Why is circuit breaker ground fault max delay = 0.5s? 2

[JensenDrive]

Why do most low voltage breaker ground overcurrent units (at least the ones I have dealt with) seem to have a maximimum delay setting of 0.5 seconds? The NEC 230.95.A says delay can be as long as 1 second @3000A. What piece of the story am I missing?

 

[unclebob]

Not the best answer, but allowing too much time to clear the fault may result in a ground fault developping in a 3-phase fault.

 

[alehman]

I don't know either. Some stand-alone ground fault relays will go up to 1 sec.

 

[Zogzog]

Thats 0.5S at a much higher current than the GFPU value, the actual delay time will also depend if I2t is on or off.

 

[jahrtrix]

Jensen,
From memory there seems to be a UL or ANSI requirement of
a 3 cycle limit. That plus the actual quenching of the arc
could be the total 1 second limit?

 

[JJHorak]

Does anyone know if there is an international standard that calls for 0.5s max delay for ground faults?

 

[jahrtrix]

I made a mistake that should have been 30 cycles for 0.5
seconds.

 

[burnt2x]

JensenDrive,
Ground fault protection is basically provided to prevent electrical damage to equipment. You have to know what is the maximum permissible energy that the electrical equipment can take before it is damaged (specially during arcing ground faults). We should limit the time allowed for the damaging fault current to flow in circuit, i.e. trip activation be as short as possible before damage occurs!
Here's my take on this matter:
1) Permissible electrical energy is expressed usually in "kW-cycles". I often us the value of Emax= 20,000 kW-cycles (please check if this energy level is okay with your equipment protected)
2.) Systems operating 150V or more wrt groung will have arcing voltages (during SLG) at around 70V-150V. Let's take 100V as the arcing voltage level (ave. of the min and max). Dividing the permissible energy 20,000 kW-cycles with the arcing voltage level 100V will give me 200,000 amp-cycles!
3.) If you know what your maximum phase-to-phase fault current at the CB location, you can deduce the maximum available ground fault say 1/3 of your L-L fault current; For example the available line-line fault on a LV CB location is Ip-p = 25kA, Islg = 8,333A. Dividing the 200,000 amp-cycles permissible by 8,333A will give us the allowable trip time delay of = 24 cycles!

That is less than 0.5 seconds(60 Hz), 0.4 seconds to be exact.

Your specific application may not be the same as mine, but the idea is clear. If you think CB ground fault setting trip delay is too long, you may want additional ground fault relaying that could offer faster shunt trips to your breaker.

Hope this helps!

 

[JensenDrive]

Thanks for the answers thus far.

Jartrix:
I suspect you are on the track of the reason, and it is what I was hoping to hear some details on. I have looked around a bit for some standard that might mention the 0.5s limit, but it is a needle in a haystack. Any ideas?

ZogZog:
Sure, time delays will be long for low fault currents if I^2t is in use and the delay starts its slope up from 0.5s @ 3000A. But still that does not quite explain the 0.5s limit.

UncleBob and Burnt2x:
I agree that there are good engineering reasons to limit the delay. I agree that there is an increased risk of ground faults evolving, esp. if you have longer delays. I like Burt2x's calculation on limiting arc energy, though in more common practice I think the standards on arc flash analysis are used to do what you are suggesting, though arc flash analysis considers only 3 ph faults.
While it can be bad to let a fault hang on a long time, it just does not seem to be a satisfying answer for the virtually universal 0.5s limit, yet.

 

[unclebob]

Having up to 1 sec of delay may be useful if you have to coordinate several devices of different voltages in series.

MV Ground fault could be set between 0.5 and 1.0 sec to give enough time to LV devices to trip first.

Exemple:

25 kV Ground Fault set at 0.7 sec

600 V main breaker GF set at 0.5 sec
600 V feeder breaker GF set at 0.3 sec
...

 

[jahrtrix]

Jensen, The best I can do is ANSI 37.17 which part of it sets calibration points for these breakers. NEC may say
1 sec, but the manufacturer who uses ANSI or IEC may have to
use their standards to pass qualification tests. Like many
here I don't have the financial means to be able have these
manuals or standards on hand. Good luck.

 

[dpc]

I suspect there are UL requirements for max trip delay for breakers sold in the US.