Leviton GFCI nuisance tripping and circuit analysis

Status
Not open for further replies.

tersh

Senior Member
Location
new york
I am pondering the following:

The GFCI chip is powered by connection to the AC line through a resistor, a diode, and the 'trip solenoid' (which I presume is supposed to do double duty as a filter choke). The capacitor from pin 6 is a decoupling capacitor to filter the rectified AC.

The circuit has a residual current detection circuit that uses an op-amp connected to a sense coil, using standard op-amp feedback through Rset. The output of this amplifier connects to a pair of comparitors so that when the output of the RCD is high enough the system is triggered.

Also connected to the _output_ of the op-amp is the 'grounded neutral detection' circuit. The way this works is that if the neutral is grounded upstream of the GFCI, the combination of the ground-neutral loop, the various capacitors and the second toroid creates a feedback path that makes the sense op-amp oscillate and drive a detectable current through the ground-neutral loop.

I am wondering if particular inductive loads can interact with the grounded neutral detection, and dump current into pin 1 via the power supply connection, and trip the circuit not by creating an actual residual current, but rather by creating a current spike at pin 1 of the chip.

Tersh, if you want to continue to research this, I think that the next step would be to build a proper spice model of the circuit, and simulate what happens with various inductive loads, _or_ probe the circuit using a good oscilloscope. You might also hire someone to do this research.

Very, very interesting, since you have clearly demonstrated a case where there is no 'residual current', but you are consistently tripping this circuit.

-Jon

First. My Siemens GFCI Breaker doesn't trip after 3 days with refrigerator on it. Good. While last year with the FM2141 chip GFCI outlet. It tripped twice a day or so hence I thought it was the fridge that was the problem.

Btw.. My course in college was electronic engineering. Spent 5 years in it. So I am supposed to be familiar with it. But that was long ago and changed interests to the medical field and then to physics now. Shaded pole motors and Leviton GFCIs are very common in the US. Can people with these try them on their oscilloscope? Also there is this person with electric fan that trips on his two Leviton GFCI. He has a theory how it occurs that it's because each of the two wire has actual inbalance current. Please comment on his theories and what could be wrong with it. He said in:

https://www.justanswer.com/electrical/2g0be-gfci-outlet-trips-even-when-clear.html

"
It seems very unlikely that the problem is caused by a loose connection because the nuisance trips occur only when one opens a switch to the load, not at random as one would expect if the problem were caused by a loose connection.


After giving the problem a good deal of thought, I have concluded that it is due to a transient that occurs when the load is switched off. It does not occur when the load is purely resistive (e.g., an iron). It occurs when an inductive load , such as a
motor
is switched off. In the GFCI, the two wires that feed the load pass through a small current transformer that produces no output if the currents in the two wires are equal in magnitude and opposite in direction. In other words, the GFCI trips if the net current through the transformer is not zero.

When the load is purely resistive, the currnet in both wires falls to zero instantaneously when the switch is opened. Hence the GFCI does not trip.

Now suppose that the load is inductive (equivalent to a resistance in series with an inductance). When the switch is opened, one leg of the circuit consists of only a wire and hence the current in it falls instantaneously to zero when the switch is opened. In the leg of the circuit containing the load , what happens depends upon the currrent in the indictance at the moment when the switch opens. (Remember that the current in the inductance varies sinusoidally before the switch opens.) If the current is zero when the switch opens, nothing more happens and the GFCI does not trip because the currents passing through the transformer are equal and opposite at all times. If the current in the inductance is not zero when the switch opens. it does not fall to zero instantaneously when the switch opens because of the relationship E=L di/dt, which says that the voltage across the inductor would have to be infinite to cause the current to fall to zero instantaneously. Therefore, one concludes that the current in the wire to the load does not fall to zero instantaneously and hence for a brief time there is an imbalance in the currents in the two wires passing through the transformer and the GFCI trips.

This analysis indicates that the GFCI trips only some of the time when a motor load is switched off. "
 

winnie

Senior Member
Location
Springfield, MA, USA
Occupation
Electric motor research
That 'inductance causing imbalance' theory doesn't fit with Kirchoff's laws. Current doesn't keep flowing in one wire and stop in the other, with a small exception: creation of fields and electromagnetic waves.

IMHO one should be looking for a (possibly unintentional) circuit causing the problem. Though to be open minded the problem could be some sort of RF effect.

