Equipment grounding conductor sizing for increased phase conductors

[bob]

the 541 bolted fault amps would be a multiplyer of 3.6 for the Square D QOD breaker chart, this would give the 150 amp breaker a clearing time of 1.125 seconds to 30 seconds. The 5 second ICEA 150?C rating for #6 is 621.27 amps which means the insulation will be damaged in about 6 to 7 seconds at 541 amps, since it falls between the 1.125 to 30 seconds it looks like it can be damaged if I'm reading the QOD chart correctly, I can't belive it has that big of a window, but it seems to? I dont remember insta-trip breakers having this?
but a #4 looks like it would do just fine
If the voltage to X0 was 277 volts, the #6 would not be a problem, even at 240 volts.
Guys I'm learning too so go easy on me

The problem comes with an arcing fault. When you have arcing fault, the magnitude of the fault is reduce significantly. With the arc impedance included, your 541 amps might be reduce to 300 amps or less. The time to trip the breaker might be in the minutes. This presents a hazard not only to the equipment but more importantly to humans working near the equipment.

 

[rbalex]

I?m sorry I couldn?t respond earlier; I?m attending a local IEEE conference this week.

I^2t is a basic short circuit concept. It essentially says for short periods of time damage is proportional to current squared times time. It is based on some well established physics and a few reasonable assumptions:

For a resistive fault,

V=IxR
Power = VxI = IxRxI = I^2xR
Energy = Power x t = I^2xRxt

Therefore, for short periods of time, with constant R, energy released and consequential damage at the point of the fault is proportional to I^2xt. Note as Smart $ has pointed out R will eventually change over time, which is why ?short periods? usually 2 seconds or less is necessary for it to be valid.

What is important to recognize is ALL faults are resistive at the point of contact or no energy would be released at all. Coincidental damage may occur of course, such as insulation damage caused from conductor overcurrent, but it is still the resistive element of the conductor that creates the heat.

What is more telling is a 1500A fault could take as long as 2.5 seconds on ­any of the curves shown ? and upsizing the EGC won?t solve it.

As I have said, and every bit of math that anyone else has thrown at it has confirmed, the only practical case where upsizing the EGC has any hope of actually solving a problem is on 120V and below circuits where the OCPD responds only to overcurrent ? and it can?t even guarantee it will be successful in that case either. Therefore, upsizing the EGC has no basis being a general requirement. Other, usually more expensive, techniques (that actually work) need to be applied, and I shouldn?t have to waste my client?s money on upsized EGCs.

 

[SG-1]

While looking for the time/current characteristic curves I came across current limiting breakers. I did not even know they existed before. Their curve seems to fill in most of the protection gap down to the 2X rated current, but there is still a small gap present in the overcurrent protection.

Anybody have any experience with these ?
I have not checked the prices yet.

 

[bob]

While looking for the time/current characteristic curves I came across current limiting breakers. I did not even know they existed before. Their curve seems to fill in most of the protection gap down to the 2X rated current, but there is still a small gap present in the overcurrent protection.

Anybody have any experience with these ?
I have not checked the prices yet.

In this discussion, the problem is generating enough current to cause the breaker to trip. You would not want to limit it. These type breakers are use to limit fault current so that it does not exceed panel or breaker ratings.

 

[Smart $]

...I shouldn?t have to waste my client?s money on upsized EGCs.

With no offense to your credibility to calculate an appropriate sized EGC, perhaps a happy medium would be to include an exception to 250.122(B). Perhaps something to the effect of...

Increased-in-size equipment grounding conductors of the wire type shall not be required where the equipment supplied by the ungrounded conductors is bonded by at least one other equipment grounding conductor of the following types: 250.118 [permitted types list numbers here].

It may need some polish... but I think it conveys the intent of my proposal.

 

[ActionDave]

In this discussion, the problem is generating enough current to cause the breaker to trip. You would not want to limit it. These type breakers are use to limit fault current so that it does not exceed panel or breaker ratings.

So why is it OK to put the circut on a standard breaker with a higher ampacity?

 

[SG-1]

In this discussion, the problem is generating enough current to cause the breaker to trip. You would not want to limit it. These type breakers are use to limit fault current so that it does not exceed panel or breaker ratings.

Bob, the time/current characteristic curve on these breakers is such that if I had a 100 amp breaker only 150A of fault current would cause it to trip in under 11 seconds. With 200 amps of fault current the trip would take place between .02 & 6 seconds. With 250 amps it would interrupt the fault in the first half cycle.

With the standard breakers that only have a thermal element the trip times could be several minutes.

The idea is to provide protection with lower levels of fault current. This is just one option if ground fault protection is not used. I agree that ground fault protection will provide the maximum level of protection here.

I do see an issue if there is 2X inrush current on a current limiting breaker.

We have established that upsizing the EGC will not get enough fault current to trip a standard breaker in a time we are comfortable with. This type breaker limits fault current it by turing it off in the first half cycle using much less fault current simular to a current limiting fuse.

Steve

 

[rbalex]

No offense taken.

Even if there is a general problem caused by upsizing the CCCs, upsizing the EGC is still not a general solution. Since there isn?t a general problem in the first place though, upsizing definitely shouldn?t be a general requirement.

Auxiliary bonding is simply more of the wrong solution since it is no more certain to solve the alleged problem than upsizing.

