Simple GFI tester question

Status
Not open for further replies.
PetrosA

PetrosA

Senior Member
Location
Lancaster, Pennsylvania
I recently had a customer who is selling her house call me to install GFI receptacles in her upstairs bathrooms. The master bath had one, but the home inspector called out all the other ones. My three prong GFI tester also showed a lack of GFI protection on all the bath recepts, so I started installing the GFIs.

Without going into details that would incriminate me for stupidity, laziness and poor safety practices, I discovered that these outlets WERE GFI protected.

So my question is, why would two GFI testers (mine and the inspector's) fail to trip a GFI device, indicating that these receptacles weren't protected? Does a typical three-prong GFI tester become less effective at a certain distance from the GFCI device? Could the GFI receptacle have been so far out of spec that the resistors in the GFI testers wouldn't set it off? Is there some other possibility that I haven't thought of?

Thanks
 
Would the receptacles happen to be feed by an old two wire system? The receptacles do not have to have a ground attached to them to work, but the tester does.
 
The plug in GFCI tester use the EGC to cause a ground fault to test the GFCI.

If there is no EGC the plug in tester cannot make a ground fault.

The only recognized way to test a GFCI is the test button on a GFCI device or breaker. Both of which do not need an EGC to do the test.
 
120629-0856 EDT

Within a GFCI there is a 15,000 ohms resistor that is used for testing the GFCI. Probably the same size is used in GFCI testers.

Within the GFCI this resistor, when activated, is connected between the input neutral and the output hot of the GFCI, or vice versa. This creates a bypass current around the GFCI current transformer.

An external GFCI tester does not have access to these points. Only the output hot is available. Thus, something else is needed to provided the bypass path. The EGC is used for this bypass path. If there is no EGC or it is open somewhere, then no path.

If you simply put a 15,000 ohm resistor from the GFCI output hot to anywhere you could find an alternate path back to the main panel neutral, then you would have a test path. Possible this could be the sink faucet or drain if metal piping is used.

.
 
gar said:
If you simply put a 15,000 ohm resistor from the GFCI output hot to anywhere you could find an alternate path back to the main panel neutral, then you would have a test path. Possible this could be the sink faucet or drain if metal piping is used.
Click to expand...

You could but that seems a bit convoluted over just pushing the test button on the device. It is also not an accepted way to test per 110.3(B).

  • Check if outlet under scrutiny is live.
  • If yes push the test button wherever it may be.
  • Check outlet again to see if it is now dead.
  • If it is dead go back and reset
  • Verify the outlet is live again


Check out the testing section of this pdf.


http://www.cpsc.gov/cpscpub/pubs/099.pdf
 
120628-0955 EDT

iwire I agree, but if you have multiple ordinary outlets downstream from a GFCI, and the inspector wants to check that these are GFCI protected, then a bypass path needs to be created.

Also how do you easily prove that some regular receptacles are connected to a particular GFCI receptacle?

If a two wire to three wire adapter is used between the socket to be tested and the tester and the green grounding wire is connected to a wire that can be grounded somewhere, then the 15,000 ohm resistor of the tester can be used, and maybe the inspector will be satisfied.

i don't like the idea of a three prong socket not having the EGC pin connected as expected. If it is there it ought to serve its purpose.

.
 
gar said:
Also how do you easily prove that some regular receptacles are connected to a particular GFCI receptacle?
Click to expand...
Easy. Hit the trip button on the GFCI device, then check to see that the downstream receptacles lost power. Hit the reset button on the GFCI device and check to see that the downstream receptacles regained power. I think this is the only proper way to verify that GFCI protection is installed and is working.
 
gar said:
120628-0955 EDT

iwire I agree, but if you have multiple ordinary outlets downstream from a GFCI, and the inspector wants to check that these are GFCI protected, then a bypass path needs to be created.
Click to expand...

You cannot comply with 110.3(B) and 'create a method' The method must be the method the manufacturer provides.

Also how do you easily prove that some regular receptacles are connected to a particular GFCI receptacle?
Click to expand...

I covered that exact situation.

