Earth leakage

So what ?

Well, straight away I can tell you that there is no significant L-E fault or N-E fault. the exact figure is not especially meaningful, but too large raises an alarm.

What would be 'too large' ?
Compare this to  a system that maybe would be fed by a 300mA RCD  say a farm with many stables etc spread out over a few hundred square metres, - we would be well under the 300mA - which is an arbitrary figure, but is set considering the small risk of a single point fault getting hot enough to cause overheating or start a fire. (300mA ~ could be up to 80 watts going astray. In this example, only 25mA so more like half a dozen watts.)
Actually anyone who has ever touched a 60W light bulb will appreciate that an 80W resistive fault in a confined space could well ignite something.

Of course here we know its not a single point fault, but showered around as lower level leakage on each of many sub circuits so that heat risk is not relevant here.

If there was a lot of leakage, I might also suggest to look at current in main earth bonds and the supply CPC - there could be diverted neutral currents as well that contribute.

We don't of course know the phase relationship of the earth leakage current to the line voltage, but I'd put money on it being nearly at 90 degrees, and capacitive on the single phase final circuits. As far as I know, no cheap device exists for verifying this, but all wiring has some capacitance core to core - the plates of the capacitor are the wires and the insulation acts as, err, the insulation. But the main source of line to earth capacitance is modern electronic devices.

Twin and earth usually manages about 50 to 100pF per metre, depending on the size and manufacturer, round flex is rather higher.

To pass 10mA say, at 230V needs 23k ohms. at 50Hz this is a capacitance of ~ 0.15uF or 150,000 pF - so requires a rather unlikely couple of km of cable. 
But a few hundred metres may be credible on a large site, so single figure milliamps are certainly possible, from cables alone. 
(1uF = 3140 Ohms. 1nF 3.14M Ohms 1pF 3.14Gig Ohms- 'j' ohms not hot ones... )

But almost all electronics now has a switch-mode power supply (rectifying the mains, and then chopping the DC at 100kHz of more - to allow smaller, lighter, & cheaper transformers ). The inductance of the mains means it cannot supply current that stops and starts a hundred thousand times a second or more, and if it tries the mains leads form an antenna and radiate RF interference.
So, to make it work reliably (and to meet EMC requirements), a local 'tank' that can charge and discharge very fast is needed. (like the tank on the loo, which spreads the impulse load of the flush out over time and saves on a fat pipe all the way to the reservoir !! )
Enter the capacitors, a big one L-N and a couple of smaller ones L-E and N-E  to tide over and attenuate fast pulses that might otherwise travel between these lines and cause havoc elsewhere.
At frequencies of 100KHz and up these capacitors essentially hold the voltage between lines from changing suddenly, and prevent them 'doing the splits' in voltage terms in response to  the switching waveform. But at 50Hz, they are still capacitors, and some current then flows L-E.
In a 3 phase system with well balanced loads this all cancels, but of course things are never really in balance. What remains it what you see on the meter.

In summary I see a system with some electronic loads connected, but nothing to raise alarm. 

If there was ten times more leakage, or even a bit less than that on a really small installation, or if you said all loads had been removed, it might warrant investigation. but this is fine.

Mike

So what ?

Well, straight away I can tell you that there is no significant L-E fault or N-E fault. the exact figure is not especially meaningful, but too large raises an alarm.

What would be 'too large' ?
Compare this to  a system that maybe would be fed by a 300mA RCD  say a farm with many stables etc spread out over a few hundred square metres, - we would be well under the 300mA - which is an arbitrary figure, but is set considering the small risk of a single point fault getting hot enough to cause overheating or start a fire. (300mA ~ could be up to 80 watts going astray. In this example, only 25mA so more like half a dozen watts.)
Actually anyone who has ever touched a 60W light bulb will appreciate that an 80W resistive fault in a confined space could well ignite something.

Of course here we know its not a single point fault, but showered around as lower level leakage on each of many sub circuits so that heat risk is not relevant here.

If there was a lot of leakage, I might also suggest to look at current in main earth bonds and the supply CPC - there could be diverted neutral currents as well that contribute.

We don't of course know the phase relationship of the earth leakage current to the line voltage, but I'd put money on it being nearly at 90 degrees, and capacitive on the single phase final circuits. As far as I know, no cheap device exists for verifying this, but all wiring has some capacitance core to core - the plates of the capacitor are the wires and the insulation acts as, err, the insulation. But the main source of line to earth capacitance is modern electronic devices.

Twin and earth usually manages about 50 to 100pF per metre, depending on the size and manufacturer, round flex is rather higher.

To pass 10mA say, at 230V needs 23k ohms. at 50Hz this is a capacitance of ~ 0.15uF or 150,000 pF - so requires a rather unlikely couple of km of cable. 
But a few hundred metres may be credible on a large site, so single figure milliamps are certainly possible, from cables alone. 
(1uF = 3140 Ohms. 1nF 3.14M Ohms 1pF 3.14Gig Ohms- 'j' ohms not hot ones... )

But almost all electronics now has a switch-mode power supply (rectifying the mains, and then chopping the DC at 100kHz of more - to allow smaller, lighter, & cheaper transformers ). The inductance of the mains means it cannot supply current that stops and starts a hundred thousand times a second or more, and if it tries the mains leads form an antenna and radiate RF interference.
So, to make it work reliably (and to meet EMC requirements), a local 'tank' that can charge and discharge very fast is needed. (like the tank on the loo, which spreads the impulse load of the flush out over time and saves on a fat pipe all the way to the reservoir !! )
Enter the capacitors, a big one L-N and a couple of smaller ones L-E and N-E  to tide over and attenuate fast pulses that might otherwise travel between these lines and cause havoc elsewhere.
At frequencies of 100KHz and up these capacitors essentially hold the voltage between lines from changing suddenly, and prevent them 'doing the splits' in voltage terms in response to  the switching waveform. But at 50Hz, they are still capacitors, and some current then flows L-E.
In a 3 phase system with well balanced loads this all cancels, but of course things are never really in balance. What remains it what you see on the meter.

In summary I see a system with some electronic loads connected, but nothing to raise alarm. 

If there was ten times more leakage, or even a bit less than that on a really small installation, or if you said all loads had been removed, it might warrant investigation. but this is fine.

Mike