-Jon
 

winnie

Senior Member
Location
Springfield, MA, USA
Occupation
Electric motor research
For DC inductive loads, it is common to place a diode across the load so that when the switch opens the current has a place to go. Obviously this won't work with an AC load, but one might be able to use an 'RC snubber'. Unfortunately I don't have the design experience to suggest proper component values.

-Jon
 

tersh

Senior Member
Location
new york
If you use surge protector to put between the shaded pole motor (in fridge) and GFCI outlet to protect against the inductive kick. Won't the surge protector MOV keeps getting hit? If the MOV has capacity of only 3000A surge current. How many inductive kick switchings would it take before the MOV reaches the limit and goes into thermal runaway?


Could the Siemens GFCI breaker not tripping be due to more MOV inside or a series surge filter which doesn't use MOV? Here there is a technology that doesn't use MOV.

https://zerosurge.com/

4K664T.jpg


Theoretically putting it between the fridge (with shade pole motor in the fan) and GFCI outlet can eliminate the inductive kicks?

Is there an oscilloscope that can shows the peak of the inductive kick, say the UL 6000v, 3000A 8/20 microsecond surge? What is usually the voltage, ampere peak of inductive kicks?
 

winnie

Senior Member
Location
Springfield, MA, USA
Occupation
Electric motor research
So it appears that packaged snubbers are sold specifically for AC motors. The design details are pretty sketchy; for example i'd expect the datasheet to specify the maximum motor load or inductance which could be served.

However if inductive kick from small motors is tripping GFCIs, then this might be a solution. I don't think you could buy a UL listed snubber that you can just install as a separate part of an electrical system; at most I think you will find UL recognized components that could go into a UL approvable design. You would specifically want something rated for 'across the line' use, meaning that if it fails it is designed to fail in a safe fashion.

Examples:
http://www.redlion.net/sites/default/files/213/3926/SNUB Product Manual.pdf
http://www.cde.com/resources/catalogs/Q-QRL.pdf

-Jon
 

tersh

Senior Member
Location
new york
So it appears that packaged snubbers are sold specifically for AC motors. The design details are pretty sketchy; for example i'd expect the datasheet to specify the maximum motor load or inductance which could be served.

However if inductive kick from small motors is tripping GFCIs, then this might be a solution. I don't think you could buy a UL listed snubber that you can just install as a separate part of an electrical system; at most I think you will find UL recognized components that could go into a UL approvable design. You would specifically want something rated for 'across the line' use, meaning that if it fails it is designed to fail in a safe fashion.

Examples:
http://www.redlion.net/sites/default/files/213/3926/SNUB Product Manual.pdf
http://www.cde.com/resources/catalogs/Q-QRL.pdf

-Jon



newreply.php


I tried to look for snubbers at amazon but couldn't find it. In the USA. What is the larger online retailers for electronics?

3 days ago I let an electrician installed a Siemens GFCI breaker in the old panel to test the fridge that regularly trips on the Meji GFCI receptacles.

newreply.php

newreply.php
newreply.php


dm1w01.jpg



N1ZWiI.jpg



You can see it's messy as a 220v-110v autotransformer is needed to power the Siemens 110v electronics. How could we do it to millions of homes. We don't have so much skilled electricians to do it.

Siemens engineering is still deliberating whether their PL loadcenter is compatible with the gfci breaker spring clips or studying their tolerances so I still couldn't install it in their PL loadcenter.

But last week I bought 6 Meiji GFCI receptacles because I decided that even without auto-monitoring self-test, it's better than another subpanel (as I was mulling how to protect the subpanel as all our homes in the Philippines don't have EGC (all the fridge, washing machines are only two prong)). So I thought pressing the self test manually each month would be safer than putting another panel with autotransformers. But I didn't expect the fridge could only trip on the Meiji and not the Siemens.

If I could find snubbers that work. May eliminate all the Siemens GFCI breakers.

However. I'd like to know one thing. In the USA. How come you don't put your bathroom heaters in the GFCI breakers? The heater is inside the bathroom or near it and the metallic faucet is connected to the heater via the water. Does this need GFCI protection? This was my last justification of needing the Siemens GFCI breakers, to protect our bathroom heaters. But is it necessarily? Are there owners in the USA who put their bathroom heaters in the GFCI breakers?
 
Last edited:

tersh

Senior Member
Location
new york
newreply.php


I tried to look for snubbers at amazon but couldn't find it. In the USA. What is the larger online retailers for electronics?

