Current leaking to ground pin, how?

[mwhens]

Hi all!

I'm renovating my house, and the ground net wasn't connected to anything yet.
I was working while standing on wet ground outside, and I felt a few times a "shock" or more like a low amp AC current when I touched a grounded metal on the lamp of fan.
So I made a priority to finally connect the new ground pin to the houses ground net.

Before connecting I checked what happened when the ground net wire and ground pin wire touched, I could see very tiny sparks.
I wanted to test what circuit breaker group was the issue, but I did not get anything useful out of those tests.. I think...

(We have 230v AC in the Netherlands)
I first measured the voltage when all breakers were on, using a multimeter between the two wires, it was 49,5v AC, and measuring current using the same multimeter at 3,3mA.
I guess this was enough to "shock" me...?

Then I started measuring with all breakers off, and is was 2,5v at 0mA
Then measured every breaker solo (only one breaker on at a time)...

breaker 1 on solo = 97,8v - 0,8mA
breaker 2 on solo = 114,6v - 3,3mA
breaker 3 on solo = 121v - 4,7mA
breaker 4 on solo = 62,4v - 3,7mA
breaker 5 on solo = 52,3v - 3,7mA
breaker 6 on solo = 53,2v - 3,8mA
breaker 7 on solo = 53,2v - 3,3mA

So I see no single breaker/circuit being the cause for the ground leak, every circuit has some leakage it seems.
I also tested all breakers on and turning each one on and off, sometimes the leakages becomes bigger, sometimes less when turning a breaker on.
But still, it's enough to "shock" me. I'm not sure if I measured correctly, the amps are a short, and voltage is not.
Is this leakage normal?
At most it's around 0.5watts of energy flowing to ground it seems... but why?

Any insight is useful

[BillyO]

I'm not sure how the wiring is done in the Netherlands, but here in NA we have a line (or live) a neutral and a ground.  The ground is connected to the neutral at the main breaker panel.

Only the line is passed through the breakers.  Neutral does not go through a breaker.

If you measure from line to ground or line to neutral you should see the full supply voltage (120V for us).  If you measure from neutral at any outlet to ground there should be 0V.  If you do measure a voltage that means a bad or broken neutral connection somewhere along the line from the panel to the outlet.  But that is for how our houses are wired.

[exe]

Very interesting. I'm also in NL. I'm probably having a different issue but may be it's somehow related. When installing light bulbs I noticed there's some 60-80V present between two wires even if the light switch is off (DMM impedance is ~10M). I measured this in multiple rooms. Not sure what's going on. Leaky light switches? Capacitive coupling??

[Kim Christensen]

Not sure about your country but here in Canada, the power company also has their neutral connected to ground at various points along the distribution system. Since "ground" is just soil and rock, their neutral will not be at the exact same potential as your ground (There is a max voltage spec for that by law). In the main breaker panel the neutral is connected to the house ground net (buried metal rod or plate in the ground). So it would be natural that there would be a potential difference before you do the final connection to the house ground net, though your voltages seem a quite high.

Ah... I reread your post. exe's comment gave me the key. It's most likely the capacitive coupling between the wiring in your house. Also, some electronics have small capacitors between live/neutral/ground for RFI suppression so anything plugged in can make it worse.

[themadhippy]

Is this leakage normal?

maybe,what sort of loads are on the circuits? how old is the wiring? is it damp?is there signs of rodent damage  or  101 other things that could cause a leakage To test things properly you really need a megger to test the insulation resistance of each circuit

Capacitive coupling??

yep

[jmelson]

In general, there should be an "earth" wire from the main power feed that is brought to the main panel (we call it service entry in the US).
This will likely be grounded with a long ground rod somewhere outside the residence, like next to a powr pole.  You will likely also have a ground rod driven next to the service entry.  Since these are connected to the power distribution system, there will be some voltage on these "grounds" as the earth is not a great conductor.  So, this is probably normal.  There may be some limits established in the Netherlands as to what is the acceptable amount of residual voltage there.
Jon

[TimFox]

Capacitive coupling at 50/60 Hz:
This is clearly seen when using a high-impedance voltmeter to measure the voltage where zero is expected with an open switch.
Modern DMMs tend to have 1 to 10 megohm impedance at this frequency.
To get roughly half the supply voltage across the meter load requires a capacitive reactance on this order (1 to 10 megohm), or 300 to 3000 pF at 50 Hz.
Some electrician DMMs have a switchable low-impedance mode to avoid this situation where wiring, line filters, or switch protection capacitors obscure the problem.

[pcprogrammer]

Do I understand it correctly that you are measuring between the ground (green/yellow) wire from the house installation and a freshly installed ground pin?

In the Netherlands the power company supplies, for as far as I know, for a single phase connection, a neutral and a live wire. These come into the house onto the meter board and go into the main fuse, which you have no access to, and then goes into the meter. From the meter the two wires go into the breaker board. In most cases it goes to one or multiple RCD's (Aardlekschakelaar for the Dutch OP). After the RCD's come the breakers, all double pole switched. From a breaker the neutral and live go into the house. The ground wire to go with it comes from a common ground bar that is connected to a earth pin.

This ground wire should normally be free of any voltage whether it is connected to the ground pin or not. I don't think that capacitive coupling within the wire guides leads to producing sparks when connecting the internal ground system to the external ground pin.

