GFCI 2-wire Experiment

[jim dungar]

...in which OCPD and GFCI (in context of the experiment) would just laugh and watch you frizzle to death.

It appears that your hypotheses is 'GFCIs do not protect against people for being electrocuted by all electrical faults'.
Your proof consists of creating an experiment that demonstrates that current flows through water from a 'hot' conductor to a 'neutral' conductor (I could not tell if an actual GFCI was involved).
And you imply that many of the forum members, or just those that disagree with your proof, are ignorant of electrical theory.

 

[FionaZuppa]

btw, if you want to convince us GFCIs are useless, an experiment (actual, not theoretical) like mbrooke posted many months back showing an instance of a failed (read: FAILED) GFCI still allowing current to pass would be much more convincing. In that, I believe the ungrounded contacts had welded together, so it tripped and reset by the buttons, but in a tripped state allowed current to pass. GFCIs are not failsafe, however they are hardly useless. Aside from possibly double insulation, GFCIs are probably the best safety item to protect the end user from shock that the electrical industry has seen to date.

so, this is not about "useless" or "failed" GFCI btw. in the context of my experiment, the GFCI was "useless" in the way 1- cap cord is used (the GFCI cannot provide the hazard coverage). remember, mbrooke eluded to why outside US 30mA is std, if you tell folks you are safer by swapping 1- outlets for a 5- GFCI that has no EGC you are creating false notion as to how much safer it will be. if i see 5- GFCI i expect EGC to be there, and when i plug in 5- cap cord into 5- GFCI i expect the EGC to be there for that added hazard coverage. so now go back to post #1, thats the whole point here. imho, if there is only 2-wire then the GFCI should be a 1-R receptacle, not a 5-R with a missing EGC.

It appears that your hypotheses is 'GFCIs do not protect against people for being electrocuted by all electrical faults'.
Your proof consists of creating an experiment that demonstrates that current flows through water from a 'hot' conductor to a 'neutral' conductor (I could not tell if an actual GFCI was involved).
And you imply that many of the forum members, or just those that disagree with your proof, are ignorant of electrical theory.

almost, let me correct this for clarity. 1) i did make note that my cord was in a GFCI. 2) much of the argument was around people making notion that the amps hitting the water will not conduct via any of the infinite paths that exist in a liquid conductor, hit the body, traverse the body, and back to the wire. there is a real hazard there, etc. so yesssssss, its a line to N issue, but not by way of user grabbing onto the line and N wires.

folks who disagree with me? well, bring on the disagreements, but 1st understand what the experiment shows. i found 1 valid disagreement, which i fixed thanks to someone actually looking at the data and scratching their head because he applied some math. i still dont think anyone understands that the ~18k inline R is meaningless in the experiment.

 

[Ingenieur]

ah, nice catch.

for clarity
R inline is ~18,115
amps of wire in water was 0.006601A (6.601mA)
this tells us the gap presented an additional ~518.54 ohms

does that change the experiment in your view?

i mentioned water quality in post #1. and yes, your point of getting hit with amps is exactly the point, in which OCPD and GFCI (in context of the experiment) would just laugh and watch you frizzle to death.

with you numbers
R total = 123/0.0066 = 18636 Ohm
R water = 18636 - 18115 = 521 Ohm
water v drop = 521 x 0.0066 = 3.5 V

although your point is doesn't trip
this is true

total I w/o limiting R = 120/521 ~ 230 mA
but in water, even with 120 across the terminals, only a small fraction would flow across the person
body plus water vs the water between the terminals, no way to estimate accurately
so in this case the person may see 30-100 mA (say person is 1000 and water path thru person another 1000 since it will be >>> than the terminal path distance, I person = 521/(2000) x 230 ~ 60 mA)
so yes, there are situations where someone falls thru the gap, but you will find 100's that are saved
if he is on the ground, it will trip

 

