Searching \ for '[EE] GFCI PROBLEM with lightening' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=gfci+problem+with
Search entire site for: 'GFCI PROBLEM with lightening'.

Exact match. Not showing close matches.
PICList Thread
'[EE] GFCI PROBLEM with lightening'
2008\05\31@143949 by Carl Denk

flavicon
face
For those not familiar a GFCI (Ground fault circuit interrupter) is
required by most building codes on circuits that could be near a
grounding point reachable by the human person extremities. This includes
near a water source like kitchen and bathroom, and exterior outlets. The
GFCI monitors the outgoing (black wire) and incoming or common (white
wire) current, and if there is a difference, in the 6 to 30 ma. range as
a trip point, the GFCI will open up the circuit. The idea being, if a
human is receiving current, then it is not returning to the GFCI, and an
imbalance of amperage is detected. This all happens very quickly where
in general the human life is spared.]

Now for my situation. A natural gas well is located 200' from the house,
and a single 120 volt 20 amp circuit goes there via. a 3 wire (black
hot, white common, and green ground) 12 gauge THHWN stranded wire in
3/4" plastic underground electrical conduit. The wire is new a year ago,
but had the problem before that for 30 years. The circuit is only
grounded via the green wire at the main breaker panel, and not at the
well end either with a ground wire or the 120' deep 8" steel well
casing. Load is a .5 amp, 6 volt wall wart with no ground terminal
powering a PIC18F1320 that monitors the well pressures and enclosure
temperature which is communicated back to the house via fiber optic
cable in the conduit.

Occasionly several times a year when a thunderstorm passes near with
lightening strikes maybe several miles away, the GFCI opens the circuit.
Resetting then stays in for months until possibly another storm passes
near. There may or may not be also on the circuit, 2 heat traces that
have the usual 120 volt grounded plugs. The heat traces are Raychem mfg.
self regulating temperature type, and a small resistance heater with fan
of the under desk type. All this is in a weatherproof enclosure. I have
thoroughly checked the circuits and loads for even the slightest short
to ground with a DMM on milliohm scale. There was no sign of the
slightest leakage.

Have thought about grounding the circuit to the well casing (think
that's a much more effect than a ground rod. The soil conditions are a
firm clay. Any thoughts??

2008\05\31@152023 by Carl Denk

flavicon
face
Forgot to note: The GFCI was replaced with a new one 6 weeks ago.

Carl Denk wrote:
{Quote hidden}

2008\05\31@154514 by Dr Skip

picon face
I have noted that outdoor, remote circuits like this tend to trip because of
water in the conduit, outlet box, or somewhere. When lightening strikes, is it
also raining?

I suspect that in real use, even a little condensation somewhere does it.

I would try replacing the breaker with a standard one, and putting a GFCI
outlet at the pump if it's all enclosed and housed. I believe that will satisfy
the code requirements too. If it goes away, it was probably water in the
conduit. I like the idea of grounding to the well, and it may fix things for
you if it is EM related, but that may not be approved. HOWEVER, if the pump
pumps to the house through metal pipe, then it would be OK IMHO - just bond the
well to the pipe electrically (not just through the pump) and to the house
ground as probably already exists.

To test, find another GFCI outlet at the house and run an extension cord out
there and put various combinations of load on it. See what triggers it if
anything. It will be more receptive to interference being above ground, so if
IT doesn't trip, then it's the line.

And if it isn't critical, just a pain to go out to check it only to find the
breaker tripped, run some sensor wires out there or an IR link to an indicator
at the house so you know before you need to know. ;)


Carl Denk wrote:
> Forgot to note: The GFCI was replaced with a new one 6 weeks ago.
>
> Carl Denk wrote:

2008\05\31@154913 by Dave Tweed

face
flavicon
face
Carl Denk wrote:
> Now for my situation. A natural gas well is located 200' from the house,
                         ^^^^^^^^^^^ ???

Seriously? I've never heard of a private home with its own gas well before.

However, from the rest of your post, it sounds like you're really talking
about a water well. (Why else would you have heat tapes?)

> Occasionly several times a year when a thunderstorm passes near with
> lightening strikes maybe several miles away, the GFCI opens the circuit.
> Resetting then stays in for months until possibly another storm passes
> near.

I think that with nonmetallic conduit, you've got quite a bit of coupling,
both capacitively and inductively, between the wires and ground currents
induced by lightning that run parallel to them. These common-mode currents
are exactly what the GFCI is looking for, and when they reach a value over
its threshold, it trips.

Short of replacing the run with metallic conduit, about the only thing I
can think of to stop these nuisance trips is to isolate the circuit with
a transformer.