3 days ago I let an electrician installed a Siemens GFCI breaker in the old panel to test the fridge that regularly trips on the Meji GFCI receptacles.

newreply.php

newreply.php
newreply.php


dm1w01.jpg



N1ZWiI.jpg



You can see it's messy as a 220v-110v autotransformer is needed to power the Siemens 110v electronics. How could we do it to millions of homes. We don't have so much skilled electricians to do it.

Siemens engineering is still deliberating whether their PL loadcenter is compatible with the gfci breaker spring clips or studying their tolerances so I still couldn't install it in their PL loadcenter.

But last week I bought 6 Meiji GFCI receptacles because I decided that even without auto-monitoring self-test, it's better than another subpanel (as I was mulling how to protect the subpanel as all our homes in the Philippines don't have EGC (all the fridge, washing machines are only two prong)). So I thought pressing the self test manually each month would be safer than putting another panel with autotransformers. But I didn't expect the fridge could only trip on the Meiji and not the Siemens.

If I could find snubbers that work. May eliminate all the Siemens GFCI breakers.

However. I'd like to know one thing. In the USA. How come you don't put your bathroom heaters in the GFCI breakers? The heater is inside the bathroom or near it and the metallic faucet is connected to the heater via the water. Does this need GFCI protection? This was my last justification of needing the Siemens GFCI breakers, to protect our bathroom heaters. But is it necessarily? Are there owners in the USA who put their bathroom heaters in the GFCI breakers?

By the way. For the snubbers:

2y7j8U.jpg


Would it work by putting it outside the refrigerator, maybe by putting the snubber inside a box connected parallel to the 2 prong plug? Or does it have to be put inside the refrigerator in parallel to the shaded pole motor? If the latter is so, it's not an option because not only can it void the refrigerator warrantee. It would be very dangerous to add something to the circuit of the refrigerator even a super expert electrician would do it. I'd only trust the manufacturer to put it and not anyone else.
 
Location
NE (9.06 miles @5.9 Degrees from Winged Horses)
Occupation
EC - retired
newreply.php


I tried to look for snubbers at amazon but couldn't find it. In the USA. What is the larger online retailers for electronics?

3 days ago I let an electrician installed a Siemens GFCI breaker in the old panel to test the fridge that regularly trips on the Meji GFCI receptacles.

newreply.php

newreply.php
newreply.php


dm1w01.jpg



N1ZWiI.jpg



You can see it's messy as a 220v-110v autotransformer is needed to power the Siemens 110v electronics. How could we do it to millions of homes. We don't have so much skilled electricians to do it.

Siemens engineering is still deliberating whether their PL loadcenter is compatible with the gfci breaker spring clips or studying their tolerances so I still couldn't install it in their PL loadcenter.

But last week I bought 6 Meiji GFCI receptacles because I decided that even without auto-monitoring self-test, it's better than another subpanel (as I was mulling how to protect the subpanel as all our homes in the Philippines don't have EGC (all the fridge, washing machines are only two prong)). So I thought pressing the self test manually each month would be safer than putting another panel with autotransformers. But I didn't expect the fridge could only trip on the Meiji and not the Siemens.

If I could find snubbers that work. May eliminate all the Siemens GFCI breakers.

However. I'd like to know one thing. In the USA. How come you don't put your bathroom heaters in the GFCI breakers? The heater is inside the bathroom or near it and the metallic faucet is connected to the heater via the water. Does this need GFCI protection? This was my last justification of needing the Siemens GFCI breakers, to protect our bathroom heaters. But is it necessarily? Are there owners in the USA who put their bathroom heaters in the GFCI breakers?

We could put the heaters on a GFCI. We can put a GFCI on anything we choose. We don’t have to. The trick is we generally have a functional Equipment grounding conductor along with a neutral at each building. A properly installed EG reduces the shock hazard considerably. Protection by a GFCI does not eliminate the possibility of shock from an improperly grounded piece of equipment.
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
1901010-1116 EST

tersh:

As an electrical engineer you still just don't get down to basics. You need to think using basic knowledge. Too much of what you say just does not connect with your background.

Some basics you don't seem to consider are:

1. An MOV or other similar transient limiter is simply a not very good clamping device at some threshold voltage. This has no effect on the rate of rise of the transient voltage until limiting starts.

2. A snubber, a series RC circuit, won't reduce initial rate of rise unless there is external series impedance between the source and the snubber.

3. A filter circuit takes input energy and stretches that energy over time to reduce rate of rise, but doesen't change the amount of energy except thru internal losses.

You need to setup controlled experiments to allow you to find out how the transient conditions you have couple to the gate of the SCR. Certainly a good scope could help with the experiments. Also some basic understanding and judgements.

