Earth Sparkin (not ur usual......)

[Ipswich Sparky]

Testing an installation - RCD's not testing properly (can get a good test occasionally.?)- Newish Megger recently calibrated
just tested RCD's at home and all good

All MCB's off, RCD's off, main switch off and main fuse removed, no alternative supplies

can get a good tingle off of the earth - disconnected earth bonding to water (no gas or oil etc) and main earth bond no voltage on consumer unit earth bar measured

PME installation - overhead supply - old uninsulated neutral style (N-L 243v , L-E 243v)

ALL circuits test perfectly IR, R1R2, ZE, ZS etc (supply 0.16 ohm Ze same as main bond approx 450mm long)

Can get a goodly spark to the earth bar in the consumer unit from either.... - have a video of this - showing exactly above
cannot measure a voltage across either and earth in the consumer unit common but they spark

Realise there will be earth potential on water pipework etc

Any ideas ?

 

[Ipswich Sparky]

Surely if you have circulating currents on a pme system when a second earth via water pipes etc is introduced this would affect the rcd ?

Surely the circulating current is a closed loop until the secondary earth via water pipes is connected ?

 

[pc1966]

Surely if you have circulating currents on a pme system when a second earth via water pipes etc is introduced this would affect the rcd ?

The CPC/earth is not passing through the RCD, it looks only at the final circuts' L&N which ought not to have connected to the CPC in any way.

If you have a high circulating current in the PME to earth bonds it might change the L-N volts slightly but it is not going to have any significant impact on the RCD behaviour.

As you have done a swap and it seems not to be the RCD, and the IR is fine so there should not be any oddities due to a N-E fault, you should look to see if there is any odd loads on the circuits that might be causing some stray current.

Do you have a clamp ammeter that can measure down to mA to see what sort of imbalance is currently happening?

Another possibility is there is a faulty load causing a DC fault that might blind a 'type AC' RCD, but again that sort of thing ought to show up on the IR tests. Though maybe try the IR test with both polarities in case you just happen to have tested a diode fault the wrong way for any significant leakage.