Is it OK to drive a ground rod at a machine if switch gear & buss plug have no ground

[Smart $]

Unlike a metal plane, the ground resistance has a resistance area beyond which it has insignificant resistance. So voltage drop across one ground resistance has no effect on the other if it lies beyond the resistance area of the latter.

Sounds like a lot of double speak on my end... :blink:

(Do not construe my statement as solicitation for elaboration )

 

[Sahib]

That is correct, and kind of has been what I have been trying to say all along, the equipment grounding conductor is more important than some isolated grounding electrode ever will be.

It is not a matter of which one is important: EGC or an isolated ground. What is important is life safety: in some condition as discussed above the isolated ground may provide better protection together with EGC.

Current on a EGC will raise voltage from that EGC to ground, I agree. If that EGC is in good condition it would need to be pretty significant current to have 60 volt drop (on a 120 volt circuit). If that kind of current is flowing to produce that kind of voltage drop, hopefully it is well over the trip setting of the overcurrent device and all we are waiting for is the time needed according to the trip curve to open the overcurrent device.

Take a 400A, 120V circuit. 750Kcmil conductor.The EGC size is #3. Its resistance per 1000ft is 0.3 ohm. So for 60V drop,current is 60/0.3=200A only.

That said a faster operating overcurrent device is probably going to limit touch voltages better than a ground rod will, either way still requires an EGC in good condition.

Agreed.

 

[kwired]

It is not a matter of which one is important: EGC or an isolated ground. What is important is life safety: in some condition as discussed above the isolated ground may provide better protection together with EGC.

Take a 400A, 120V circuit. 750Kcmil conductor.The EGC size is #3. Its resistance per 1000ft is 0.3 ohm. So for 60V drop,current is 60/0.3=200A only.

Agreed.

OK I'll concede the voltage to ground in the vicinity will be lowered by this ground rod, but will not subscribe to the fact that many will think they are getting more protection than they really are from a ground rod driven at such equipment. The voltage zones around that rod are not very large especially at only 120 volts supply. Increase supply to 277 volts to ground and your chances of having a high resistance fault that doesn't rapidly decrease in resistance seem to go down pretty dramatically. In many cases there is alternate grounded paths that also lower resistance of fault paths so we have more than just the EGC to work with and that makes the ground rod fairly pointless if there are other reliable paths for fault current to travel.

 

[jaggedben]

FWIW, this video from Mike Holt partially addresses the issue. It's long though.

http://www.youtube.com/watch?v=YuDqXFvRv94

 

[dereckbc]

Take a 400A, 120V circuit. 750Kcmil conductor.The EGC size is #3. Its resistance per 1000ft is 0.3 ohm. So for 60V drop,current is 60/0.3=200A only.

Nothing in your example is real.

1. No where I have ever seen or witnessed a 400 amp 120 volt circuit 1000 feet long. Anything requiring that much power at that distance is going to be 3-phase operating at much higher voltage.
2. 1000 feet of #3 AWG is .2 Ohms not the .3 you mention.
3. If the supply is 1000 feet of 750 the total combined impedance is of the 750 and #3 is .217 Ohms impedance resulting in 120 V / .217 ohms = 552 amps. Even if you could drive a rod to get 10 ohm's impedance would do nothing.
4. At the point of contact the 750 MCM impedance is so much lower than 3 AWG the voltage will be 552 amps x .2 ohms = 552 amps x .2 ohms = 110 volts at the point of the fault.
5. A 1000 feet circuit using 750 MCM with 400 amp load current would have a voltage drop of 14 volts or 12 % voltage drop.
6. Not compliant to any known code or practice. Minimum EGC for this make believe circuit would require a 250 MCM which would create a 2400 amp fault current which would operate the OCPD instantly.

Nough said.

 

[don_resqcapt19]

The general theory of local ground rods/rings is the same as equipotential bonding. During a line-to-ground fault it raises the voltage of nearby earth to the voltage potential on the grounded side of the fault.
To use the voltage divider analogy mentioned earlier, grounded metal can be at 60V (one-half of a 120V circuit) to electrical ground at the service, but if the nearby earth is raised to that same 60V potential, shock potential is reduced significantly.

But the area around the grounding electrode where the voltage is raised is so small that there really isn't any improvement in safety.

There is only one thing that can make an under 600 volt system safe under ground fault conditions and that is an effective fault clearing path that will let the OCPD quickly clear the fault. The hazard will exist until the fault is cleared.

 

[dereckbc]

But the area around the grounding electrode where the voltage is raised is so small that there really isn't any improvement in safety.

There is only one thing that can make an under 600 volt system safe under ground fault conditions and that is an effective fault clearing path that will let the OCPD quickly clear the fault. The hazard will exist until the fault is cleared.

Amend, that is the point being missed by a few. From a deignPOV, one wants to induce roughly 6 x OCPD fault rating to trip the device instantaneously.

Life safety keeps coming up but is a secondary concern because there is no amount of impedance that is going to keep the voltage low enough to prevent a shock hazard to personnel. The point is to remove the voltage as fast as possible, not limit the voltage.

 

[Smart $]

But the area around the grounding electrode where the voltage is raised is so small that there really isn't any improvement in safety.

I agree. As I said, it is the theory. I did not say it was fact. However, one major discernible issue is we are talking about a single ground rod here. If the electrode were of a highly conductive lattice design (mesh; grid), the elevated voltage would not be isolated to that small area around the rod.