We are guessing that an inductive circuit with current flowing at the time of turning off the current is the source of a transient that trips your GFCI . This interruption can occur both when you close a switch, contact bounce, as well as when you open the switch.

For the purpose of determining how tripping is occurring the NEC has nothing to do with your experiment. When you find the cause, then you consider how to solve the problem within the NEC rules.

.

.
 

tersh

Senior Member
Location
new york
We could put the heaters on a GFCI. We can put a GFCI on anything we choose. We don’t have to. The trick is we generally have a functional Equipment grounding conductor along with a neutral at each building. A properly installed EG reduces the shock hazard considerably. Protection by a GFCI does not eliminate the possibility of shock from an improperly grounded piece of equipment.

In the Philippines. All of our appliances (Refrigerators, Washing Machines, Heaters), etc. have no grounding or EGC. All the plugs are two prong. For example. This is the heater we commonly used in our bathroom.


MeJBnc.jpg


The plug is 2 prong made in Malaysia. Available at https://wschoo.en.ecplaza.net/products/water-heater-vizz-colour-series_646890 It's multi-point heater 27Ampere.

What do you mean "Protection by a GFCI does not eliminate the possibility of shock from an improperly grounded piece of equipment". Can you give example? Since the Multipoint heater doesn't have EGC. I guess putting it to the Siemens GFCI breaker would be great idea (rather than nothing). The Meji FM2141 (used by Leviton) based GFCI is only up to 20A. The heater is 27A. Millions of homes installed it that way without any EGC or GFCI. 99.99% of filipinos don't know the meaning of GFCI.
 

LarryFine

Master Electrician Electric Contractor Richmond VA
Location
Henrico County, VA
Occupation
Electrical Contractor
I don't know whom you're quoting, but a GFCI can't protect against a line-to-neutral shock.

A shock is required for a GFCI device to function; it's designed to prevent electrocution.
 

tersh

Senior Member
Location
new york
I don't know whom you're quoting, but a GFCI can't protect against a line-to-neutral shock.

A shock is required for a GFCI device to function; it's designed to prevent electrocution.

Neither can an EGC protect against a line-to-neutral shock which is equivalent to putting your fingers at the two holes of outlets. Of course this is extreme that happens rarely. Your EGC only protects any hot touching the metal chassis which trips the breakers. It doesn't protect against line-to-neutral shock.

In our country. Since we don't use 120v neutral as I detailed before, and we used your USA equivalent of red and black line to line of 240v. Then I know a GFCI doesn't protect against a line-to-line shock. Neither does any EGC.

But with GFCI guaranteed to trip at 5mA. It can protect even against shock and electrocution (provided only one line touched against any ground (natural concrete or EG or GEC), isn't it. Shock is when the current is above 10mA. I was able to verify that the Meiji outlets trip at 5mA by using this equipment at their head office (it's the technican fingers holding the GFCI).

8DWxnW.png
 

tersh

Senior Member
Location
new york
I used a ‘snubber’ on the bath fan that was giving me grief about 9 years ago. Not a problem since.

IIRC, it didn’t matter whose GFCI device I used. Different circuits. No interconnection until the neutral bar.

When the following is put between the shaded pole motor and the GFCI outlet (with FM2141 chip used also by Leviton). There are no more trips after repeated switchings (whereas with it removed, it trips 4 out of 5 times).

KT1Uon.jpg


The red stuff is capacitor with rating of 0.0042 uF.
The coils are wound on a ferrite (semi-metallic) core.
It seems to be LC filter type, designed to mainly filter out electrical interference.

But something confuses me.

zhusvf.jpg


The cover said "Voltage Surge Protector". There is no MOV (metal oxide varistors). Do you considered LC filter as surge protector too? It's made locally in the Philippines and I can't find it anymore. Here we don't have any UL, CE or don't consider them before accepting it. Is it a fake surge protector when it's just an LC filter?

What kinds of surges can LC filter remove versus that of MOV? Sorry I forgot my lessons 25 years ago. We were not taught about MOV. So I'm quite confused especially with the paradoxical labelling above.. Thank you.
 