[FionaZuppa]

with you numbers
R total = 123/0.0066 = 18636 Ohm
R water = 18636 - 18115 = 521 Ohm
water v drop = 521 x 0.0066 = 3.5 V

although your point is doesn't trip
this is true

total I w/o limiting R = 120/521 ~ 230 mA
but in water, even with 120 across the terminals, only a small fraction would flow across the person
body plus water vs the water between the terminals, no way to estimate accurately
so in this case the person may see 30-100 mA (say person is 1000 and water path thru person another 1000 since it will be >>> than the terminal path distance, I person = 521/(2000) x 230 ~ 60 mA)
so yes, there are situations where someone falls thru the gap, but you will find 100's that are saved
if he is on the ground, it will trip

with a logical 1-R GFCI and many scenarios where fault current to an EGC may be way lower than line-N fault amps, the GFCI is "useless". lets take the scenario where the hairdryer falls in on top of the person and is 4ft away from the EGC of the metal drain piping, you think you'll get 6mA out to the EGC? maybe, maybe not. will you get 6+mA to cross the water, through the body, back to wire in dryer? very possibly yes since the dryer is in very close proximity to the body in the water.

if i were to re-run same experiment but use a 5- cap cord, i pretty sure GFI would pop.

hence, the reason why 2-wire GFCI does not cover all hazards, some potentially deadly. hence, the reason why all GFCI should be 3-wire. hence, the reason why all "bathroom" appliance should be NEMA 5- !!!

in other words, dont make a 5-R GFCI into a 5-R no EGC (a logical 1-R).

 

[mbrooke]

btw, if you want to convince us GFCIs are useless, an experiment (actual, not theoretical) like mbrooke posted many months back showing an instance of a failed (read: FAILED) GFCI still allowing current to pass would be much more convincing. In that, I believe the ungrounded contacts had welded together, so it tripped and reset by the buttons, but in a tripped state allowed current to pass. GFCIs are not failsafe, however they are hardly useless. Aside from possibly double insulation, GFCIs are probably the best safety item to protect the end user from shock that the electrical industry has seen to date.

Yup, here is the thread:

http://forums.mikeholt.com/showthread.php?t=178511


Leviton actually took it apart and officially gave word on its failure (welded hot contact).

 

[JFletcher]

so, this is not about "useless" or "failed" GFCI btw. in the context of my experiment, the GFCI was "useless" in the way 1- cap cord is used (the GFCI cannot provide the hazard coverage). remember, mbrooke eluded to why outside US 30mA is std, if you tell folks you are safer by swapping 1- outlets for a 5- GFCI that has no EGC you are creating false notion as to how much safer it will be. if i see 5- GFCI i expect EGC to be there, and when i plug in 5- cap cord into 5- GFCI i expect the EGC to be there for that added hazard coverage. so now go back to post #1, thats the whole point here. imho, if there is only 2-wire then the GFCI should be a 1-R receptacle, not a 5-R with a missing EGC.

Okay, let me try it this way. GFCI receptacle connected to 2 wires. Lets assume for a moment it works and at 6ma. I come by and step barefoot on a frayed cord (2 or 3 wire, ungrounded conductor only is frayed) plugged into it. One of two things will happen: if I am electrically isolated from ground or anything that is grounded (lets say Im stepping on a rubber mat), I am not getting shocked. Period.

The second possibility is that I am grounded, lets just, for argument's sake, say I am sticking my tongue on the end of a copper 4/0 cable going back to the panel's ground buss. iow, as grounded as possible. Now, I only have my body's resistance, which will be quite low for a wet tongue (and lets say a wet foot too), to regulate the current passing thru me (and probably my heart). That current is going from ungrounded, thru me, to another ground, not thru the neutral. The GFCI will trip.

Are you arguing that a 3 prong appliance like a washing machine is less safe in a 2 wire GFCI receptacle than on a 3 prong w/o GFCI protection? 3 prong receptacles on 2 wire circuits are dangerous because there is a false sense of security there (they are also not allowed by the NEC). a 3 prong GFCI on a 2 wire circuit is not unsafe; the degree of safety improvement one could reasonably advertise would be dependent on duplicating faults of every kind, or case studies. Not one sink experiment.

It doesnt matter if there is an EGC or not. If there is a refrigerator plugged into a two wire GFCI receptacle, and there were a fault to the frame, leaving it energized at 120V, when I touch it, I am either touching something else conductive or not. The fault current either travels thru me, not the neutral, or doesnt. If the fault is >6ma, the receptacle trips, EGC or not.

Your arguments re line faults are irrelevant; they are GFCI, not LFCI. Most L-N faults will trip OCPD. Yes, there are cases where if you get in the way of a L-N fault, a GFCI wouldnt trip, and it could kill you. an EGC would be absolutely worthless in those cases too.

