Different Kind of Failed GFCI

Status
Not open for further replies.

Little Bill

Moderator
Staff member
Location
Tennessee NEC:2017
Occupation
Semi-Retired Electrician
I've found and replaced several GFCIs that had gone bad. I think I may have found one that had failed and was still "live". But by far most fail and are dead.
Well the other day I was asked to check a receptacle that wasn't working while I was there to move a light.
I traced the receptacle's (non-GFCI) feed to a GFCI recep. on the same wall in the kitchen. The GFCI was working and hadn't tripped. But what I found was the load side had failed.
That was new to me.
Have any of you found one that failed this way?
 

hurk27

Senior Member
Prior to the new UL requirements I think some where around 1997 or 1998, many GFCI's didn't have the fail open and would keep on supplying power even when the GFCI was bad, most today will fail open, but I have found a few China ones that would not, and sent them into UL for them to take care of.
 

Little Bill

Moderator
Staff member
Location
Tennessee NEC:2017
Occupation
Semi-Retired Electrician
Prior to the new UL requirements I think some where around 1997 or 1998, many GFCI's didn't have the fail open and would keep on supplying power even when the GFCI was bad, most today will fail open, but I have found a few China ones that would not, and sent them into UL for them to take care of.

This house is only 5 years old. The GFCI was functioning (according to the test button) other than the load side. The receptacle on the GFCI itself had power, just none downstream.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
120810-1238 EDT

Little Bill:

Open the defective GFCI. There should be a separate set of contacts for the load side to produce the effect that you observe. Something relative to these contacts failed. The contacts are probably riveted.

.
 

Dennis Alwon

Moderator
Staff member
Location
Chapel Hill, NC
Occupation
Retired Electrical Contractor
I have seen many of the older GFCI's fail while still being live. That is why that test button is there to make sure that the unit still operated.
 

Little Bill

Moderator
Staff member
Location
Tennessee NEC:2017
Occupation
Semi-Retired Electrician
120810-1238 EDT

Little Bill:

Open the defective GFCI. There should be a separate set of contacts for the load side to produce the effect that you observe. Something relative to these contacts failed. The contacts are probably riveted.

.

I would do an autopsy as well. And I would post pics here, too. Hint, hint.

Unfortunately, I can't seem to find the old/defective GFCI. I suspect it went in my trash box and got thrown away. At the end of the day I'm usually worn out and things don't get put where they belong until later. I keep a box for trash and when I change out devices I usually just toss them in there. I think that's where this one ended up.

I have seen many of the older GFCI's fail while still being live. That is why that test button is there to make sure that the unit still operated.

I have seen GFCIs fail live but not like this. I did use the test button and it worked. I was able to reset it. Only thing wrong was the load side failed. The GFCI receptacle had power and would trip/reset but nothing downstream off the load side.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
120811-1226 EDT

Little Bill:

If you go to my web site at http://beta-a2.com/GFCI.html , then go to photo P8, but it is labeled P5 presently, you will find a photo of the contacts inside a Leviton 7899 GFCI. After the actual P5 photo all are labeled P5. So count from the actual P5 to get to P8.

This photo is titled "Current Transformer Location and Other Components". The "Hot Load Terminal & Fixed Contact" is a small assembly of a contact and terminal for the GFCI load terminal. For there to be no power to the load terminals something has to happen to the hot or neutral contacts or the spring beams. Different manufacturers possibly have different configurations.

.
 

Little Bill

Moderator
Staff member
Location
Tennessee NEC:2017
Occupation
Semi-Retired Electrician
120811-1226 EDT

Little Bill:

If you go to my web site at http://beta-a2.com/GFCI.html , then go to photo P8, but it is labeled P5 presently, you will find a photo of the contacts inside a Leviton 7899 GFCI. After the actual P5 photo all are labeled P5. So count from the actual P5 to get to P8.

This photo is titled "Current Transformer Location and Other Components". The "Hot Load Terminal & Fixed Contact" is a small assembly of a contact and terminal for the GFCI load terminal. For there to be no power to the load terminals something has to happen to the hot or neutral contacts or the spring beams. Different manufacturers possibly have different configurations.

