Installing GFCI in garage with no ground

[Greg Fretwell]

Originally Posted by Mike Whitt

The GFCI never trips by use of the grounding electrode or the equipment grounding conductor in normal operation.

What makes it trip is an inblance between the Hot and Neutral conductors. This is why it is not proper to test a GFCI device with one to those plug-in testers.

It would trip a GFCI if you had a phase or neutral fault to the EGC and that EGC was connected to a ground rod. Normally that would be a fault to the main bonding jumper but on a 2 wire circuit that path is not available. At least, if you had it tied to earth ground you would not have energized equipment waiting for a person to create that imbalance. It would trip the GFCI immediately on the first fault. The case of your equipment would also be at "local ground" potential. That is why we have to drive a rod at a secoind building in the first place.
I agree, the normal fault path is via the EGC to the MBJ but you don't have that connection here.


As a totally off topic conversation starter, wouldn't a bootleg ground, connected to a ground electrode be the same as a 250.32(B)(2)???

[rromoser]

How about using "wiggys" solenoid type voltage tester? One lead goes to the hot side of the GFCI. The other goes to "Earth" ground (waterpipe, metal faucet, etc.) It had better trip the GFCI

[Mike Whitt]

Greg and Randall

The connection to earth through a grounding electrodes weather a rod, water pipe, concrete encased, ground ring or what ever used to connect to earth plays no role in opening anything be it the fuse, breaker or GFCI device.

The sole purpose of an earth ground is for:
250.4 General Requirements for Grounding and Bonding.
The following general requirements identify what grounding and bonding of electrical systems are required to accomplish. The prescriptive methods contained in Article 250 shall be followed to comply with the performance requirements of this section.
(A) Grounded Systems.
(1) Electrical System Grounding. Electrical systems that are grounded shall be connected to earth in a manner that will limit the voltage imposed by lightning, line surges, or unintentional contact with higher-voltage lines and that will stabilize the voltage to earth during normal operation.

The purpose of the rod at the second building is the same as it is for the first building and has nothing to do with the operation of any type of safety device be it a fuse, breaker or GFCI.

The overcurrent device opens in a fault condition due to the enormous amount of current being drawn due to the bonding of the equipment grounding conductor back to the neutral at the service.

A GFCI device opens when there is more current on either the hot or the neutral than on the other. The amount of current difference can be no more than .006 amps.

The use of the plug-in tester does not limit the current to .006 amps and sometimes will allow enough current to flow that it will open a 15 amp fuse.

The GFCI device could care less what is causing the imbalance or where the current is going. It only cares that an imbalance of .006 amps or greater exists and if properly installed will open the circuit.

The only true test that the GFCI device is operating at the .006 amps is to use the test button on the device as mentioned in this thread.
For someone to use a plug-in tester and then sign off that the GFCI is working properly is leaving their self open for litigation as these devices are not approved to test GFCI devices.
They are alright to check to see if other receptacles are on a circuit protected by GFCI but it should NEVER be the only test done.

[Mike Whitt]

Originally Posted by rromoser

How about using "wiggys" solenoid type voltage tester? One lead goes to the hot side of the GFCI. The other goes to "Earth" ground (waterpipe, metal faucet, etc.) It had better trip the GFCI

Yes this will open a GFCI device but can you tell at what amperage?

[Greg Fretwell]

Mike are you saying a fault to a ground rod (not bonded to the service) will NOT operate a GFCI?

[David P. Valley]

Originally Posted by Greg Fretwell

Mike are you saying a fault to a ground rod (not bonded to the service) will NOT operate a GFCI?

I hope not...

[Greg Fretwell]

BTW Mike the QO GFCI seems to be tested at the 12ma level, not 6

[Greg Fretwell]

David, why would you think 120v to ground, any ground would not create a 6ma current? Dropping an extension cord in the wet grass seems to defy that notion.
My only point in "grounding" your EGC of a 2 wire connected receptacle if you don't have a conductor back to the service MBJ is a short to the box WOULD trip the GFCI without having to have a fault through a person.

Lets take the example of the dead kid. If they had that garage door grounded to the earth and a GFCI installed it would have tripped immediately and not waited for a wet kid to complete the circuit.

[Mike Whitt]

Originally Posted by Greg Fretwell

Mike are you saying a fault to a ground rod (not bonded to the service) will NOT operate a GFCI?

Might and might not. A rod with more than 2000 ohms would not open the GFCI device.
Is it possible to have an eight foot rod with more than 2000 ohms? Yes especially in my area.
I have driven rods in lab experiments that didn’t open a GFCI breaker

Originally Posted by Greg Fretwell

BTW Mike the QO GFCI seems to be tested at the 12ma level, not 6

May be true if you are figuring the resistor only but the impedance of the complete circuit will bring it down to the UL requirement of 6 miliamps.

Originally Posted by Greg Fretwell

David, why would you think 120v to ground, any ground would not create a 6ma current? Dropping an extension cord in the wet grass seems to defy that notion.

As outlined above it is very possible to have a grounding electrode system with to much resistance to open the GFCI.
An electrode system that has been installed for several years could possibly be deteriorated completely away.