 

[FionaZuppa]

Your arguments re line faults are irrelevant; they are GFCI, not LFCI. Most L-N faults will trip OCPD. Yes, there are cases where if you get in the way of a L-N fault, a GFCI wouldnt trip, and it could kill you. an EGC would be absolutely worthless in those cases too.

you say "most L-N faults will trip OCPD", well, i guess my crude experiment doesnt fall into your "most" category. in my experiment it doesnt (even if you take out that 18k inline), OCPD is perfectly happy supplying amps to a faulty ckt. what type of fault is it when you have current flowing via non-intended path(s), or potentially a hazard path? why does everyone think a current fault has to be back to EGC? the experiment i have shown is parallel to L-N path in the water, not in series with L or N, etc. from the dryer perspective, the switch is in off position yet there is amps on the wire, is this not a fault of some type?

the fridge frame is something different, its not a liquid conductor commonly found in all resi bathrooms :thumbsup:

 

[K8MHZ]

For those of you that got popcorn, enjoy this.

Kids, don't try this at home!!!! (This guy does not understand electricity)

 

[FionaZuppa]

For those of you that got popcorn, enjoy this.

Kids, don't try this at home!!!! (This guy does not understand electricity)

For those of you that got "NEC" popcorn, this one is way better. why you say he doesnt understand? what doesnt he understand?

 

[K8MHZ]

why you say he doesnt understand? what doesnt he understand?

 

[JFletcher]

Here's the bottom line: if you got your wish of 2 prong GFCI receptacles, the first time someone goes to plug in a 3 prong plug, they are going to use an adapter or change the receptacle to a 3 prong GFCI on a 2 wire circuit. There are stickers that cover NO EGC CONNECTED to 3 wire GFCI on 2 wire circuits.

 

[kwired]

Fiona,

First let me say I only read the first 40 or so posts then just skimmed over the rest to this point, so I apologize if I missed something other then people thumbing noses at each other in the previous 40 or so posts.

I am assuming you did this experiment because of a discussion that started in another thread, and I recall being the one that asked you to try such an experiment. I don't feel you succeeded at simulating the conditions that were being discussed in that thread though.

In that thread (from the best of what I can recall) you were claiming a person sitting in an isolated (from ground) tub of water would be shocked, and probably at lethal levels, if an energized appliance were dropped in the tub. I think most following there agreed that if the tub were grounded that a GFCI would likely trip, but your main claim that prompted the asking of you to do some experimenting was that in the isolated tub that current would flow out one supply lead, through the water, through the victim, back through the water and to the other supply lead (presumably at reasonably close distance from the first supply lead.

I told you something to the effect that current will take all paths, but also would be relative to the resistance of each path. I also said the resistance from L1, through water, through victim, back through more water, and into L2 would be significantly high enough that current in the victim would be very low compared to current flowing in the short distance of water directly between L1 and L2. This all assuming you are in an insulated tub and there is no current to any grounded objects outside of what gets introduced within the appliance.

If you want to have a more accurate simulation and measure how much current the victim might see, get rid of the 18k resistor (it won't be there in the real world, and is a current limiter for what you have set up or for what I will suggest. you can keep your power supply cord, maybe fix the ends so they can't touch one another but still keep them close together, anywhere from your fraction of an inch up to 3 or 4 inches apart, or maybe even try both ways. You then need to place something in the water to simulate the victim. placing it series in the cord is not the same place the victim would be, placing it in series in the cord does assure the current does flow through the victim, the question was how much if any current will flow through the victim.

Maybe use a ham as a victim, cured and salted probably would be lower resistance then an actual person with no skin abrasions.

put meter probes on opposite ends of the ham. Not really sure how you can accurately determine current as there would be many potential paths through it just like a person, you could start with it still wrapped and put a metal probe in each end of some sort that would be more less an electrode between water and ham and measure current entering the probe(s) - not quite a true simulation but will let you know what the current is for that situation. Otherwise you could just probe for voltage across various positions on the ham, if you have current you must have a drop in voltage across whatever it is flowing through. If you never find any voltage I think it is safe to say the vicitim is not being electrocuted. Are they still in a dangerous situation - yes, any move that makes contact with enough difference in potential (maybe even across voltage gradient in the water only) it can be lethal, but as long as they remain in same position they are likely safe.

