Are you required to install ground indicating lights on an ungrounded system? Seems like it's optional and I can't seem to find anything in the code to indicate (no pun intended) that I need them.

George: Check out 250.21 in the 2005 NEC.
Creighton

But it doesn't say ground indicating lights.
just detectors..

Are you refering to 250.21(3)(D)?

Dnk..

George Little,
I would question if it is required of not because it does not make any sense not to.
If you draw a simple diagram of an ungrounded delta feeding 2 3p breaker2 that is supply power to motors in different part of a facility 100 feet apart. The motor winding in one of the motors fails and the 'A' phased contacts the frame of the motor which is grounded mechanically. Will there be a ground fault? No, because you have not provided a path for current to flow back to the source. Do you know that it has failed? Most likely not because what indications would you have?
Now, the other motor just happens to have a failing winding also except that it involves the 'B' phase. Now the 'A' and 'B' phase have contact with the ground through the metal frames of each motor. Does the current stop there? No. The current is going to seek a conductive path through the facility form one motor to the other, though anything that is conductive.
You never want that to happen so don't even question is GF indicating lights of any other approved system is optional. It doesn’t make sense not to.
BUT, GF indicating lights are useless unless there is a qualified person present to react to a GF indication. That's when he springs into action to identify and isolated the the grounded line be there is a catastrophic failure.
So, it doesn’t end with the lights but must follow through with personal who are qualified to recognize and respond to the event.
Dave

If I read this correctly, ground fault indicating _lights_ are not required, just some sort of ground fault detector.

IMHO ground fault indicating lights are the _simplest_ and least expensive form of ground fault detector that one could use.

But if a facility had a different sort of ground fault detector, then that would probably meet the requirements of the code.

Ungrounded systems are usually used for continuity of service. But what about an ungrounded system used for the purpose of reducing energy dissipated in a fault. If the breakers were arranged to detect and trip on a ground fault for an _ungrounded_ system, then IMHO this would meet the requirements of having a ground fault detector.

One could imagine a system that used coordinated ground fault detection so that the breaker nearest the fault would trip, at the same time never having large ground fault currents flowing.

-Jon

Jon- I guess I don't follow your scenario. If you have a system that is not grounded then a single fault to ground will not trip the overcurrant device. Now I can buy into the option of having either an audible or visual indication of a fault. Could be a display on a computer screen and it wouldn't have to be a light. Please elaborate on the overcurrent protection option you referenced.

It is also very important to never forget that it is with almost with all certainty that you can get enough current to flow through the ground form one point of contact to another to trip a breaker. The NEC requires that when a single ground rod is driven that there is 25ohms or less resistance. If one applies that to an ungrounded delta where 2 lines go to ground in different locations in a facility us 25 ohms as an example of the L-L resistance to determine if enougn current will flow to trip a breaker, not. The resistance may very likely be much higher.
Remember that GF indicators of any type are worthless unless somebody is notified and that person has the qualifications to locate the failure.

I understand that in an ungrounded system, a single ground fault would not cause an overcurrent condition that would trip a breaker.

But overcurrent is not the only thing that a breaker can be designed to detect. If a breaker were designed to detect and trip on a ground fault in an ungrounded system, then IMHO this would serve the purpose of ground fault detection.

I can think of several plausible ways to implement such breakers, though I don't know the precise details nor if anyone sells systems that would do this.

For example, something like a GFCI type breaker would detect the capacitive charging current flowing through the ground fault. A system using coordinated ground fault detecting breakers might be sufficient to detect and _trip_ on ground faults. Such a ground fault system would function with an impedance grounded electrical system, I don't know if it would function properly with an ungrounded system.

-Jon

But if it tripped on the first fault of an ungrounded system, the whole idea of an ungrounded system is destroyed. Not tripping on the first fault is one of the reasons you install an ungrounded system.

winnie,
If one elects to use a GFCI, what device did you have in mind? Ungrounded deltas are commonly 480v and in industrial facilities.
Would that device be list for application in such a system at 480v and would it work?
The purpose of and ungrounded delta is the continuity of service. As an example a line that produces glass. If it's shut down with no warning what happens? With an ungrounded system, as soon as there is a ground indication it affords the opportunity to locate and isolate the failure or shut the process down in an orderly fashion.
Dave

Ground lights are not required, but they are nice to have.

