ECN Forum
Posted By: renosteinke How long can a circcuit be? - 11/01/06 02:57 PM
Another thread raised the issue of breakers not tripping. It was suggested that impedance was the cause.

So- how long can a circuit bem before impedance becomes an issue in breaker performance?

The answer seems to be available in the GEMI study. Free software, that lets you compare different solutions, is available at:
http://www.steelconduit.org/gemi.htm
Posted By: SteveFehr Re: How long can a circcuit be? - 11/01/06 04:33 PM
There is no length restriction; the breakers are designed to protect the cable, which fails thermally. The instant trip is there to kill the circuit immediately if a short circuit is detected (and may not ever trip if the cable impedance is too high) but thermal trip should always work regardless of the fault circuit and protect the cables. If the breaker doesn't trip (and isn't malfunctioning), it's because it didn't see levels of fault current that would pose a danger to the building cabling, merely elevated fault current. It's not smart enough to know in-rush from a cheap compressor from a kid with a fork in the socket; AFCIs aren't even that smart. It just looks at current and trips if there's too much for too long.

On better breakers, all the trip settings are calibrated and manually configrable, so it's just a matter of taking system impedance into account when dialing in the trip settings.

[This message has been edited by SteveFehr (edited 11-01-2006).]

[This message has been edited by SteveFehr (edited 11-01-2006).]
Posted By: JBD Re: How long can a circcuit be? - 11/01/06 06:28 PM
Steve brings up an important point. Branch overcurrent protective devices (fuses and breakers) are designed to protect the branch circuit wiring not the wiring after the receptacle.

This is a paraphrased quote from a major manufacturer concerning BOLTED short circuits:
If a low voltage phase conductor is properly sized per NEC 240, it is not possible to damage the phase conductor during short-circuits below the AIC rating of the breaker protecting the phase conductor. UL tests to verify the short-circuit rating of a circuit breaker are performed considering 75C cable. The corresponding phase conductor is sized according to the NEC and must pass the fault tests without compromising its integrity."
Posted By: RODALCO Re: How long can a circcuit be? - 11/01/06 07:33 PM
Cable length is an interesting point raised here.
If a circuit is very long and the maximum load on the circuit equals the circuit braker value, the voltage at the outlet may be half the voltage and the other half of the voltage can be dissipated as heat in the cable or connection.
When motors are connected, lower voltage, less torque is available, motors may not be able to start properly and draw excessive start current adding to the problem.

Under short circuit condition the cable impedance may be high enough not to trip the braker immidiately and excessive heating in the cable may occur and possible cause of a fire.

Certainly cable length is an important fact to bear in mind.

My $0.02 worth on the topic

Regards

Raymond ( RODALCO )
Posted By: JBD Re: How long can a circcuit be? - 11/01/06 10:01 PM
Breakers are designed and tested to protect the branch circuit conductors under all conditions. This is one of the reasons that the UL test for circuit breakers (and fuses mounted in holders) includes 4' of conductor. This way we know the conductor is protected up to the AIC rating of the device and to the thermal rating of the termination. If there is enough impedance to limit the fault current below that of the short circuit portion of the protective device then its thermal portion will protect the conductor.
Posted By: wa2ise Re: How long can a circcuit be? - 11/01/06 10:18 PM
If a cable is long enough to have enough impedance to cause its breaker not to clear a fault, is it also too long (for the gauge of wire used) for acceptable amounts of voltage drop when it's used normally? Say #14 loaded by a 13A load at the far end and the voltage drop seen by the load excessive? If you upsize the wire to get acceptable voltage drop, would this excessive impedance issue go away? Or are there issues involved if ferrious conduit or such us correctly used?
Posted By: SteveFehr Re: How long can a circcuit be? - 11/01/06 11:14 PM
Fault current is far in excess of normal load current, and that exacerbates voltage drop. A cable that sees 5V drop at 13A will see 50V drop at 130A. This is unrelated to normal conditions, though, where the voltage is within generally acceptable limits.
Posted By: resqcapt19 Re: How long can a circcuit be? - 11/01/06 11:51 PM
JBD,
Quote
If a low voltage phase conductor is properly sized per NEC 240, it is not possible to damage the phase conductor during short-circuits below the AIC rating of the breaker protecting the phase conductor.
The 1/4 cycle withstand for #14 is 3370 amps. That is below the AIC rating of all breakers.
Don
Posted By: gfretwell Re: How long can a circcuit be? - 11/02/06 01:00 AM
I think it is around 1300 feet that you could wirenut the 2 conductors of a 14ga circuit together and the breaker would never trip. You could still have good performance out there if you were just running a 100w light.
I don't know why I know this but it was either a "puzzler" or a real question on one of these BBs.
Posted By: frank Re: How long can a circcuit be? - 11/02/06 01:03 AM
Can be as long as the voltage drop allows it to be is what i always thought but i guess thats wrong.