A properly-rated and -installed transformer should eliminate the need for
the GFCI altogether.

> I have thoroughly checked the circuits and loads for even the slightest
> short to ground with a DMM on milliohm scale. There was no sign of the
> slightest leakage.

Actually, you want to use your highest megohm scale to look for leakage,
and as much voltage as your ohmmeter can muster. But I doubt that this is
the problem anyway.

> Have thought about grounding the circuit to the well casing (think
> that's a much more effect than a ground rod. The soil conditions are a
> firm clay. Any thoughts??

That would probably make things worse, by allowing the lightning-induced
currents to flow in the grounding (green) conductor, which makes no direct
difference to the GFCI, but that wire probably has even tighter coupling to
the other two wires, increasing the pulses you see there.

-- Dave Tweed

2008\05\31@170316 by Dr Skip

picon face
OOPS. didn't catch that! Disregard my previous suggestions...



Dave Tweed wrote:
> Carl Denk wrote:
>                           ^^^^^^^^^^^ ???
>
> Seriously? I've never heard of a private home with its own gas well before.
>

2008\05\31@171012 by Carl Denk

flavicon
face
Thanks for all the replies. :)

Yes, it is a natural gas well, for a house, we tried 3 holes for water,
which were all dry, took the last one down 1000' and have heated
(including kitchen, clothes dryer, and garage heat) for 31 years. This
is not terribly unusual in our area 40 miles West of Cleveland, Ohio,
USA. The water vapor dew point of the gas is right about freezing, and
we have had the inside of the pipe frost over shutting off the gas. Not
good to come home when it is below zero F outside and house is below
50F, but we have backup electric resistance heat that automatically
comes on. That system is a topic in itself. The PIC sends data to a
standby generator's PLC, to control the electric heat and alarm on loss
of enclosure heat. If there is no communication (PLC master, PIC slave)
from the PIC, the PLC is programmed to use the electric house heat. The
PLC also monitors a sump pump and sets a audible alarm and sends an
E-mail to our cell phone if needed. Wouldn't be hard to add alarm for a
tripped breaker. The primary issue is, when the PIC has no power, the
PLC transfers the house heat to electric, which is expensive, and to
have the daughter drive 30 miles to reset the breaker is a pain if I am
out of town..

Don't know if an isolation transformer is acceptable code wise, but if
one leg got grounded and touched the other might not be good. I could
put a plain breaker in, and use a GFCI duplex outlet to satisfy the
code, but I would prefer to stay with the GFCI in the breaker box.

On the DMM, yes I used the megohm scale, sorry about that.

Yes there could be rain around during the problem. Last night we got
1/2", but never woke up from thunder. Likely is some water in the
conduit, but with the impervious clay we have around here, it is
unlikely to drain or evaporate anytime during the year. Rely on the
integrity of the "W" insulation of the wire. So what we have wet wise is
there all the time. I did look for spider nests, and that was OK. Have
had trouble with dial-up line previously during damp weather, but this
enclosure is very weather tight including thermal insulation, 2' wide,
2' high, and 6' long. And the electrical is all sealed as a 2nd line of
defense.

I don't have enough extension cord to go out there, but it's 99% sure
only happens when there is lightning near (not necessarily close, maybe
at least several miles).

Dr Skip wrote:
{Quote hidden}

2008\05\31@171914 by Carl Denk

flavicon
face
That's OK, we are just different around here. :) :)  Well cost $8,000 31
years ago, hows that for an investment. But it's not all gravy. It took
a few years to get a handle on heating pipes, pressure regulators,
automating the switch to backup electric heat, etc. But it's been stable
for some years. Previous to the PIC, there was a 1/8" nylon tubing is
the electric conduit with a pressure switch in the basement. I was
getting a little paranoid about a gas leak blowing up the house. The
fiber optic cable is much safer. :)  Within a 3 mile radius there are
another 1/2 dozen houses with wells, and another 1/2 dozen with
production wells owned by some gas company, and the house gets free gas
plus a royalty, but if the well is plugged for lack of production, those
people are out of luck, and usually they switch to propane.

Dr Skip wrote:
{Quote hidden}

2008\05\31@172015 by jim

flavicon
face
You must be near Sandusky?

