformer about 30 ft away. Fun show lets get the H*** out of here. The only damage the site received was it blew the lamp out in the top strobe. The bonding is the same at this tower as the one I am working on the only difference is the contractor did as he should and ran an EGC with the phase conductors to the disconnect and did not re-bond the N-G. Now In my way of thinking this gen. only has a bond from the ring. The neutral is isolated from the frame. If the tower takes a hit the currant flowes on to the ring. With every thing else it has a free path to flow. But on the gen. it has no where to go. It cannot flow through back to the main bonding point which in theory, I think, would bring every thing to the same potential lesson the chance of damage. Am I off base, or have I been drinking too much sweet tea:D Also I am posting another thread in this topic I would like your opinion on. Too much time driving and thinking.
 
Status
Not open for further replies.
Top