[This message has been edited by frank (edited 11-01-2006).]
Posted By: JBD Re: How long can a circcuit be? - 11/02/06 02:24 AM
Don,

Quote
The 1/4 cycle withstand for #14 is 3370 amps. That is below the AIC rating of all breakers.

It may have to do with the dynamic impedance of the protective device.

Regardless, during a test UL requires 10" (I was mistaken in my previous statment of 4') of rated conductor to be subjected to the same fault current as the breaker. This is to prove that the breaker protects the conductor.

This is a quote from a George Gregory article:
Quote
Every time a circuit breaker is tested under consistent circumstances, it demonstrates that the circuit breaker design will protect from the following:
• lateral forces that tend to break circuit breaker cases;
• in-line forces that tend to pull wires out of connectors;
• heating that tends to melt insulation.
Posted By: resqcapt19 Re: How long can a circcuit be? - 11/02/06 02:32 AM
JBD,
Quote
Regardless, during a test UL requires 10" (I was mistaken in my previous statment of 4') of rated conductor to be subjected to the same fault current as the breaker. This is to prove that the breaker protects the conductor.
The AIC testing is just to prove that the breaker itself can withstand the fault current. In many cases the conductor will be damaged if the fault current is any where near the AIC rating of the breaker. Now if you are talking about the overcurrent trip points of the breaker, then the conductor, if properly sized, would be protected. As I recall UL Standard 489 for breakers only calls for trip testing at 135% and 600% of the breaker rating. There is no test or requirement that the breaker have a magnetic or instantaneous trip function.
Don
Don
Posted By: renosteinke Re: How long can a circcuit be? - 11/02/06 02:35 AM
I beg to differ with some sentiments expressed so far.

Just because the code sets no rule does not mean something is a good idea. Npw, if you are saying circuit length is a design issue- well, I couldn't agree more.

As for protecting the conductros:

We base our voltage drop calculations on the load imposed. A small load will result in little voltage drop.
Probably the biggest weakness in voltage drop calculations is that they are based upon the RESISTANCE of the conductor. Resistance is fine- when you're using DC. We're not- we use AC. That means IMPEDANCE is an issue.

A real world example I once saw: A 350 ft. circuit, #12 wire in 1/2" EMT, had a short between the wire and the pipe. Rather than trip the breaker, the fault continued.... in the process, it cut an extermely neat slot in the side of the EMT. After 'plasma cutting' for about 1/2 inch, it appears the wire bent too far from the pipe for the arc to continue, and the fault ended. Now, this exposed wire end was still "hot"; but, as the wire had been severed by the fault, parts of the circuit downstream from it no longer had power.

Using the GEMI program, one can see that, at the length involved, one could no longer rely upon the breaker to function under a 40 amp fault.

Now, for those of you who know how to weld, it is no surprise the 120 volts and 40 amps is plenty of electricity to work with.

I submit that an arc that will burn through steel is certainly a safety hazard. After a certain length of run, we need to consider whether the breaker will operate when needed.
Posted By: JBD Re: How long can a circcuit be? - 11/02/06 04:40 AM
Don,
It is a fact that to pass a UL test for an AIC rating the breaker must protect 10" of rated conductor as well as itself.