Originally Posted by Greg Fretwell

My only point in "grounding" your EGC of a 2 wire connected receptacle if you don't have a conductor back to the service MBJ is a short to the box WOULD trip the GFCI without having to have a fault through a person.

I don’t understand what you are saying here.
If the box is not bonded to the service where will the current go?

Originally Posted by Greg Fretwell

Lets take the example of the dead kid. If they had that garage door grounded to the earth and a GFCI installed it would have tripped immediately and not waited for a wet kid to complete the circuit.

This is very iffy. It would depend on the resistance of the soil between the rod and the ground at the transformer.

The bottom line is that a rod DOES NOT operate the GFCI device.

[Greg Fretwell]

If a rod wouldn't operate a GFCI how can that extension cord in the wet grass do it? BTW not to be pedantic but 20K ohms should operate your GFCI quite nicely

120/24000=0.005

If you are saying an 8' ground rod won't get you <24k why bother to drive any at all?

[jtedesco1]

Testing a ground rod

[Mike Whitt]

Originally Posted by Greg Fretwell

If a rod wouldn't operate a GFCI how can that extension cord in the wet grass do it?

Probably by causing a imbalance between the hot and neutral as this is what the GFCI is looking for.

Originally Posted by Greg Fretwell

BTW not to be pedantic but 20K ohms should operate your GFCI quite nicely

Originally Posted by Greg Fretwell

120/24000=0.005

Well the device is listed for 6 mil not 5 mil so it could possibly hold

Originally Posted by Greg Fretwell

If you are saying an 8' ground rod won't get you <24k why bother to drive any at all?

The original post was about a system that was installed before the equipment grounding conductor was required.
It would be nothing unusual to find a ¾ inch water pipe driven a couple of feet into the earth and mostly rusted away on a system that old.
I would not want to trust a GFCI device to operate on a system that might be in this condition, would you?

[Greg Fretwell]

In backward order,
it is not up to me to "trust" a GFCI on this system (2 wire, no ground) 406.3(D)(3)(b) says I have to.
I said 20k ohms would trip your GFCI, that IS 6ma (you said 2k wouldn't and that is 60ma, double the amount for a GFP) The math example 24k gave you the nominal 5ma number
My whole point was to provide a path to ground if the case of the equipment became energized BEFORE a person completes that path.
If you don't think a ground rod would give you <20k ohms to ground why would you ever bother to drive one? I bet you a cup of coffee I could get enough current to trip a GFCI by simply shooting a 2" tap con in the concrete floor and putting a wire under it.

[Mike Whitt]

Greg

The one thing that 406.3(D)(3)(b) does not require is a connection to an equipment grounding conductor. As a matter of fact the use of one is prohibited. “An equipment grounding conductor shall not be connected between the grounding-type receptacles.”

You are correct about the 20K ohms and this is what I meant to type but kept leaving one of the “0s” off.
I have found many of a 2 foot piece of ¾ galvanized pipe used as a ground rod and when I tried to pull it out of the ground it came out in pieces. This rod would not have been beneficial to any part of the system including the operation of a GFCI device.

As to the tap-con I would agree that it might open a GFCI device and will agree that concrete is a better way to ground than earth by itself. I would also agree that under the right circumstances that the tap-con would have no affect on the GFCI device.

What I am saying is that we should NEVER use the earth or even a concrete pad as a fault conductor especially for the operation of a GFCI device.

We must also remember that the GFCI device does not rely on a ground path or a connection to earth in order to operate; it functions by reading a difference between the current carrying conductors.

Let’s use you example of the rod at the second building, “That is why we have to drive a rod at a second building in the first place.” Found here

If an equipment grounding conductor is installed with the feeders to the second building then the grounded (neutral) conductor is isolated form the equipment grounding conductor at the second building and the equipment grounding conductor is bonded back at the main of the first building in the service. The equipment grounding conductor is also bonding to panel and the rod at the second building in the event of a lightning strike.

If no equipment grounding conductor is installed with the feeders then the grounded (neutral) is bonded to the equipment grounding conductors, the panel and the rod at the second building.

The reason for these two different types of installations is simple. There MUST be a low impedance connection between the equipment grounding conductor and the source of power, the neutral.
What we never want to do is have a high impedance connection using things such as earth ground.

We have no control over the climate and what the weather brings us on a daily basis. In extremely dry conditions there could be no path at all through a rod connected to earth and the source in which case a GFCI would not operate at all and especially if it is an old system that has rotted away.

I am not debating whether or not a newly installed rod or even a tap-con in concrete would open a GFCI or not but instead I am trying to impress is the fact that the earth or the tap-con in concrete is not a conductor that we should be relying on to operate a GFCI device. There is way to many variables involved to put the safety of someone on depending on a rod to open a GFCI device.
If this was a safe procedure then it would be allowed to be installed by the NEC but this is not an acceptable installation using the rules outlined in the NEC.

[David P. Valley]

Wow,

You two are really getting deep with this GFCI situation. I'm learning quite a bit on this thread. Thanks to the both of you.