If a two wire to three wire adapter is used between the socket to be tested and the tester and the green grounding wire is connected to a wire that can be grounded somewhere, then the 15,000 ohm resistor of the tester can be used, and maybe the inspector will be satisfied.
Click to expand...

All that is true, but it does not meet code.

i don't like the idea of a three prong socket not having the EGC pin connected as expected. If it is there it ought to serve its purpose.
Click to expand...

I am not big of that either, the NEC requires the outlet be marked 'no equipment ground' but that is rarely done and even less understood by homeowners if it is done.
 
iwire said:
I am not big of that either, the NEC requires the outlet be marked 'no equipment ground' but that is rarely done and even less understood by homeowners if it is done.
Click to expand...

The NEC requires the receptacles, not the outlets, to be marked. See 406.3

It does not say the markings have to be visible, but it does specifically say the receptacles (not the covers) have to be marked.

I just had this conversation with a very strict inspector.
 
For clarification:

This is a large home (Custom McMansion style house) built in 1995, EGCs were present and spliced properly. The only GFI I could find on the second floor was in the master bath and looked original to the date of the house. The other four baths were evidently on the load side of the GFI, but weren't tripping it using my Greenlee GFI tester, nor did they trip when the home inspector used whatever GFI tester he had. The other receptacles in the master bath also failed to trip the GFI using the tester (no more than a 20' run from the GFI recept.). There were also two powder rooms on the 1st floor that failed a GFI test, but I suppose they could also have been on that same line. Either way, I think it's a horrible design to daisy chain five second floor bathrooms in a $2.5 million dollar house on one GFI receptacle and even worse to use a single 20A circuit for that many full baths - even in 1995.
 
Let's take a standard GFCI receptacle and take it apart. Inside you will find an electronics board (an IC to some of you), a sensor, and a contact. Since the innards of an actual GFCI are quite complicated, I'll use my crude drawing to make things (hopefully) a little more clear.

GFCI1.jpg




Now, let's plug in a power tool, with the assumption that the tool is safe to use.​

GFCI2.jpg


When you turn the drill on, it draws current through the entire ciruit.... the 'hot' and 'neutral' conductors both have the same amount of current flowing through them, only in opposite directions.​

GFCI5.jpg

Since the same amount of current is flowing through both the hot and neutral, they will cancel each other out and the sensor will detect 0.0 amps. If this sensor detects more than 0.005 amps (5mA), then the electronics picks this up and will open the contact, turning power off to the outlet.​

So let's do just that. Let's say there's a problem with the cord or the metal case of the drill (yea, I know, the drill looks like it's plastic, but let's pretend it's metal), and a ground fault exists between it and a ground. Now, current will flow through the hot, and to the ground fault. Let's say the fault current is 0.05 amps.​

GFCI6.jpg

The drill will still run properly, but part of the current it consumes does not return through the proper path (the neutral). Instead, it flows through the ground fault.​

The sensor will detect an imbalance of 0.05 amps, and the electronics will open the contact and turn of the power.​

This is the primary and sole function of GFCI protection. It is assumed a ground fault is going through a human being and turn the power off. Only when the fault is removed will the ability to restore power with the 'Reset' button work. The GFCI does not care whether there actually is a human getting shocked or not. It could be the end of a cord is lying in a puddle of water. It could be a fault the operator of the drill is not a part of. Nonetheless, once more than 0.005 amps flows through a ground fault, the GFCI opens.​

Now, in order to test a GFCI receptacle, the manufacturers put in a handy-dandy Test button. What this test button does is create a small current flow (using a resistor to simulate a load with a ground fault) that intentionally bypasses part of the sensor, forcing it to sense that imbalance. If the GFCI sensor, contact, and electronics are functioning properly the power will be shut off.​

GFCI7.jpg



Now, up until this point, you will notice the 'Ground' wire has not been mentioned, nor has any current flowed along it. Point is, none needs to. The ground wire is there to open the breaker or fuse if there is a fault within the wiring beyond the receptacle. If the drill was shorted internally, then the ground would carry enough current to cause the breaker or fuse to open and turn off the power. The ground wire is for the operation of the breaker or fuse, NOT the GFCI. If no ground wire existed to feed this receptacle, the GFCI will still function as designed...... sensing imbalances and turning power off in a ground-fault situation.​