The quote I provide is:
from
IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS, VOL. 35, NO. 1, JANUARY/FEBRUARY 1999 page 135

titled
Short-Circuit Ratings and Application Guidelines for Molded-Case Circuit Breakers

written by:
William M. Hall, Senior Member, IEEE, and George D. Gregory, Senior Member, IEEE
Posted By: Trumpy Re: How long can a circcuit be? - 11/02/06 07:39 AM
As far as New Zealand goes (and Ray might remember these) there were tables in the ECP (Electrical Code of Practice) 28, for the installation of cables, with a maximum length given for 5% voltage drop in a given run.
We have a test here in NZ, called an Earth-Loop Impedance test, that tests not only the connections to your panel, but back to the transformer.
It gives you a reading of the total impedance of the total loop, from that, you can measure your supply voltage and find your fault current.
Fault currents as mentioned above, are in the order of kilo-amps.
I did a test the other day at work and found that a simple fault will cause 250,000A, flow through cords and the like, if a breaker does not trip.
The Sub-station(66kV) is just over the fence, with a tranny just out the door (11kV/400V).

{Message edited for 2 typo's}


[This message has been edited by Trumpy (edited 11-02-2006).]
Posted By: RODALCO Re: How long can a circcuit be? - 11/02/06 10:34 AM
Mike, are you sure about the 250000 Amps??
That is looks more like the 250MVA level of the 11 kV bus!!
If your supply transformer is 100 or 200 kVA you get in the 10's of kA's if the supply cable is heavy gauge and short.

You can't even protect that faultlevel with a 120 kA HRC fuse.
Posted By: SteveFehr Re: How long can a circcuit be? - 11/02/06 12:05 PM
What size is that 400V transformer? If it's 25kVA-50kVA, it probably isn't capable of more than 20kAIC or thereabouts. The impedances of the wires and cable and extension cords will drastically reduce the peak fault current from the maximum the transformer is theoretically capable of.

- Start with the transformer impedance. Then, add up all the line resistances between the transformer and your fault point using NEC tables. You'll certainly need to take the complex impedance into account, especially if your fault current is approaching the limit of your breakers.
Posted By: resqcapt19 Re: How long can a circcuit be? - 11/02/06 12:27 PM
JBD,
I just don't understand how the conductor is protected by a breaker with a high AIC unless the breaker is of the current limiting type. As I said before the lowest breaker AIC rating of 5000 amps will cause damage to a #14 conductor unless the breaker opens in less than 1/2 cycle (this is beased on the Insulated Cable Engineers
Association (ICEA) damage current level). Do standard breakers open that quickly? Not trying to be a pain, but just trying to apply what I think that I know to this case.
Don
Posted By: SteveFehr Re: How long can a circcuit be? - 11/02/06 04:08 PM
resqcapt19, I don't want to say that it's impossible for there to be conditions that could possibly put 5000A into a #14 cable, but even if you hooked that #14 cable to an ideal 120V 0-impedance source, if the romex exceeded 4', resistance in that cable is too high to allow current levels to reach 5000'. In reality, transformer and feeder impedance drop current levels WELL below that point.



[This message has been edited by SteveFehr (edited 11-02-2006).]
Posted By: JBD Re: How long can a circcuit be? - 11/02/06 04:28 PM
Don,
It primarily has to do with system impedance. All OCPDs exhibit some current limiting effect as soon as they begin to open effectively changing the bolted fault into an arcing fault. And, yes small breakers can and do operate in less than 1/2 cycle.

Here is more information from the article:
Quote
When a small MCCB is placed in a system with a high available fault current, the circuit breaker with any connecting impedance will be large in comparison with the system impedance and will result in a significant reduction of available fault current on the load side of the circuit breaker. This limitation is not the case with very large circuit breakers, which will permit virtually the full available fault current to flow through them until it is interrupted.

Quote
XV. CONCLUSION
1) The rationale for the short-circuit test with “leads” is strong, in that it demonstrates not only protection of the insulated wire, but also demonstrates the ability of the circuit breaker to retain the wires without damage to the wires or to itself during the heavy forces of a short circuit.

