EGC, fault current path, and ungrounded 480V 3ph Delta

Status
Not open for further replies.
P

phaset

Member
Location
Atlanta
Hey all,

I'm trying to square the code in regards to fault current with what common sense and physics tells me:

Imagine an industrial circumstance (no general public access) where the service is 3ph 480 3wire with no neutral, nor a grounded leg. I don't currently know if this is the case, but this possibility was tossed out there by another party involved with the project in advance of seeing the site (state of California - CEC is based on NEC 2008).

Further suppose that equipment is tied to an EGC created by the bonded conduit back to the service (per 250.118), and there's a GEC and grounding electrode at the service, and they are connected to the egc per 250.24(D).

---

Since the secondary is not grounded, there is no fault current return path to the xfmr. Also, since the voltage is less than 600V, the grounding electrode is excluded from being the fault current path (250.4).

Practically this means that a fault to EGC only serves to raise the bond path GEC, and GE to phase voltage, and more than likely not tripping the overcurrent and clearing the fault, leaving an energized grounding path.

How does one square up this reality with the code's grounding, bonding, and fault current path requirements? Is there a specific section and/or exceptions in the code that allow for the operation of an ungrounded 3wire system of this type (beyond 250.20)?

P.S. If the current infrastructure turns out to be floating, it will have to remain so. There will be no opportunity for transition to grounded delta or use of a zig-zag ground.

Thanks in advance for any and all clarity!
 
phaset said:
Hey all,

I'm trying to square the code in regards to fault current with what common sense and physics tells me:

Imagine an industrial circumstance (no general public access) where the service is 3ph 480 3wire with no neutral, nor a grounded leg. I don't currently know if this is the case, but this possibility was tossed out there by another party involved with the project in advance of seeing the site (state of California - CEC is based on NEC 2008).

Further suppose that equipment is tied to an EGC created by the bonded conduit back to the service (per 250.118), and there's a GEC and grounding electrode at the service, and they are connected to the egc per 250.24(D).

---

Since the secondary is not grounded, there is no fault current return path to the xfmr. Also, since the voltage is less than 600V, the grounding electrode is excluded from being the fault current path (250.4).

Practically this means that a fault to EGC only serves to raise the bond path GEC, and GE to phase voltage, and more than likely not tripping the overcurrent and clearing the fault, leaving an energized grounding path.

How does one square up this reality with the code's grounding, bonding, and fault current path requirements? Is there a specific section and/or exceptions in the code that allow for the operation of an ungrounded 3wire system of this type (beyond 250.20)?

P.S. If the current infrastructure turns out to be floating, it will have to remain so. There will be no opportunity for transition to grounded delta or use of a zig-zag ground.

Thanks in advance for any and all clarity!
Click to expand...

If it is now floating, a fault to EGC will not raise the EGC, it will reference the fault point on the circuit to earth ground. No current will flow.
A second phase to EGC fault can then send beaucoup current through some part of the EGC raising it well above ground and hopefully tripping the OCPD.
For this reason NEC requires a ground detector to alert to the first fault. 250.21(B).

Bonding refers only to the EGC bonding continuity requirement which is unchanged and the grounded conductor (not a neutral) to EGC bond which will happen only in one place at the service point just like with a wye or 120/240 3-wire (plus EGC) service.
 
With a 480 volt ungrounded system, you have a definite need for bonding by a 250.118 means.
Without proper bonding you could conceivably have to pieces of adjacent equipment with ground faults developing a potential 480v between each piece.

The GEC, as with other systems, gives you a reference to earth. Someone else may show me differently, but I fail to see it's importance on an ungrounded system relative to grounded system..
 
phaset said:
Practically this means that a fault to EGC only serves to raise the bond path GEC, and GE to phase voltage, and more than likely not tripping the overcurrent
Click to expand...

Not tripping the over-current device is the entire reason ungrounded systems are chosen.