The measured current is not high enough to trip the RCD's, but it is still an indication that there is a fault in the wiring or some connected appliance.

For a three phase system it is not much different. The incoming supply consists of a neutral and the three phases. Ground is setup in the same way.

My advise is to call in someone with a proper insulation tester and check the wiring.

[Gregg]

What and where are all of the connections to your ground net both directly (wired) and indirectly connected?  By indirect I mean things like water pipes, gas pipes, metallic conduits, heating ducts that touch things that may be grounded etc.

It appears that your ground pin is at a different potential from your ground pin which isn’t as uncommon as you might suspect.  Unless your system has a dedicated grounding conductor all the way back to the distribution transformer (where the neutral and a good ground should be connected together) you will probably find a potential difference between ground pins a few meters apart.  If you or a neighbor have a poor neutral connection, the neutral will try to return via the ground pin making the point to point earth potential much greater.

Capacitive coupling to the ground conductor varies as each circuit is powered which makes it seem more complicated than it really is.

All of your grounding conductors should all be connected at only one point to avoid circulating currents within your grounding system. 

[pcprogrammer]

This ground wire should normally be free of any voltage whether it is connected to the ground pin or not. I don't think that capacitive coupling within the wire guides leads to producing sparks when connecting the internal ground system to the external ground pin.

I was doubting my own comment about no voltage in light of the fact that you measure a shadow of mains voltage when touching an oscilloscope probe tip. So I conducted a simple test. Took a ~15m 3 wire extension cord and plugged it in a non grounded outlet I made. Measured between the ground wire of the extension cord and the ground of the house wiring with a modern digital multi meter and got a reading of ~112V. Mains here is ~240V.

But I did not measure any current and did not get shocked either when touching both grounds.

So I still believe there is a fault in the original posters system.

[exe]

To get roughly half the supply voltage across the meter load requires a capacitive reactance on this order (1 to 10 megohm), or 300 to 3000 pF at 50 Hz.

Oh boy, never thought that it takes so little capacitance... Thanks, Tim!

[golden_labels]

Do you have any bigger device with a switched power supply connected? Like a PC or a TV? Y capacitors in them form a voltage divider, with earth being its output. That’s why such devices have a strict requirement to have the earth pin grounded. Otherwise the user will get zapped. Current passing through those is not high enough to cause direct harm to a human, but can be unpleasant.

[Brianf]

You will likely also have a ground rod driven next to the service entry.

Not necessarily. This situation only applies to a 'TT' supply arrangement.

Others, such as TN-S, TN-C-S, TN-C-S (PME) and TN-C use an earth connection provided by the supply company. There is no consumer supplied earth point.

[SeanB]

All the above supply methods also require equipotential bonding, to all extraneous metal in the house, like metal water and gas pipes, and any metal frame in the structure. This ensures there is no step potential in the house itself, in case of a fault to a metal case, so that the RCD devices can operate properly, That OP felt a tingle from a disconnected floating ground is kind of normal, coming from the cable capacitance, and this is why you see the small spark, as there is a capacitive current flowing normally, and more so if there is a fault.

Most supply authorities will also, at point of supply, also have a ground point, simply to make sure that faulty neutrals show up as a fault current in the substation, either done deliberately by having a ground rod, or incidentally by having a galvanised steel frame on a concrete base that is in contact with ground, bonded to the neutral bus, or in many cases acting as the neutral busbar itself.

[EPAIII]

This is an ELECTRONICS BB, not one for electricians. IMHO, neither the building's electric system nor any appliance that is not defective should provide enough leakage for you to get a shock, but I am far away from you, across a big puddle.

Call a good, local electrician and ask if it is a problem. Better safe than sorry.

[golden_labels]

This is an ELECTRONICS BB, not one for electricians. IMHO, neither the building's electric system nor any appliance that is not defective should provide enough leakage for you to get a shock, but I am far away from you, across a big puddle.

The OP is not searching for an advice, but explanation of an observed situation. The installation will be fixed and no reason to think it will not be conforming to the local electric code.

Any device, defective or not, may produce enough of leakage or even a direct short, if it is improperly wired. And OP has clearly stated this is exactly the case: the earth wires are floating.

[Gyro]

This is an ELECTRONICS BB, not one for electricians. IMHO, neither the building's electric system nor any appliance that is not defective should provide enough leakage for you to get a shock, but I am far away from you, across a big puddle.

In fact the forum does have some industrial electricians among its numbers. It's not the first time mains installation related questions have come up. eg: https://www.eevblog.com/forum/chat/post-a-picture-of-your-fuse-box/ (started by Dave) among others.

[pcprogrammer]

This is an ELECTRONICS BB, not one for electricians. IMHO, neither the building's electric system nor any appliance that is not defective should provide enough leakage for you to get a shock, but I am far away from you, across a big puddle.

As a matter of fact, I have formal training in that field, but it is old, >40 years. Worked in the field too for a while. Did several installations according to code, which in the Netherlands is the NEN1010. Unfortunately I lost my copy of it long ago.

Can't recall receiving information about voltage and current flow on (floating) ground wires due to capacitance though, and what the OP mentioned sounds like to high a currents for just cable capacitance. In the Netherlands the cabling most often is based on 3 separate pvc insulated wires within a pvc tube. It depends on how old the house is of the OP if this is the case, but chances are it is because this system has been around a long time. It was what I was taught.