.

Thanks Gar!
Great pics and info. I wish I had not lost/thrown away my defective GFCI.:(
 

jmellc

Senior Member
Location
Durham, NC
Occupation
Facility Maintenance Tech. Licensed Electrician
I have seen that a time or 2. A conductor in the GFCI burns out or vibrates loose, etc. Most GFCI outlets are going to fail at some point, in this way or some other. I have seen a few last 10 years or more, but have seen a lot only last a year or 2. Have seen a few fail after a few months.

I wish GFCI breakers were not so expensive. They give superior protection and I have seldom seen them fail, except for GE brand. SqD, CH, Siemens, or CHbr have all done well for me. Outlets are cheaper up front, but breakers are cheaper in the long haul and are much less subject to nuisance tripping. Pigtailing can also be done at all outlets, which is the only right way to do them, IMHO.
 

George Stolz

Moderator
Staff member
Location
Windsor, CO NEC: 2017
Occupation
Service Manager
You think that's weird? I saw one that functioned on the load side and would not supply power on the contacts of the GFCI itself. That one weirded me out.

Jmelic: you'd rather install a dozen devices to fail versus one with duplexes on the load side? Sounds like you'd pay for your breaker anyway. :)
 

jmellc

Senior Member
Location
Durham, NC
Occupation
Facility Maintenance Tech. Licensed Electrician
You think that's weird? I saw one that functioned on the load side and would not supply power on the contacts of the GFCI itself. That one weirded me out.

Jmelic: you'd rather install a dozen devices to fail versus one with duplexes on the load side? Sounds like you'd pay for your breaker anyway. :)

I think you missed what I said. If I install a GFCI receptacle and protect others downline with it, that receptacle takes more load. With a breaker, I pigtail at each receptacle and each one carries only its own load. 1 receptacle going bad does not kill others downline. I also see receptacles trip from appliance loads that do not trip breakers. Some receptacles trip when power is restored after an outage. That does not happen with breakers, not that I have ever seen. On top of that, receptacles give out far more often than breakers. Breakers are the cheapest way in the long run.
 

kwired

Electron manager
Location
NE Nebraska
I think what has likely happened is some connection within the device failed - located after the contacts that open when it trips, but someplace where it branches from contact to the receptacle to the load terminal screw. Very possible it was just one conductor it happend to and not both hot and neutral. A little more testing may have indicated you only lost continuity to one conductor to the load terminals. A plug in tester would not light up at all if all that was lost is the hot conductor, if the neutral was lost it would indicate open neutral. A volt meter would still read voltage from line side hot to load side neutral if load side hot is all that is lost.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
120812-2103 EDT

kwired:

In the case of the Leviton there are quite thick conductive parts between the fixed contacts on the load side, the terminal, and the riveted silver contact. Same for the fixed (non-moving) silver contacts for the receptacle socket members. The most likely failure point is a poor crimp on the rivet joint. However, improperly heat treated or work hardened materials in the conductive elements could have fractured in forming, and be easily broken under high current stress.

My photos, that I previously referenced, clearly show the construction of the Leviton unit.

The spring members supporting the moving contacts are a weaker mechanical element and might not produce correct contact pressure if incorrectly formed or bent at assembly. Also the rivet joint on these springs is a potential problem.

Whenever someone has a failure like this it is important to disassemble the defective unit and look for a cause.

.
 

kwired

Electron manager
Location
NE Nebraska
120812-2103 EDT

kwired:

In the case of the Leviton there are quite thick conductive parts between the fixed contacts on the load side, the terminal, and the riveted silver contact. Same for the fixed (non-moving) silver contacts for the receptacle socket members. The most likely failure point is a poor crimp on the rivet joint. However, improperly heat treated or work hardened materials in the conductive elements could have fractured in forming, and be easily broken under high current stress.

My photos, that I previously referenced, clearly show the construction of the Leviton unit.

The spring members supporting the moving contacts are a weaker mechanical element and might not produce correct contact pressure if incorrectly formed or bent at assembly. Also the rivet joint on these springs is a potential problem.