I would prefer that upsizing as a general requirement be removed altogether; but that won?t happen. An effective exception would recognize alternate ground fault detection means instead - if a general problem were established in the first place.

 

[Smart $]

Even if there is a general problem caused by upsizing the CCCs, upsizing the EGC is still not a general solution. Since there isn’t a general problem in the first place though, upsizing definitely shouldn’t be a general requirement.

I agree, but not entirely... I see a problem in one area, but it exists even when the circuit conductors aren't upsized. That is when wire-type EGC is the only grounding means (other than earth) for the circuit.

Auxiliary bonding is simply more of the wrong solution since it is no more certain to solve the alleged problem than upsizing.

I'm not suggesting that auxiliary bonding be added, but rather accept coincidental bonding be recognized... such as when the circuit is run in metal conduit, where the conduit is acceptable as an EGC by itself without the need for a wire-type EGC.

I would prefer that upsizing as a general requirement be removed altogether; but that won’t happen. An effective exception would recognize alternate ground fault detection means instead - if a general problem were established in the first place.

Regardless of there being a general problem, why not propose ground fault detection as an exception, being effective as you purport it to be.

The general idea here is to eliminate EGC upsizing where it is more obviously not necessary

 

[hurk27]

I would prefer that upsizing as a general requirement be removed altogether; but that won’t happen. An effective exception would recognize alternate ground fault detection means instead - if a general problem were established in the first place.

But then we are adding to the very cost we are complaining about, with the GFP I would assume it would cost more then the cost of up sizing the EGC?

I would think a table that would be more lineal to the actual possible available fault current, more like a % ratio of up sizing the ungrounded to a % of required up sizing of the EGC:

lets say if you up size the ungrounded 300% then the EGC needs to be up sized 200% but that would have to have a limit on length because of larger ratio of resistance in the smaller wire. so maybe some kind of chart with length included. but we now need an EE to devise such a chart lol:grin:

 

[rbalex]

But then we are adding to the very cost we are complaining about, with the GFP I would assume it would cost more then the cost of up sizing the EGC?...

The Bob Alexander Criteria for a Professionally Engineered Design:

1. It is safe to the Unsuspecting Public
2. It is safe to the Qualified Worker
3. It does the Job Required
4. Within the context of and subject to each of the above priorities in order, it is Cost Effective
5. I get paid for it

(Item #4 doesn?t necessarily mean it will be inexpensive, because Item #3 can be a bitch with some clients and projects ? and occasionally so can satisfying some AHJs; but not nearly as often . Items #1 & 2 are not negotiable. Item #5 ? so far so good.)

My complaints about cost are always subordinate to my complaints about safety and/or effectiveness (part of #3) which upsizing isn't. Only when I say something is ineffective and someone defends it by saying, "It doesn't hurt," do I complain about cost. Oddly enough, I could decide upsizing an EGC is the most cost effective approach; I don?t believe I ever have because I usually end up providing some other form of ground fault protection, either required or highly recommended, anyway. Of course EGCs still get upsized because 250.122 is also currently part of #3.

All it takes for low voltage systems is a 1000A board and GFI becomes a requirement to begin with - regardless of circuit lengths or EGC sizes. Above 600V, detection just isn't a problem, but we tend to have some form of GF detection installed anyway. GFCI protected circuits don't care about voltage drop and, if AFCI's do become the general standard, they don't care either. Impedance grounded systems already limit GF current and are required to have auxiliary detecting means.

For any practical design, not one of the cases above would become any more effective by upsizing the EGC but they all require it anyway. Conversely, no one that defends the requirement can say failure to upsize will definitely cause a problem in a practical installation of the above cases; i.e. reasonable loads, circuit lengths and other circuit parameters. In fact, while it may seem intuitive, even voltage drop is a highly unlikely candidate for causing 120V GF problems for most practical circuits.

And if the EGCs are in parallel in multi-conductor cables it becomes a nightmare. The last bit of relief for conductors in parallel was stripped away in 2008.

It has long been recognized that parallel conductors are a special case for EGCs. While the load current may be equally divided among the CCCs, an individual conductor from an EGC formed from multiple conductors is very likely to be called upon to carry the entire ground fault current. Therefore, each of those conductors needs to be sized per Table 250-122.

However, before 2008, Section 250.122(F)(2) permitted an engineered solution with GF protection specifically designed for the circuit. It had a fatal flaw though, before 2005, Section 250.122(F)(2)(3) said, ?The ground-fault protection is listed for the purpose.? The kiss of death came in 2005 when it became, ?The ground-fault protection is listed for the purpose of protecting the equipment grounding conductor.? Since UL had no Standard for it, the 2008 response was to scrap the engineered solution and bye-bye Section 250-122(F)(2) entirely. The idea that it could have been worded any number of other ways and retain the concept just didn?t seem to occur to anyone: ?The ground-fault protection is listed?, ?The ground-fault protection is identified for the purpose of protecting the equipment grounding conductor.?

There is a ray of hope for 2011. If left unchanged, 250.122(F) will recognize using a single EGC outside the raceway or cable; however, there will need to be a coordinating change in 300.3(B) since running outside the cable or raceway doesn?t clearly meet any of the permissives in 300.3(B)1 through (B)(4). Of course, this only solves parallel conductor problems ? the other effective engineered solutions are still out in the cold.