Theoretical GFCI question, Fridge should never be on one?

[mbrooke]

Ok help me understand once and for all. Can an inductor or capacitor big enough trip a GFCI or is psychically impossible? My understanding is that capacitors and inductors changes phase angle, however, my understanding is that this can not and will not cause an imbalance between 2 wires that could be perceived as a ground fault? Even though steady state and transient phenomenon take place, this cant cause electrons to differ between hot and neutral?

What got me thinking about this was a question that keeps coming up or more a statement by other electricians/engineers that refrigerators should never be put on a GFCI. The theory is that because the motor (compressor) starting is highly inductive it can cause a GFCI to nuisance trip. I will admit that I have seen many older GFCIs trip on a fridge as well as many newer GFCIs trip on old fridges. One case I lived in an apartment where the countertop GFCI fed the fridge. Never a problem, but much to my surprise on day as I opened the fridge the light bulb burnt out... much to be surprise the internal arc over in the bulb tripped the GFCI:blink::blink::huh:

There are many times where code under some NEC cycles forces me to add a GFCI behind a fridge/freezer (such as a garage with an accessible outlet), but are the concerns really valid? And why?

 

[ptonsparky]

Put the fridge on the gfci. If the old one trips it, buy a new fridge. GFCIs work.

 

[jumper]

Put the fridge on the gfci. If the old one trips it, buy a new fridge. GFCIs work.

Agreed.

 

[iwire]

I also agree with the others.

Much larger motors that are cycled much more often on construction sites work just fine on GFCIs.


Besides, in many cases there is no choice in using GFCIs with refrigerators.

 

[mbrooke]

I know, but what is the science behind it?

 

[ptonsparky]

I know, but what is the science behind it?

Old refrigerators that need to be replaced. Simple as that.

 

[ptonsparky]

When gfci protection first became available many appliances had in excess of 5ma leakage current. Design, age or both. The old fridge worked just fine until those new fangled gadgets showed up. Most of us knew that you stood on the " board" before opening the door.

 

[mbrooke]

When gfci protection first became available many appliances had in excess of 5ma leakage current. Design, age or both. The old fridge worked just fine until those new fangled gadgets showed up. Most of us knew that you stood on the " board" before opening the door.

Makes sense. But what caused the higher leakage current? Ive heard many times that terminals on defrost heaters weren't well designed and sometimes defrosting ice bridged them to ground momentarily.


And can a capacitor or inductor large enough trip a GFCI?

 

[GoldDigger]

A capacitor to ground can trip a GFCI; a capacitor across line and neutral should not.
Switching transients and arcs between line and neutral can cause high frequency current components (harmonics).
These current harmonics combined with voltage drop from series resistance or inductance can cause voltage harmonics on the conductors.
Higher harmonics can cause more current to ground in stray capacitance to ground that would not pass enough current at 60Hz to trip the GFCI.
The higher current harmonic flow in the stray capacitance can then trip the GFCI.
Alternatively, there could be a high resistance path through the EGC or other ground which is parallel to the neutral. It will carry a fixed fraction of the neutral current.
This fraction of the FLA could be less than 6ma while the same fraction of the LRA could be above 6ma.
Just some hypothetical explanations with no experimental backup.

 

[kwired]

Inductive kickback is sometimes a problem with GFCI's. Most reputable GFCI's made by Leviton, P&S, Cooper... and most circuit breaker type have some ability to overlook kickback, but will still trip when a certain threshold has been reached. Cheap GFCI's you get at big box stores suffer more "nuisance trips" from kickback then the good ones do.

P&S has a GFCI receptacle with an audible alarm that sounds when tripped - comes in handy for refrigerators and freezers in garages, basements, etc. where they may go unnoticed when tripped, I have started using them in those kinds of places - cheap insurance against having a freezer full of thawed food.

 

[charlie b]

I know, but what is the science behind it?

Current goes to the load on the ungrounded ("hot" or "phase") conductor and returns on the grounded ("neutral") conductor. The GFCI device has a wire loop that encircles both conductors. If the current in both conductors is the same value and in opposite directions, then the net magnetic field detected by the wire loop will be zero. If some current that left the source on the ungrounded conductor does not return via the grounded conductor (i.e., there is leakage current somewhere along the run), then the wire loop will see a net magnetic field. If the intensity of that magnetic field exceeds a predetermined value (i.e., corresponding to a leakage current of 3 - 5 milliamps), the GFCI device will trip.

Any appliance that trips a GFCI is in need of repair or replacement. The trip indicates the existence of leakage current.