[This message has been edited by JBD (edited 11-02-2006).]
Posted By: iwire Re: How long can a circcuit be? - 11/02/06 04:39 PM
Quote
And, yes small breakers can and do operate in less than 1/2 cycle.

Isn't that assuming a low impedance circuit.

In the field I see bolted faults on 20 amp breakers that take seconds to open the breaker as the circuit length may may be 400' out and back on 12 AWG.
Posted By: iwire Re: How long can a circcuit be? - 11/02/06 04:44 PM
Quote
Steve brings up an important point. Branch over current protective devices (fuses and breakers) are designed to protect the branch circuit wiring not the wiring after the receptacle.

They may not be designed with that in mind but IMO they are used for that purpose.

Consider the rules in 422.11 that prevent me from suppling an appliance with a circuit that has a much greater capacity than the appliance requires.

For example I can not supply a water heater that has a 24 amp current draw with a 100 amp branch circuit..

It is my view that this rule is in place so that the wiring inside the appliance is also protected.

Or consider the rules in 210.23 that prohibit me from supplying a receptacle (other than 15s) with a circuit of higher rating.

[This message has been edited by iwire (edited 11-02-2006).]
Posted By: JBD Re: How long can a circcuit be? - 11/02/06 08:22 PM
Don,

Quote
Isn't that assuming a low impedance circuit.

In the field I see bolted faults on 20 amp breakers that take seconds to open the breaker as the circuit length may may be 400' out and back on 12 AWG.

Of course, if there is not enough fault current to open a breaker instantaneously then it could take seconds for it to operate thermally. But this low of current would be below the damage point of the conductor. And, the thermal curve of a UL489 breaker should mimic that of the conductor when it is at the reference temperature of the breaker (typically 25C or 40C).

The OCPD does not have to clear the fault in 1/2 cycle to protect the conductor it only has to reduce the current below the damage point.

Bob,
Quote
It is my view that this rule is in place so that the wiring inside the appliance is also protected

I don't believe so. For example a water heater might require a 40A circuit but after the incoming terminal block only 20A conductor might be run to each element.

[This message has been edited by JBD (edited 11-02-2006).]
Posted By: iwire Re: How long can a circcuit be? - 11/02/06 08:32 PM
Quote
I don't believe so. For example a water heater might require a 40A circuit but after the incoming terminal block only 20A conductor might be run to each element.

Than can you explain why the NEC (and UL) will not allow supplying that WH with a 100 amp feeder?

Or why HVAC equipment has a maximum breaker size listed on it?

Just because the 'protection' does not mirror 240 and 310.16 does not mean that the breaker does not protect the appliance wiring from bursting into flames before the OCPD opens.

In your water heater example what would you say does protect the conductors from short circuits if not the branch circuit OCPD?

I am not trying to bust your chops, we just see this differently.

Bob.
Posted By: JBD Re: How long can a circcuit be? - 11/02/06 10:21 PM
Bob,

What I am saying is that a branch breaker is designed and tested to protect branch circuit conductors. The fact that they can also protect some end use equipment is secondary.

I can understand why the manufacturer of end use quipment would design and test their product to use the branch breaker, then they would not need to incur the cost of internal protection.
Posted By: iwire Re: How long can a circcuit be? - 11/02/06 10:42 PM
JBD doesn't the NEC actually require the use of a OCPD for the short circuit and ground fault protection of motors?

As shown in 430.51.

IMO both the NEC and the NRTLs expect the branch circuit OCPD to play a part in the protection of many types of utilization equipment.

I would not call it 'secondary' I would say the tasks are equal.

JMO, Bob
Posted By: gfretwell Re: How long can a circcuit be? - 11/03/06 12:37 AM
In small conductors 240.4(D) provides overload protection greater than necessary for simple overcurrent protection or short circuit protection.
Posted By: JBD Re: How long can a circcuit be? - 11/03/06 02:32 AM
Bob,

I agree the tasks may be equal, but the design of the breaker is to protect the branch circuit short circuits AND overloads. Supplementary breakers are designed to protect loads. The fact that some loads get to use the features of the branch breaker (i.e. overload for some appliances and short circuit for motors) is a function of the load design not that of the breaker.
Posted By: iwire Re: How long can a circcuit be? - 11/03/06 11:01 AM
JBD I am interested as I do not really understand what your saying.