Whenever someone has a failure like this it is important to disassemble the defective unit and look for a cause.

.

I did not look at your photos, and how a particular unit is constructed is not extremenly important as I don't think OP even told us what model of GFCI he had. My point was that an internal component is where the open circuit is between the receptacle on the face of the unit and the load terminals. They both are connected to load side of the trip mechanism and that appears to still work, all that is left to fail is whatever connects the two user outputs together.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
120815-0830 EDT

kwired:

I believe most new GFCIs have a somewhat similar construction, but different ways of testing. This is probably determined by UL specifications.

The new devices seem to have 4 output poles with a single throw action. Two output poles derived from the input hot leg and two from the neutral. In the tripped state all poles are open and in the reset state all poles are closed.

When the GFCI is shipped from the factory it has a tape across the output terminals and it is in the tripped state. The tape should be sufficient to get the source (supply) connected to the input terminals, although it won't prevent neutral and hot being interchanged.

Suppose the tape is removed and someone connects the power source to the output terminals, then the receptacle sockets will not be powered because because there is no connection between the output terminals and the receptacle sockets in the tripped state. Further in the case of Leviton it can not be reset because there is no power to the electronics. In this device the reset function is not simply a mechanical latch.

If power is applied to the input terminals and the GFCI is latched, and then input power is removed the GFCI the GFCI will remain latched. Now remove the power source from the input terminals, remove the tape, and connect power to the output terminals, then power will exist at the original input terminals and the receptacle sockets. Connect no load to the original input terminals and this GFCI will normally never trip even with leakage current in a device plugged into the GFCI. This is because no unbalanced current flows thru the GFCI current transformer.

If the Leviton is wired backwards and is in the latched state, then operating the test button will trip the GFCI, and it can not be reset after this. The green light remains on with back fed power, but no power is connected to the receptacle sockets. A high impedance meter will read about 10 V because of capacitive coupling.

.
 

kwired

Electron manager
Location
NE Nebraska
120815-0830 EDT

kwired:

I believe most new GFCIs have a somewhat similar construction, but different ways of testing. This is probably determined by UL specifications.

The new devices seem to have 4 output poles with a single throw action. Two output poles derived from the input hot leg and two from the neutral. In the tripped state all poles are open and in the reset state all poles are closed.

When the GFCI is shipped from the factory it has a tape across the output terminals and it is in the tripped state. The tape should be sufficient to get the source (supply) connected to the input terminals, although it won't prevent neutral and hot being interchanged.

Suppose the tape is removed and someone connects the power source to the output terminals, then the receptacle sockets will not be powered because because there is no connection between the output terminals and the receptacle sockets in the tripped state. Further in the case of Leviton it can not be reset because there is no power to the electronics. In this device the reset function is not simply a mechanical latch.

If power is applied to the input terminals and the GFCI is latched, and then input power is removed the GFCI the GFCI will remain latched. Now remove the power source from the input terminals, remove the tape, and connect power to the output terminals, then power will exist at the original input terminals and the receptacle sockets. Connect no load to the original input terminals and this GFCI will normally never trip even with leakage current in a device plugged into the GFCI. This is because no unbalanced current flows thru the GFCI current transformer.

If the Leviton is wired backwards and is in the latched state, then operating the test button will trip the GFCI, and it can not be reset after this. The green light remains on with back fed power, but no power is connected to the receptacle sockets. A high impedance meter will read about 10 V because of capacitive coupling.

.

I have never considered just how they make them so they will not work if line/load are reversed, what you said makes sense though and probably is a fairly common scheme for all. So the OP probably had a failed contact or component after the contact on the portion going to at least one of the "load" terminals.

I always wondered why P&S GFCI's always trip first time you power them up - this may be why. Even if you press reset if it is in trip position the button locks as if it was reset, but when you power it for the first time it still trips immediately. I would bet if I reversed line and load on one of these and powered it up I could get it back into this state it is in when new. Now I have to try that, just out of curiosity.
 
Status
Not open for further replies.
Top