I got some trouble with our wiring. When I worked on some circuits of our late 1970ies-panel I inadvertently touched the ground bus with the back of my hand and got a nice shock. When i pulled out the dishwasher ground wire it sparked heavily. No other ground wire did that. There's definitely quite an amount of juice flowing in the ground, but neither the old 100mA nor the new 30mA RCD have tripped.
Can you give me any help here? When I threw the dishwasher breaker (1+N) the arcing stopped as well.

Is the installation a TT system? If yes, then did you measure the earth resistance? If it's too high, then your RCD won't trip that easily.

TN-C-S with the ground wire bonded at the house entrance. As far as I know no ground rod, only an old lead water pipe bonded with an old rusty clamp.
There's a single long piece of NM cable running from the panel to the receptacle behind the dishwasher, so not much room for mixups (total wire run maybe 10m, the panel's located in the hallway, just outside the kitchen door). The dishwasher is fairly new (2 or 3 years I think). Part of the wiring is NYM cable run in surface-mount conduit, rest is same cable buried in plaster.

To quote Mr. Spock, "Fascinating!"

Assuming that both RCDs aren't faulty, the loop impedance would need to rise to a substantial level to prevent them from tripping (something over 2200 ohms for the 100mA type nad 7.3K for the 30mA one). With TN-C-S the only practical way I can see for this to happen is a very bad bond between neutral and ground at the service entrance.

The only problem is that with the current down to less than 30mA, I wouldn't expect to see that much of a spark when lifting the dishwasher ground.

Just had another thought on this.

Are the RCDs the type that will respond to pulsing DC as well as pure AC?

A badly corroded bond could act as a rectifier. If the RCDs you have don't respond to pulsing DC then you could have significant fault current flowing on only positive or negative half cycles without the RCD tripping.

What did you get a shock to? - that is what else are you touching ? Any ground bus that can give a shock (to earth) simply isn't a ground ! No amount of leakage current could raise the voltage that much.
I would suggest a thorough checkout of the wiring associated with the panel, including RCD test, and a PAT test on the machine.

Is there equipotential bonding? Mains bonding?

Is the neutral bonded to earth, via an earth rod or similar, except at the transformer?

If the electrical earth isn't in contact with mother earth, you'll have the full voltage drop in the neutral as a potential between the earth wire and the building. This pretty much requires that the water pipe is plastic somewhere, or else a water heater or similar will provide a path for earth.

I wasn't touching anything at that moment, except for the plastic handle of an insulated screwdriver. I was standing on an aluminum ladder, but I had rubber soles.
The wiring is rather suspect. There's a bonding bus bar in the basement where gas and water mains enter. The new gas mains isn't bonded, the new water mains is plastic. Cable TV is still bonded, as is a weird stub of old lead pipe disappearing in the ground. The bus bar is rusty and threatening to fall apart (half of it already did). From the bus bar a ground wire runs to the main fuse enclosure. At the PoCo disconnect ground and neutral are simply under the same screw, from this point all distribution is 5w. From the cutout a new cable goes to the main fuse box mentioned above. Then the risers go to every floor where a short lead to the apartment main fuses is tapped off. The riser wiring is late 1970ies PVC wires in FX conduit. Maybe the ground is disconnected somewhere, which would at least explain the RCD not tripping.
The old RCD was definitely working, it tripped on some occasions (direct hot-to-ground faults though). We carefully tried touching the ground bus bar with the back of our hands but never got shocked since. There was no voltage between Neutral and ground, but between the lifted-off dishwasher ground and the bus we measured a fluctuating voltage up to 200V.
All water pipes except for the last few m in the apartment are plastic (makes for nicely floating grounds in the apartments with older wiring )
Voltage drop is the only explanation that seems possible.

Yes, Fascinating, Mr Spock!
There must be a problem in your ground cable somewhere along the way. Maybe even in the PoCo box. It can happen that the screw is loose.

[This message has been edited by Belgian (edited 07-20-2003).]

[This message has been edited by Belgian (edited 07-20-2003).]

So with the dishwater circuit ground lifted at the panel you were reading nothing between neutral and the ground bus, but up to 200V between neutral and the dishwasher ground?

It sounds as though there's a double problem here: (a) A ground-fault/leakage somewhere in the dishwasher or its branch circuit, and (b) a bad bond on the main grounding system.

No. We got the voltage between the floating ground wire and the ground bus. No voltage between ground and neutral at all.

Before thinking any further on this, were you using a high-impedance digital meter?

In my job I work on 230 V-dc systems and it is standard for neither line be earth grounded. Thus both the line and neutral float above ground. If I take a voltmeter and measure the voltage from each line to ground and there is no leakage current favouring one side, I will get an exact split reading of 115 V-dc.

It just seems that your neutral has to be floating somehow and is not tied to the earth ground or not connected well.

On the other hand, are you sure you got an electrical shock? I've experienced a sensation similar to a shock once when my elbow rubbed slightly against a metal beam.