The breaker does not know where the premise wiring ends and utilization equipment starts.

All the breaker 'knows' is the current passing through it.

It seems to me that saying the design of the breaker is to protect the branch circuit short circuits AND overloads is to limited.

A breaker is designed to open under predetermined conditions of short circuits and overloads. It does not mater if that overload is in a feeder, branch circuit, or utilization equipment.

I could also point to 240.5 that rightly or not extends the job of the OCPD beyond the receptacles and outlets.

Not trying to be a pain I just see a lot in the NEC that requires a branch circuit breaker to protect the equipment beyond outlet as well as the branch circuit itself.



[This message has been edited by iwire (edited 11-03-2006).]
Posted By: JBD Re: How long can a circcuit be? - 11/03/06 03:09 PM
Bob,

The point I am trying to make is not about the requirements of the NEC that the branch OCPD be used to protect certain pieces of equipmentand appliances, which it clearly does.

My point is dealing strictly with the design and testing of branch OCPDs. Their primary purpose is to protect conductors. Their thermal curve is designed to be below the thermal damage curve of a properly rated conductor. Their instantaneous region is tested to protect the conductor from the forces of a short circuit. These design criteria are why it is difficult (and why the NEC includes special allowances) to chose a standard thermal magnetic breaker that will coordinate with motor and transformer inrush currents. A properly operating listed branch OCPD will protect a properly rated and installed conductor during a low impedance short circuit fault.
Posted By: Trumpy Re: How long can a circcuit be? - 11/03/06 07:41 PM
Ray'
Quote
Mike, are you sure about the 250000 Amps??
Oops,
That should have read 15kA, not sure how that slipped under the radar.
Just got a new laptop and the keyboard is VERY different to my normal computer keyboard.
Fault currents are very high at work, due to the size of the mains installed and the proximity of transformers.
The PoCo used more tranny's than usual with this area of town, to maintain the voltage at a respectable level, given the loads involved.
Posted By: resqcapt19 Re: How long can a circcuit be? - 11/04/06 12:38 AM
Steve,
Quote
I don't want to say that it's impossible for there to be conditions that could possibly put 5000A into a #14 cable, but even if you hooked that #14 cable to an ideal 120V 0-impedance source, if the romex exceeded 4', resistance in that cable is too high to allow current levels to reach 5000'.
My numbers say that 8' (4' of NM) of 14 will permit 4777 amps to flow if the system can supply it. Yes that is less than 5000 amps, but the 1/2 cycle damage point for #14 is 2384 amps.
Quote
In reality, transformer and feeder impedance drop current levels WELL below that point.
So there is never a need for a small breaker with an AIC greater than 5,000 amps?
Don

[This message has been edited by resqcapt19 (edited 11-03-2006).]
Posted By: OreElect Re: How long can a circcuit be? - 11/05/06 10:13 PM
Just for fun, try doing a voltage drop calculation on a 150 foot run of 14 with a 12 amp load.
Posted By: SteveFehr Re: How long can a circcuit be? - 11/07/06 02:26 PM
Quote
So there is never a need for a small breaker with an AIC greater than 5,000 amps?
Don
Well, that depends greatly on the distribution system and the size of the transformer. In this example, yeah, 5KAIC would be all that's required. But we could easily come up with another example (say, a 15A breaker on a 400A panel bus-fed from a 3000A switchboard/1MVA transformer) feeding a receptacle 6" from the panel. In THIS case, there is potential to exceed 5KAIC and damage the cable.

But in a typical residence, shorting out the receptacle? Not likely at all. The engineer really has to calculate all this and specify the KAIC rating of the breakers if it's expected to be an issue.

[This message has been edited by SteveFehr (edited 11-07-2006).]
© ECN Electrical Forums