Location
NE (9.06 miles @5.9 Degrees from Winged Horses)
Occupation
EC - retired
An equipment ground during a fault may only carry millamps. It may carry hundreds of amps and it may carry that current forever, dependent on the size of the overcurrent protection ahead of it. A 16 amp fault to the tank of the water heater and Equipment Ground will not trip a 20 amp breaker. As far as the breaker is concerned that is just part of the load. A properly installed EG reduces the shock hazard should some one touch the water heater. You may not even feel one. The potential is reduced. Remove the EG and you have full potential. Add a GFCI to an EG and the current flow is reduced to the 5ma on the EG. Remove the EG and the potential rises. You know this.

Your definition of shock appears to be what we call electrocution, ie death.
 

tersh

Senior Member
Location
new york
No, unless the surge is only of frequencies above the filter's cutoff.

What components of inductive kick can be filtered by LC filter? Does inductive kick produce any RF components that can be filtered by LC filter? I'm wondering this because since the LC filter eliminate the trips of the shaped pole motor, then is it really inductive kick causing the GFCI trips or RF?

Do you know of a web site that summarizes the waveforms of inductive kick, RF interference filtered by LC filter, Surges, RC snubber, (what else) and how they differ between one another?
 

tersh

Senior Member
Location
new york
An equipment ground during a fault may only carry millamps. It may carry hundreds of amps and it may carry that current forever, dependent on the size of the overcurrent protection ahead of it. A 16 amp fault to the tank of the water heater and Equipment Ground will not trip a 20 amp breaker. As far as the breaker is concerned that is just part of the load. A properly installed EG reduces the shock hazard should some one touch the water heater. You may not even feel one. The potential is reduced. Remove the EG and you have full potential. Add a GFCI to an EG and the current flow is reduced to the 5ma on the EG. Remove the EG and the potential rises. You know this.

Your definition of shock appears to be what we call electrocution, ie death.

I was just saying the discoverer of GFCI tested different mA on people. At 5mA, one would get tinging. At 10mA (?), one would begin to get shocked.

Btw. If you don't have EG (Equipment Ground). Remember all our appliances in the country don't have Equipment Ground or EGC. I'd like to know how fast is the response time of a 5mA GFCI? If a person got accidentally exposed to one of the live wire with feet on the concrete floor. And there is potentially a 10A exposure, how fast can the 5mA GFCI relay opens and how much current would still pass through him? If the relay open longer than the current moving through it, then can you still get electrocuted? If so, then what is the use of GFCI?
 

gar

Senior Member
Location
Ann Arbor, Michigan
Occupation
EE
190111-0830 EST

tersh:

You continue to have false impressions because you do not go back to basics.

Look at the GFCI datasheet for the trip time characteristic. There is an inverse time relationship, electronic in this case, relative to current. The 5 mA is after a moderately long time relative to how little time it takes for you to detect a shock. Further you can detect a shock at a much lower level than 5 mA.

See https://pdfs.semanticscholar.org/3687/65b32d61fb81143356f034aaea68f166bb3d.pdf
1 mA looks to be a better level for detectability for ordinary people. An interesting side note is that John Swets is referenced in this discussion. I first met him in 1953 when I was in a psychology class taught by Wilson (Spike) P. Tanner. Tanner and Swets were both PhD students at the time working on signal detectability, and running vision experiments. I became a subject, about $1 per hour. That summer I started building test equipment for Tanner's experiments.

Note that a 1 mA shock could kill you indirectly. You stand on a ladder, get shocked, fall off, crack your head on the floor, and die.

At 5 mA you will feel a substantial shock.

What is a surge, transient, impulse, RF, etc. Their meanings are all dependent upon context.

More later possibly.

.
 
Location
NE (9.06 miles @5.9 Degrees from Winged Horses)
Occupation
EC - retired
I was just saying the discoverer of GFCI tested different mA on people. At 5mA, one would get tinging. At 10mA (?), one would begin to get shocked.I have talked to people that have tripped a GFCI through their body. It was not a tingle. Try it yourself if you like but remember GFCIs have a lot of footnotes about the general health, age, and conditions of the recipient surviving said 5ma tingle.

Btw. If you don't have EG (Equipment Ground). Remember all our appliances in the country don't have Equipment Ground or EGC. I'd like to know how fast is the response time of a 5mA GFCI? If a person got accidentally exposed to one of the live wire with feet on the concrete floor. And there is potentially a 10A exposure, how fast can the 5mA GFCI relay opens and how much current would still pass through him? If the relay open longer than the current moving through it, then can you still get electrocuted? If so, then what is the use of GFCI?

The GFCI is used to limit your exposure to potentially harmful or lethal currents. It does not prevent a shock especially when no EG is used. It may prevent electrocution. Change your terms.
 
Status
Not open for further replies.
Top