Jim
----- Original Message -----
From: "Carl Denk" <spam_OUTcdenkTakeThisOuTspamalltel.net>
To: "Microcontroller discussion list - Public." <.....piclistKILLspamspam@spam@mit.edu>
Sent: Saturday, May 31, 2008 4:09 PM
Subject: Re: [EE] GFCI PROBLEM with lightening


{Quote hidden}

> --

2008\05\31@183344 by Carl Denk

flavicon
face
East of Sandusky, South of Elyria, East of Oberlin. Sandusky is an hour
drive.

jim wrote:
> You must be near Sandusky?
>
> Jim
> {Original Message removed}

2008\05\31@233146 by Apptech

face
flavicon
face
A long shot, but easy enough to try if there's enough cable
slack somewhere.
Wind cable in a loop of about 1 foot diameter and a few
turns.
Seems unlikely that it would work in this case but ... .
"Real" lightning strikes do not like to transit such an
arrangement and tend to blow through the insulation to local
ground somewhere on the coil. Calculations of inductance and
impedance to not suggest that such an arrangement should
work. Various real world reports suggest it actually works.
It MAY be that the rise time of the signals concerned would
be similarly affected. You MAY be getting local sub
flashover discharges, but these would not have the same
characteristics of a real strike.



       Russell


'[EE] GFCI PROBLEM with lightening'
2008\06\01@000137 by Jake Anderson
flavicon
face
Apptech wrote:
> A long shot, but easy enough to try if there's enough cable
> slack somewhere.
> Wind cable in a loop of about 1 foot diameter and a few
> turns.
> Seems unlikely that it would work in this case but ... .
> "Real" lightning strikes do not like to transit such an
> arrangement and tend to blow through the insulation to local
> ground somewhere on the coil. Calculations of inductance and
> impedance to not suggest that such an arrangement should
> work. Various real world reports suggest it actually works.
> It MAY be that the rise time of the signals concerned would
> be similarly affected. You MAY be getting local sub
> flashover discharges, but these would not have the same
> characteristics of a real strike.
>
>
>
>         Russell
>
>  
wrap it around a grounded iron pipe perhaps? increase the inductance and
give it a handy exit path.

2008\06\01@042907 by Apptech

face
flavicon
face
>> Wind cable in a loop of about 1 foot diameter and a few
>> turns.
>> Seems unlikely that it would work in this case but ... .
>> "Real" lightning strikes do not like to transit such an
>> arrangement and tend to blow through the insulation to
>> local
>> ground somewhere on the coil.

> wrap it around a grounded iron pipe perhaps? increase the
> inductance and
> give it a handy exit path.

Maybe, but for a real strike the iron will saturate at a
tiny tiny fraction of the current involved.



       Russell



2008\06\01@161340 by Carl Denk

flavicon
face
Several items:
1: This is not an actual lightning hit anywhere nearby, but
electrostatic (or whatever one might want to properly call it) energy in
the area, maybe a difference in potential of the earth as a result of
lightning storms not close, by near (kind of double talk ~)  ).
2:  The weather enclosure is a light steel angle frame with corrugated
aluminum siding screwed on. Hadn't thought about it previously, but
maybe want to ground it to the well casing with a heavy gauge cable, say
000, in case there was a direct hit. But there are 30' high trees nearby
that probably would be hit first.
3: Thought more about the duplex GFCI and plain breaker. The breaker
with or without GFCI, the black and white wires, the duplex receptacle
with or without GFCI, and the connected load are all in series, and for
there to be current, the current needs to flow in the circle of all the
series connected components, and if the GFCI was at either location will
detect the current imbalance. Or is it the parallel wires are an
inductor/capacitor parallel with the load?
4: What's the thinking on the black and white, possibly include the
green wire, to make them a twisted pair, or maybe I can find a shielded
cable and pull that into the plastic conduit? Changing the conduit to
metallic is not an option.

Apptech wrote:
{Quote hidden}

2008\06\03@133725 by Peter

picon face
I think that having a gas well in your yard is cool in view of fuel and gas
prices. A low yield one is probably better because it will last a very long time
(think about 31 years worth of gas bills for a house at todays prices - wow,
that's about two or three cars or half a house - also 1000 foot wells at $8000
all finished are pretty cheap imho ...). I have read about places where lighting
cnadles in the basement is a bad idea ... methane seeps in from the ground.

I think that adding a line filter unit at both ends of the cable will fix your
lightning problem and might prolong the life of your pic device. A line filter
is the inline 'box' type of RLC filter used to remove EMI from mains wiring.
E.g. from Corcom etc. The effect you are seeing is likely caused by the wire
going out acting as an antenna and coupling common mode current into the GFI.
That will trip it for sure. In a way you have a radio receiver of the same type
as a coherer ... getting the 15-20 mA needed to trip a GFI from 100 feet of
'antenna' connected to the well that will pick up ground currents is probably
trivial even if lightning strikes are fairly far off.