This is just with a relatively small appliance being dropped in the tub. Put voltage leads at opposite ends of the tub and you do put the victim right in the path between the points of voltage. In that case the more conductive the water is the better chance the victim has of not being a significant enough path to be electrocuted, they likely will have some current flow through them no matter what though.

 

[FionaZuppa]

Here's the bottom line: if you got your wish of 2 prong GFCI receptacles, the first time someone goes to plug in a 3 prong plug, they are going to use an adapter or change the receptacle to a 3 prong GFCI on a 2 wire circuit. There are stickers that cover NO EGC CONNECTED to 3 wire GFCI on 2 wire circuits.

we went over this a zillion times already, this exact statement of yours.

so lets think about the scenario with a tad more meat, shall we.


an old house that has old occupants, for last 52yrs in the home they never had any 5-R's (EVER), its all 2-wire crud. ok, so we can only imagine that for many many years they have used cheaters and broken EGC pins off, yes? ok, well, now you come along and swap out a 1-R with a non-EGC'd 5-R GFCI and place a sticker on it that says "no EGC". great, "20 more then" <-- sarcasm btw. the folks are blind as a bat but notice now there's this 3rd hole in the receptacle, they really have no idea what that is, but what they might know is, must be safer to just stop breaking off the EGC pin of items they buy, because you told them the GFCI is safer. well, they are indeed safer, but not because of the missing EGC, yet for some reason they believe that extra pin makes them safer, but in the realm of reality the 5- GFCI (now a logical 1-) does not provide all of the hazard coverage a EGC'd 5- GFCI does. given the context as to why anyone is swapping out a 1-R with a 5-R GFCI, why provide the EGC pin? imho, block it out, or, provide EGC. for those who thought the amps would only flow between the two terminals of the appliance when it hots the water, thats a big WRONG as many of us know it, but some insisted no amps would flow across the unattached amp probes, and what did we find, amps in fact do flow, and in some scenarios the amps through unattached probes can be greater that the trip specs of GFCI class-A !

Fiona,

First let me say I only read the first 40 or so posts then just skimmed over the rest to this point, so I apologize if I missed something other then people thumbing noses at each other in the previous 40 or so posts.

I am assuming you did this experiment because of a discussion that started in another thread, and I recall being the one that asked you to try such an experiment. I don't feel you succeeded at simulating the conditions that were being discussed in that thread though.

In that thread (from the best of what I can recall) you were claiming a person sitting in an isolated (from ground) tub of water would be shocked, and probably at lethal levels, if an energized appliance were dropped in the tub. I think most following there agreed that if the tub were grounded that a GFCI would likely trip, but your main claim that prompted the asking of you to do some experimenting was that in the isolated tub that current would flow out one supply lead, through the water, through the victim, back through the water and to the other supply lead (presumably at reasonably close distance from the first supply lead.

If you want to have a more accurate simulation and measure how much current the victim might see, get rid of the 18k resistor (it won't be there in the real world, and is a current limiter for what you have set up or for what I will suggest.

yes, its from the other thread that went south, so i provided some real experimental data.
you may have missed my reply in this thread about your comment in red. if the EGC is too far away (could be 4ft in a tub), the ohms across 4ft of bath water may be less than GFCI trip amps, so for those who agreed in other post, perhaps they were missing some data?

as for the part i bolded in purple, the inline R does not matter. if you believe it matters then please explain in detail why it matters. the inline R is a safety device, thats all it is. everything is linear in terms of V and A. if someone want to dupe my experiment w/o the inline R just to prove that it is linear, then have at it.