Any capacitive or inductive load that has all of its current coming into it via the ungrounded conductor, and all of its current leaving via the grounded conductor, will not result in a difference in current as seen by the GFCI device. Golddigger has given some examples of capacitive loads that will and that will not trip a GFCI.

Does that help answer your questions?

 

[xformer]

I ran across a florescent fixture that would occasional trip a GFCI due to what I call Capacitor bleed off. About every fourth or fifth turn off, the GFCI would trip. I could see the same happening in an appliance that had capacitors connected to them.

 

[mbrooke]

A capacitor to ground can trip a GFCI; a capacitor across line and neutral should not.
Switching transients and arcs between line and neutral can cause high frequency current components (harmonics).
These current harmonics combined with voltage drop from series resistance or inductance can cause voltage harmonics on the conductors.
Higher harmonics can cause more current to ground in stray capacitance to ground that would not pass enough current at 60Hz to trip the GFCI.
The higher current harmonic flow in the stray capacitance can then trip the GFCI.
Alternatively, there could be a high resistance path through the EGC or other ground which is parallel to the neutral. It will carry a fixed fraction of the neutral current.
This fraction of the FLA could be less than 6ma while the same fraction of the LRA could be above 6ma.
Just some hypothetical explanations with no experimental backup.

That's been my thinking as well. I frequently hear the theory (though no one has proved it) that older GFCIs could not filter out transient and harmonics to well. The starting inrush of the compressor's locked rotor current would trip them.

FWIW, the code seems to be aware a GFCI can false trip. An example of that is I believe in the NEC handbook where wet medical process rooms mention that a GFCI can nuisance trip.

 

[mbrooke]

Current goes to the load on the ungrounded ("hot" or "phase") conductor and returns on the grounded ("neutral") conductor. The GFCI device has a wire loop that encircles both conductors. If the current in both conductors is the same value and in opposite directions, then the net magnetic field detected by the wire loop will be zero. If some current that left the source on the ungrounded conductor does not return via the grounded conductor (i.e., there is leakage current somewhere along the run), then the wire loop will see a net magnetic field. If the intensity of that magnetic field exceeds a predetermined value (i.e., corresponding to a leakage current of 3 - 5 milliamps), the GFCI device will trip.

Any appliance that trips a GFCI is in need of repair or replacement. The trip indicates the existence of leakage current.


Any capacitive or inductive load that has all of its current coming into it via the ungrounded conductor, and all of its current leaving via the grounded conductor, will not result in a difference in current as seen by the GFCI device. Golddigger has given some examples of capacitive loads that will and that will not trip a GFCI.

Does that help answer your questions?

I get that part well, but the part that confuses me are claims made by EE and sparkies that a fridge where ground leakage is less than 5ma (net current will not be high enough in the sensing coil to trip it), can still trip a GFCI from transients and non steady state phenomenon. Not sure if it even true, but my line of thinking is on par with Golddigger.

 

[iwire]

claims made by EE and sparkies that a fridge where ground leakage is less than 5ma (net current will not be high enough in the sensing coil to trip it), can still trip a GFCI from transients and non steady state phenomenon.

Did they have a recording meter on the units when they tripped?

I would bet not.

 

[mbrooke]

Did they have a recording meter on the units when they tripped?

I would bet not.

Even if I did not sure what I would be recording exactly. I have no idea what is really doing it.

 

[mjc763]

One reason I do not put them on the load side of a gfi is because homeowners will not notice right away that it tripped which could lead to hundreds of dollars in wasted food. I've heard of the nuisance tripping caused by older compressors as well.

 

[don_resqcapt19]

One reason I do not put them on the load side of a gfi is because homeowners will not notice right away that it tripped which could lead to hundreds of dollars in wasted food. I've heard of the nuisance tripping caused by older compressors as well.

There are conditions where you have no choice, if you are going to install per the rules found in the NEC.

 

[mjc763]

There are conditions where you have no choice, if you are going to install per the rules found in the NEC.

The only condition I can think of in a house would be if the fridge was within 6' of a sink and that's only if your on the 2014 code. And of course unfinished areas, and garages, but thats not typical. I've had a city inspector require the fridge to fed line side in kitchens, I'm guessing he had a bad experience.

 

[iwire]

The only condition I can think of in a house would be if the fridge was within 6' of a sink and that's only if your on the 2014 code. And of course unfinished areas, and garages, but thats not typical. I've had a city inspector require the fridge to fed line side in kitchens, I'm guessing he had a bad experience.

This thread had been about GFCIs in general, not limited to dwelling unit kitchens.

I spend my time in commecial kitchens which require all 15 and 20 amp 125 volt receptacles to be GFCI. That results in many refrigeration units being supplied by a GFCI.