All the grounded / bonded parts remain at the same potential to each other and the earth so it remains safe.
 
GoldDigger said:
If it is now floating, a fault to EGC will not raise the EGC, it will reference the fault point on the circuit to earth ground. No current will flow.
Click to expand...

GoldDigger, thanks for the reply. I'm going to double check all my thinking below:

I think I now see what I did in my mind, which was to reflexively create a current path from GE to secondary. But with a floating secondary this cannot occur, right?

A second phase to EGC fault can then send beaucoup current through some part of the EGC raising it well above ground and hopefully tripping the OCPD.
For this reason NEC requires a ground detector to alert to the first fault. 250.21(B).
Click to expand...

Because the current bridges phase to phase via the EGC, right?

Bonding refers only to the EGC bonding continuity requirement which is unchanged and the grounded conductor (not a neutral) to EGC bond which will happen only in one place at the service point just like with a wye or 120/240 3-wire (plus EGC) service.
Click to expand...

But in the absence of a grounded secondary conductor, or a neutral, where is a grounded conductor to EGC bond occurring? Isn't the EGC only going GEC -> GE, and no where else? Am I missing something here?
 
iwire said:
Not tripping the over-current device is the entire reason ungrounded systems are chosen.

All the grounded / bonded parts remain at the same potential to each other and the earth so it remains safe.
Click to expand...

Right, but how do I square this with the code requiring a low impedance path for the fault current?
 
phaset said:
I think I now see what I did in my mind, which was to reflexively create a current path from GE to secondary. But with a floating secondary this cannot occur, right? [1]

Because the current bridges phase to phase via the EGC, right? [2]

But in the absence of a grounded secondary conductor, or a neutral, where is a grounded conductor to EGC bond occurring? Isn't the EGC only going GEC -> GE, and no where else? Am I missing something here?[3]
Click to expand...
1. Right. No closed current loop.
2. Right. And since some point on the EGC is referenced to ground, the current and the resistance in the EGC will cause a voltage difference relative to that point. Which will directly energize supposedly grounded metal surfaces. Ouch!
3. I was momentarily thinking of a corner grounded delta rather than an ungrounded delta. My bad. :) (Got two threads crossed in my mind.)
 
Equipment Grounding for ungrounded and HRG is to reduce touch potential to prevent electric shock or electrocution on a single line to ground fault. Grounding Electrode Conductor may facilitate the operation of the OCPD on a second line to ground fault and for this reason GEC should have enough cross sectional area for a return path to operate the OCPD on a second line to ground fault.
 
We have quite a few older 3W ungrounded 480V pump station services, many of which were built before ground fault detection was required. All of the switchgear is bonded and grounded together. Any phase going to ground will produce no fault current. But as was mentioned above, all enclosures and conduits MUST be bonded together, otherwise opposite phases grounding unbonded enclosures can allow YOU to become a path for fault current when you bridge between the two. Probably ruin your day. I know the subject has been battered to death here before, but it is my opinion that ungrounded is safer because if you are working the gear hot, any accidental contact between you and ground probably will not be fatal. However, if one phase becomes grounded and you or a wire gets between the grounded cabinet and either of the two ungrounded phases........well, I'll leave it up to your imagination.
 
Required

Required

meternerd said:
We have quite a few older 3W ungrounded 480V pump station services, many of which were built before ground fault detection was required. All of the switchgear is bonded and grounded together. Any phase going to ground will produce no fault current. But as was mentioned above, all enclosures and conduits MUST be bonded together, otherwise opposite phases grounding unbonded enclosures can allow YOU to become a path for fault current when you bridge between the two. Probably ruin your day. I know the subject has been battered to death here before, but it is my opinion that ungrounded is safer because if you are working the gear hot, any accidental contact between you and ground probably will not be fatal. However, if one phase becomes grounded and you or a wire gets between the grounded cabinet and either of the two ungrounded phases........well, I'll leave it up to your imagination.
Click to expand...