 

[user 100]

Here's the bottom line: if you got your wish of 2 prong GFCI receptacles, the first time someone goes to plug in a 3 prong plug, they are going to use an adapter or change the receptacle to a 3 prong GFCI on a 2 wire circuit. There are stickers that cover NO EGC CONNECTED to 3 wire GFCI on 2 wire circuits.

we went over this a zillion times already, this exact statement of yours.

so lets think about the scenario with a tad more meat, shall we.


an old house that has old occupants, for last 52yrs in the home they never had any 5-R's (EVER), its all 2-wire crud. ok, so we can only imagine that for many many years they have used cheaters and broken EGC pins off, yes? ok, well, now you come along and swap out a 1-R with a non-EGC'd 5-R GFCI and place a sticker on it that says "no EGC". great, "20 more then" <-- sarcasm btw. the folks are blind as a bat but notice now there's this 3rd hole in the receptacle, they really have no idea what that is, but what they might know is, must be safer to just stop breaking off the EGC pin of items they buy, because you told them the GFCI is safer. well, they are indeed safer, but not because of the missing EGC, yet for some reason they believe that extra pin makes them safer, but in the realm of reality the 5- GFCI (now a logical 1-) does not provide all of the hazard coverage a EGC'd 5- GFCI does. given the context as to why anyone is swapping out a 1-R with a 5-R GFCI, why provide the EGC pin? imho, block it out, or, provide EGC.

And still, they are going to circumvent your 1-15 gfci with that cheater or remove a ground prong- there is no plausible benefit to having a 1-15 gfci......

You're basically saying that its best for people to be forced to circumvent something all b/c you believe that when we do the 406 2w replacement option, we may be perpetuating a "false sense of security" with respect to egcs, by providing a egc free GFCI receptacle that will accommodate 5-15 plugs, nevermind the fact that people will be protected from electrocution should they ignore the stickers/ violate 250.114 and something goes awry........:happyno:

 

[FionaZuppa]

And still, they are going to circumvent your 1-15 gfci with that cheater or remove a ground prong- there is no plausible benefit to having a 1-15 gfci......

You're basically saying that its best for people to be forced to circumvent something all b/c you believe that when we do the 406 2w replacement option, we may be perpetuating a "false sense of security" with respect to egcs, by providing a egc free GFCI receptacle that will accommodate 5-15 plugs, nevermind the fact that people will be protected from electrocution should they ignore the stickers/ violate 250.114 and something goes awry........:happyno:

you have to look at it the other way. stop being an electrician for just a minute.

if someone has 5-P and cant plug that in, they go "hmmm, wtf, why is that, my plug has this 3rd pin on it, what do i do (run to HD and ask nim-wad who advises to not break pin off but gives a cheater)". <-- ok, that was some asserted effort there, but it got the person thinking about why there is a diff, the plug just didnt fit the 1-R.

ok, option two, 5-R GFCI w/o EGC "hmm, plug right in, i am all good". <-- nope, not all good, w/o the EGC there that GFCI does not cover all the hazards that a 5-R GFCI should be covering, yet the user didnt for 1sec question plugging in the 5-P cap cord, and has no clue what the sticker means (if a sticker is even there any longer).

and in a 2-wire house where you do a swap out of 1-R to a 5-R GFCI w/o EGC, what does it matter if they use cheater or break EGC pins off, there is no EGC, right? breaking pin off at that point does not create more hazard, right? (well, breaking pin off may be bad if the two blades are not polarized, so a cheater is really best option, HD guy did good).

 

[kwired]

you have to look at it the other way. stop being an electrician for just a minute.

if someone has 5-P and cant plug that in, they go "hmmm, wtf, why is that, my plug has this 3rd pin on it, what do i do (run to HD and ask nim-wad who advises to not break pin off but gives a cheater)". <-- ok, that was some asserted effort there, but it got the person thinking about why there is a diff, the plug just didnt fit the 1-R.

ok, option two, 5-R GFCI w/o EGC "hmm, plug right in, i am all good". <-- nope, not all good, w/o the EGC there that GFCI does not cover all the hazards that a 5-R GFCI should be covering, yet the user didnt for 1sec question plugging in the 5-P cap cord, and has no clue what the sticker means (if a sticker is even there any longer).

and in a 2-wire house where you do a swap out of 1-R to a 5-R GFCI w/o EGC, what does it matter if they use cheater or break EGC pins off, there is no EGC, right? breaking pin off at that point does not create more hazard, right? (well, breaking pin off may be bad if the two blades are not polarized, so a cheater is really best option, HD guy did good).

There is no EGC in the tub of water if you only have a two wire corded appliance whether it is plugged into a grounding type receptacle or not. I don't see what the presence or absence of the EGC has to do with your experiment unless you compare bringing just the two current carrying lines to the tub of water to bringing all three lines of the receptacle to the water, you can also do both with or without the GFCI if you want, the experiment though was supposed to be how does current flow in an isolated tub of water with a live cord/appliance dropped into the tub - you keep throwing all sorts of other things at this that are outside the scope of where does current flow in the tub of water.