The only function a ground wire has in the operation of a GFCI is when a plug-in tester is used to test a GFCI. One important note here; the industry standard is to use only the built-in Test button to check for proper function. UL does not recognize using plug-in testers as a proper method of testing GFCIs.​

GFCI8.jpg
​


With a properly-wired (3-wire) GFCI receptacle, the only way a plug-in tester can safely simulate an imbalance is to induce it through the ground. It has to, since it cannot bypass the sensor using the neutral. If you placed the resistor between the hot and neutral, the GFCI would only 'see' it as any other load, as the current flow between the hot and neutral would be equal and cancel each other out at the sensor.​

With the plug-in tester, the current flows out through the hot, and back through the ground. Since current flow on the ground is not going through the sensor, only the current on the hot is detected. With no opposing return current flow to balance it out, the electronics assume a ground fault and opens the contact... turning power off.​


If you only have a 2-wire (ungrounded) circuit, then the plug-in tester cannot create a current flow on the ground as the ground is not there. Since no current can flow in this incomplete circuit, the GFCI will sense 0.0 flow and not open the contact.​

GFCI9.jpg


There are other things today's GFCIs do that are not relevant to the discussion here, but I'll mention them here.​

One is Loss of Neutral. The electronics can sense the loss of the circuit's neutral, and will open the contact. The reason being, if there is no neutral, the electronics cannot fuction in order to open the contact in the event of a ground fault. So if the neutral feeding the circuit opens, the GFCI will turn the power off.​

Another is a self-test function. If the internal wiring of the GFCI becomes damaged (either physically or, say, a surge due to lighting), the power will shut off as well. Pushing the Reset button will not reset it.... it will stay off and need to be replaced.​

Another important function is Line-Load and Hot/Neutral Miswire detection. If the power to the receptacle is mistakenly attached to the Load terminals, or if the hot and neutral are reversed on the Line terminals, the GFCI will detect these installation errors as well and will not turn on until the problem(s) is (are) corrected. This is to ensure the GFCI is wired correctly at initial install to provide proper protection for the life of the unit.​

Most GFCI receptacles today are also Tamper-Resistant (notice the TR on the face) to keep foreign objects from being pushed into the slots, and GFCIs are available in Weather-Resistant versions for installation in damp and wet locations.​



OK, school is out.
thumbsup.gif

​
 
Just a couple of observations:

My plug in tester stops working below 103 volts. Not enough fault current flowing to trip.

A Lo-Z multi-meter or meter attachment will trip a GFCI pretty good. :happyyes:

I seem to recall a study that said over half or just under half of GFCIs were not working after 5 or 10 years.
 
Last edited:
PetrosA said:
For clarification:

This is a large home (Custom McMansion style house) built in 1995, EGCs were present and spliced properly. The only GFI I could find on the second floor was in the master bath and looked original to the date of the house... .
Click to expand...
Given what you have desciribed here I would say you had a faulty device.
 
K8MHZ said:
The NEC requires the receptacles, not the outlets, to be marked. See 406.3

It does not say the markings have to be visible, but it does specifically say the receptacles (not the covers) have to be marked.

I just had this conversation with a very strict inspector.
Click to expand...

406.4(D).

Wow never read it that way before!
 
jxofaltrds said:
406.4(D).

Wow never read it that way before!
Click to expand...

I must be on a roll!

It seems I also dug up the fact that if NM is securely attached to a metal enclosure, there is no rule stating the sheath has to be at least 1/4" inside the box.

I need to start keeping track of these little tid bits, I run across them every now and then and just use them as conversation starters for the inspectors I know.
 
iwire said:
Sorry, I find outlets to be quicker to spell then receptacles but yes it is receptacles.
Click to expand...

Easier to say, too. I do it all the time.

I pointed out because that is the part of the outlet the NEC refers to as needing to be marked, while standard practice is to mark the cover, which is also a part of the outlet, but it is not the receptacle.

If you think I was trying to pick on you, you should know by now that I would never do such a dastardly thing.
 
Status
Not open for further replies.
Top