These systems are "required" to be grounded...If you find services that are really "ungrounded" please report to the local
electrical subcode or building official...can you walk away from a landmine without telling anyone?
 
clarification...

clarification...

In the past many things have been allowed...current code in 250.20 and 250.21 "requires" services to be either grounded
or have a ground detector installed. (didn't mention the ground detector in previous post..)
An ungrounded service without fault detection is a landmine...which is mostly my point.
 
[In the past many things have been allowed...current code in 250.20 and 250.21 "requires" services to be either grounded
or have a ground detector installed. (didn't mention the ground detector in previous post..)
An ungrounded service without fault detection is a landmine...which is mostly my point.]


I know this is an older post, but I just read your reply. Why do you think an ungrounded Delta service is a "landmine"? Even without ground detection, if any phase were to become grounded, nothing would happen. If two phases become grounded, it's a phase to phase fault, exactly like a phase to phase fault on any three phase service, Delta or Wye. Where's the hazard?
 
meternerd said:
[In the past many things have been allowed...current code in 250.20 and 250.21 "requires" services to be either grounded
or have a ground detector installed. (didn't mention the ground detector in previous post..)
An ungrounded service without fault detection is a landmine...which is mostly my point.]


I know this is an older post, but I just read your reply. Why do you think an ungrounded Delta service is a "landmine"? Even without ground detection, if any phase were to become grounded, nothing would happen. If two phases become grounded, it's a phase to phase fault, exactly like a phase to phase fault on any three phase service, Delta or Wye. Where's the hazard?
Click to expand...
The hazard is that phase-to-phase fault current is flowing in the EGC, potentially raising the voltage of "grounded" metal surfaces downstream of the fault. These would be machines and raceways which would not normally be affected as severely by a phase-to-EGC fault in a grounded system.
A lot depends on the details of how the entire EGC network is wired.
 
meternerd said:
[In the past many things have been allowed...current code in 250.20 and 250.21 "requires" services to be either grounded
or have a ground detector installed. (didn't mention the ground detector in previous post..)
An ungrounded service without fault detection is a landmine...which is mostly my point.]


I know this is an older post, but I just read your reply. Why do you think an ungrounded Delta service is a "landmine"? Even without ground detection, if any phase were to become grounded, nothing would happen. If two phases become grounded, it's a phase to phase fault, exactly like a phase to phase fault on any three phase service, Delta or Wye. Where's the hazard?
Click to expand...

Imagine something in a building becoming energized like a piece of machinery, or a control cabinet....etc. someone with luck like mine
will eventually come in contact with this energized equipment...with my luck I'll be standing on the top step of an aluminum ladder.....
I know it is a "what if?", but GFCI's and AFCI's guard against the "what ifs" also...
 
SparkyHC said:
Imagine something in a building becoming energized like a piece of machinery, or a control cabinet....etc. someone with luck like mine
will eventually come in contact with this energized equipment...with my luck I'll be standing on the top step of an aluminum ladder.....
I know it is a "what if?", but GFCI's and AFCI's guard against the "what ifs" also...
Click to expand...

If the cabinet you refer to is grounded, there will be no energized cabinet. All that will happen is that the grounded phase will be "brought down" to cabinet potential, which is zero to ground. That's why corner grounded delta works. Agreed a phase to ground fault on either of the other two phases can produce a phase to phase fault, but that's why we have breakers. Haven't convinced me yet! No offense...just an old, nearly retired fart who loves to argue.:p If I didn't mention it before, I was a Navy submariner, and in the Navy, NOTHING is grounded, because when you get depth charged, you don't want one ground tripping the system. Nobody died from an ungrounded system that I know of.:D Man I love this profession!

Thanks for all of your input. An educated electrician is a safe electrician. At the end of the day we all want to go home with ALL of our body parts!
 
Status
Not open for further replies.
Top