 

[Smart $]

...Life safety keeps coming up but is a secondary concern because there is no amount of impedance that is going to keep the voltage low enough to prevent a shock hazard to personnel. The point is to remove the voltage as fast as possible, not limit the voltage.

I don't agree with this assessment. The point is to remove fast AND minimize as much as feasibly and/or reasonably practicable. This premise is exhibited by equipotential bonding of pools and such.

 

[kwired]

But the area around the grounding electrode where the voltage is raised is so small that there really isn't any improvement in safety.

There is only one thing that can make an under 600 volt system safe under ground fault conditions and that is an effective fault clearing path that will let the OCPD quickly clear the fault. The hazard will exist until the fault is cleared.

I agree. As I said, it is the theory. I did not say it was fact. However, one major discernible issue is we are talking about a single ground rod here. If the electrode were of a highly conductive lattice design (mesh; grid), the elevated voltage would not be isolated to that small area around the rod.

Thanks guys, both those posts kind of sums up what I have been trying to say.

 

[dereckbc]

I don't agree with this assessment. The point is to remove fast AND minimize as much as feasibly and/or reasonably practicable. This premise is exhibited by equipotential bonding of pools and such.

Sounds to me like you agreed. Only thing you can control is the speed with the OCPD type/size and EGC properly. You have no control of potential difference and have to take whatever you get. A swimming pool is a different monster and changes the conversation to step potential differences to minimize current running through ones body.

 

[don_resqcapt19]

I agree. As I said, it is the theory. I did not say it was fact. However, one major discernible issue is we are talking about a single ground rod here. If the electrode were of a highly conductive lattice design (mesh; grid), the elevated voltage would not be isolated to that small area around the rod.

Exactly the intent of some more specialized installations such as pool equipotential bonding or a utility substation bonding grid under the gravel in the substation. The intent of both is to raise the area voltage to prevent step and touch potentials.

 

[don_resqcapt19]

If you have a ground fault, is it safe to assume that the total voltage drop of the fault circuit will be equal to the system line to neutral voltage and that in almost all cases the voltage to "remote earth" at the point of fault will be greater than 1/2 of the line to neutral voltage?

 

[Smart $]

If you have a ground fault, is it safe to assume that the total voltage drop of the fault circuit will be equal to the system line to neutral voltage and that in almost all cases the voltage to "remote earth" at the point of fault will be greater than 1/2 of the line to neutral voltage?

The voltage drop of the entire circuit has to be the line to [grounded] neutral voltage at the source, don't it?

To say it is half at the point of fault is a reasonable assumption for discussing the topic. Too many variables to nail it down a specific value, IMO. Chances are, somebody somewhere at some point in time has done some experiments and recorded empirical data. I just don't happen to have any of that data in my hip pocket. :blink:

 

[don_resqcapt19]

...
To say it is half at the point of fault is a reasonable assumption for discussing the topic. Too many variables to nail it down a specific value, IMO. Chances are, somebody somewhere at some point in time has done some experiments and recorded empirical data. I just don't happen to have any of that data in my hip pocket. :blink:

I was just thinking that the impedance of the ungrounded conductor would likely be less than that of the EGC, but maybe the metallic parallel paths, like a metal raceway would make the impedance of the fault return path lower than that of the ungrounded conductor.

I guess it is probably safe to say that in many cases the voltage to remote earth at the point of the ground fault would be a shock hazard until the OCPD clears the fault.

 

[Smart $]

I was just thinking that the impedance of the ungrounded conductor would likely be less than that of the EGC, but maybe the metallic parallel paths, like a metal raceway would make the impedance of the fault return path lower than that of the ungrounded conductor.

I guess it is probably safe to say that in many cases the voltage to remote earth at the point of the ground fault would be a shock hazard until the OCPD clears the fault.

Exactly. Multiple pathways for fault current is one of the variables I mentioned.

FWIW, I'd avoid using safe as you have, given the context of the statement...

 

[dereckbc]

Exactly. Multiple pathways for fault current is one of the variables I mentioned.

Correct but a voltage divider none the less. Keep in mind that only really applies to commercial and industrial applications where metal raceways are used. In a stick built house using plastic boxes, all circuits are IGR, so what you are left with a 1/2 voltage divider and fault voltage will be approx 60 volts. Will except in Chicago where I understand even residential electric has to be in metallic raceway (emt)?

FWIW, I'd avoid using safe as you have, given the context of the statement...

See I told you we agreed.

 

[Smart $]

Correct but a voltage divider none the less. Keep in mind that only really applies to commercial and industrial applications where metal raceways are used. In a stick built house using plastic boxes, all circuits are IGR, so what you are left with a 1/2 voltage divider and fault voltage will be approx 60 volts...

Agree with voltage divider description... but in stick-built residential, the multiple pathway variable must also include the higher probability of resistive faults.

 

[dereckbc]

Agree with voltage divider description... but in stick-built residential, the multiple pathway variable must also include the higher probability of resistive faults.

OK but what real path are you suggesting in a stick built home? Only ones I can think of would be CATV, SATV, TV antenna (maybe) and ham radio antennas. Otherwise all the circuits are IGR.

 

[Smart $]

OK but what real path are you suggesting in a stick built home? Only ones I can think of would be CATV, SATV, TV antenna (maybe) and ham radio antennas. Otherwise all the circuits are IGR.

Well those are some, but I was thinking more along the lines of ungrounded conductor faults not directly to ground (EGC, EGC bonded), e.g. metal plumbing fittings and fixtures connected to nonmetallic plumbing and such... while in commercial and industrial you typically have more metal structural framework that's bonded.