 

[FionaZuppa]

There is no EGC in the tub of water if you only have a two wire corded appliance whether it is plugged into a grounding type receptacle or not. I don't see what the presence or absence of the EGC has to do with your experiment unless you compare bringing just the two current carrying lines to the tub of water to bringing all three lines of the receptacle to the water, you can also do both with or without the GFCI if you want, the experiment though was supposed to be how does current flow in an isolated tub of water with a live cord/appliance dropped into the tub - you keep throwing all sorts of other things at this that are outside the scope of where does current flow in the tub of water.

wow, very frustrating.
there may be EGC in the tub, depends on what metal is around and what it may be tied to, etc. even with EGC in the tub somewhere, the non-EGC'd 5-R GFCI may not protect as expected.

so if it were a 5-P connected to my 5-R GFCI with EGC, how do you imagine the amps would flow in the water?? you'll get fault amps back onto the EGC. the EGC doesnt even need to be connected to anything in the appliance, just needs to be exposed and close-by to the line inside the appliance (think all plastic do-wad appliance). however, if it were a 5-P plugged into one of your 1-R swap-outs, that 3rd wire in the cord cant help you, not because of the wire, because the GFCI just cannot provide that extra coverage that everyone thinks they get when they see an outlet with funny buttons in the middle of it.

 

[kwired]

wow, very frustrating.
there may be EGC in the tub, depends on what metal is around and what it may be tied to, etc. even with EGC in the tub somewhere, the non-EGC'd 5-R GFCI may not protect as expected.

so if it were a 5-P connected to my 5-R GFCI with EGC, how do you imagine the amps would flow in the water?? you'll get fault amps back onto the EGC. the EGC doesnt even need to be connected to anything in the appliance, just needs to be exposed and close-by to the line inside the appliance (think all plastic do-wad appliance). however, if it were a 5-P plugged into one of your 1-R swap-outs, that 3rd wire in the cord cant help you, not because of the wire, because the GFCI just cannot provide that extra coverage that everyone thinks they get when they see an outlet with funny buttons in the middle of it.

What EGC? From the beginning of this in the other thread, we were presuming no grounded surfaces present in an electrically isolated tub. Throwing in a grounded object for the potential victim to contact or for current to pass to via the water changes all the parameters of the experiment, and if you do have such current flowing outside the GFCI protected conductors that is more then the trip threshold - you will trip the GFCI whether there is a EGC present or not at the GFCI location. Unless you have water with no impurities chances are with a 120 volt source the water is conductive enough to carry the necessary 4-6 mA necessary to trip the GFCI.

To prove the point you have been trying to make ever since the other thread that prompted me to ask you to do such experimenting, you need to prove that the user in an electrically isolated tub of water would be shocked if you dropped an energized appliance (with no EGC in the supply cord) in that tub. So far you have not proven that one way or another, you measured some current in the water, you need to measure current in a simulated victim to prove anything, or at least a voltage being imposed on the victim which would leave a possibility of current flowing because of that voltage. If the water around said victim has less resistance then the victim, the bulk of current that flows will flow around the victim, how much flows through the victim will depend on how much voltage is across the victim and what their resistance is. Then there is just the possibility they do feel a shock in just a leg or something but never have current flowing through the right places to stop the heart, that possibility is increased even more if just standing knee deep in the water vs being nearly completely immersed.

 

[FionaZuppa]

you measured some current in the water, you need to measure current in a simulated victim to prove anything,

actually, no i did not. if i were to measure amps of water one end of the wire would be directly connected to one amp probe.
the amp probes (less whatever body ohms you may hypothesize would be inline with probe) is the body in the experiment.

 

[Ingenieur]

actually, no i did not. if i were to measure amps of water one end of the wire would be directly connected to one amp probe.
the amp probes (less whatever body ohms you may hypothesize would be inline with probe) is the body in the experiment.

but if the probes were say 100 times further apart to allow a body to fit
say 28"
would not the Z go up by L/A times?
Say 520 Ohm (at 0.28") what would it be then?