I once ran into a situation where we had 460V between 2 pieces of eqpt that were about 2' apart. Not a pleasant thing to come into contact with.

I agree with you Dave- I worked with a customer who produced ice cream and they could not afford to have the ice cream stop moving through the pipe line that was refrigerated with amonia because the pipeline and the manifold would freeze solid. They could tolerate a single fault and then empty the line and clear the fault under scheduled maintenance. Jon's theory while is not popular could be used to detect faults and not shut down the system.

I'm kinda looking at the information in 250.21. In 250.21(3)d, it talks about ground detectors in the control circuit when the continuity of the control circuit is required. When you look at item number (5) there is no indication the detectiors are required. I agree with the comment that it makes sense to have them.

George,
I seems like I discussed this on another forum some time back and suggested that, if this is not already clearly included in the current code it should be added.
When it comes down to it an ungrounded delta can be so dangerous I can't believe that the NEC hasn't addressed the issue. Personally I feel strongly that some type of ground detection should be required and that a qualified person be available to address a ground detection issue. I don't think it should be screwed around with.
The NEC is specific on so many things it is interesting why they appear to be avoiding this application. Maybe it's simply a political issue.

250.21(3) is only for control circuits like those found on stamping presses.

In 2005, the last sentence of 250.21 specifically requires ground detectors on most ungrounded systems.

In previous code cycles this was only an FPN.

Guys,

I have to admit to stepping well outside of my field of experience here. Here is what I said on my first post on the topic:

I was suggesting that if, _as a design choice_, someone were to use an ungrounded system for a reason other than continuity of service, then the requirements for ground fault detection might be met in a fashion other than indicator lights. I was offering this as a situation where ground fault detection was provided, but ground fault indicating lights were not used, and this was intended simply as a proposed answer to George's original question. This was not offered up as the next latest and greatest system design

The design choice is the desire to detect and remove power from ground faults without having hundreds or thousands of amps flowing through the wiring and equipment grounding system in the event of a ground fault.

I do not know if listed devices are available that would function in this fashion. I do not know if anyone does this in NEC compliant permanent installations with this design goal in mind.

I know that normal GFCI devices are only used at 120V. So a standard GFCI breaker is out. But ground fault detection equipment certainly exists for 480V systems, people complain about GFP on this board whenever they encounter a ground fault system that is not properly coordinated. So 480V two pole and three pole ground fault detecting breakers are certainly _conceivable_ products even if you can't get them off the shelf.

-Jon

JBD- First of all thanks for the '05 reference. We here in Michigan are on the '02 so I didn't look in the '05. I personally think they should be installed but as an inspector I can only recommend and can't require them. You guys in Wiconsin are ahead of us I guess. Maybe Ed Lawry had something to say about it

George, yes I am sure Ed had lots to say about it.

Jon, Class A GFCIs (people protection) are only available up to 6mA and 240V. There are some manufacturers that make non-Class A devices for 10mA and greater currents up to 480V. There are breakers with internal GF protection but they usually start at 2A and go up from there. Most mA rated devices would not be practicle due to the normal system charging current.

One problem with your method, is there is very little ground fault current flowing as a result of the first ground. In fact there might even be less current than normal based on the amount of coupling capacitance that is "shorted out".

We use floating control circuits on our emergency fans. You want the fans to run to destruction in the event of a fire. One wire grounding won't take us out of service but you lose your safety margin if you don't detect it. I install AC ground detectors which drive a Ground OK "GOK" status relay. I also install a Power OK "POK" status relay. These are both N.O.H.C. relays. The "POK" relay is driven via control power through the local "On,Hand,Remote" switch and 2 phase loss relays from the 480V side. This gives me the ability to use POK & GOK together as a very informative dual status point where:

0,0 = Power Fail
0,1 = Hand switch off of remote or phase missing.
1,0 = Ground Fault
1,1 = Normal

It's all gravy when your ground detectors help you get other info that you needed anyway.

The down side if any, is the infrequent question, "Why am I only reading 60 VAC to ground?" I tell them that the ground detector is establishing a high impedance, half point reference, and suggest a better location for a meter probe.
Joe