There is a reason for which mains hauling switchgear is more intelligent than a
GFI and makes several 'self repair' attempts in such a fault case, before giving
up. Typically after a nearby lightning strike the breakers restart after 60
seconds, and try it again three times, waiting successively longer times. If
there is a short or repeated trips within that amount of time they stay off and
operator intervention is needed. I don't think that you want to copy this
behavior. The line filters will likely reduce the sensitivity of your coherer
receiver to reasonable levels. They should have no impact on the normal GFI
functionality. I don't know what your code says about outdoor electrical
installations but I am fairly sure that this type of thing (30 meters of mains
to an outside building just like that) is totally illegal in most places when
not installed and checked by an authorized electrician. I don't even want to
think about what permits would be needed to use such an electrical installation
anywhere near a producing gas well, even if it would not be next to a house.

Peter


2008\06\03@154816 by Carl Denk

flavicon
face


Peter wrote:
> ... A line filter
> is the inline 'box' type of RLC filter used to remove EMI from mains wiring.
> E.g. from Corcom etc.
That wouldn't be too hard or expensive to do, I'll have to order a
couple of filters with my next order to Mouser, Jameco, etc.

>  The effect you are seeing is likely caused by the wire
> going out acting as an antenna and coupling common mode current into the GFI.
>  
That's the sort of thing I was thinking was happening, but that being
out of my normal turf, I didn't know the words to use. :)

> There is a reason for which mains hauling switchgear is more intelligent than a
> GFI and makes several 'self repair' attempts in such a fault case, before giving
> up. Typically after a nearby lightning strike the breakers restart after 60
> seconds, and try it again three times, waiting successively longer times. If
> there is a short or repeated trips within that amount of time they stay off and
> operator intervention is needed.
We are on a rural electric cooperative power utility, and their
switchgear tries 3 time within about 10 seconds to reset before giving
up. If a tree branch or Squirrel (small mammal) briefly causes a short,
all is fine, otherwise a crew is dispatched to fix the problem. Recently
a crew was replacing a 3 phase pole, and instead of going to the
substation, and opening the breaker, I think the took a hot stick with a
ground conductor and brought the circuit down, because we got the 3
resets before it went down for the duration of their work. Now I am on
their call list before they take us down for maintenance.  The house
main has a whole house surge suppressor, and I am generous with
suppressors around the house at equipment. Last thing I lost was a 2
years ago, a dialup modem, but that came in the phone line, which now
has a suppressor on it also. Glad it wasn't an internal modem!

>  I don't think that you want to copy this
> behavior. The line filters will likely reduce the sensitivity of your coherer
> receiver to reasonable levels. They should have no impact on the normal GFI
> functionality. I don't know what your code says about outdoor electrical
> installations but I am fairly sure that this type of thing (30 meters of mains
> to an outside building just like that) is totally illegal in most places when
> not installed and checked by an authorized electrician. I don't even want to
> think about what permits would be needed to use such an electrical installation
> anywhere near a producing gas well, even if it would not be next to a house.
>  
Actually electric power at a gas well is common. The production wells
that produce oil also have a "pump jack" (typical equipment seen at an
oil well, big counterweighted shaft going around maybe 4 RPM, pulling a
rod going down the well bore. It's a piston pump pulling the oil up in
many stages.) Usually there is an electric motor maybe 5 or 10 HP. As
for the electric near the natural gas, that's not an issue, unless there
is a defect in the well casing or piping, which is not wanted to loose
the gas there is no hazard. State regulations require 200' to a
building, but once the well is in, maybe you could build much closer. In
cities we see once in a while a well being plugged in city neighborhoods
with small lots(maybe 50' x 140') . In some instances the conduit might
want to be explosion proof, but I'm not sure that is required in this
case. Electrical inspection in our area is lax to non-existant. With the
blind flange off the top of the well, if ignited, the flame would be a
modest maybe 4' high, and a nice hand warmer for the well driller. The
driller used a cable tool rig to drill the well, and when the bit was
going up and down in the well, the driller would light the gas, and keep
his hands warm. :)

2008\06\04@125617 by Peter

picon face
I know just enough about gas/oil operations to scare me. Your drill operator
must have had cojones to light up the gas at the well head while drilling, and
he was almost certainly not an oil/gas guy. Of course he 'knew' by clairvoyance
that he would never hit a small void containing just enough gas to make a
fireball that would bake him medium rare. By the way, afaik some drills used for
water wells are not ok for oil/gas wells because the string parts can spark
while drilling, drill trucks are expensive, and one can't just hop in and drive
away when the drill string is in the well ... Either that or he planned to make
his widow rich with his life insurance in case something did happen ...

Anyway, good luck with the line filters,

Peter


More... (looser matching)
- Last day of these posts
- In 2008 , 2009 only
- Today
- New search...