Right chaps/ladies- this seems to be an appropriate place and thread to get educated better in my understanding. 

Please may someone help me and my understanding of generator earthing requirements. 

I don't do Generators all day every day, but periodically do for construction site set ups and temporary supplies for construction site compounds. This is not my day in and day out job, but not uncommon either. 

So most Generators larger than 10KVa arrive as standard, here in the UK as a TNS system. I believe that is standard operating procedure around here.

So far all mine have. Hired from Large national hire companies.

So the generator runs and operates as a TNS System. Earth fault path is a TNS system. 

However, I understand if the Generator is NOT earth staked, the neutral and earth systems, because they are common at the Genset - can float up and down with reference to terra firma - the ground next to the Generator - and RCDs don't know when a fault is occurring as this floating value confuses the RCDs......, yes?

All of my generators have come with a earth fault relay RCDs that I'd normally set up, (I commonly call these "viggi units") as described in the above sort of way, with a 300mA 0.25 second time delay and cascading RCD settings out from there.

Clearly, this needs to have an earth reference point to tell the RCD what value earth is. Of course it does - no dispute there. 

So - I think I'm all on track so far......

Now, to the earth reference point; the earth stake, matt, or commonly for me, on railway sidings or construction sites out in the country side where a metal fence has been installed, a metal fence that is "owned by" and managed by us, and I have permission to use - then I'd use the metal fence, I like a metal fence................... and its required value in ohms.

So, hypothetically,  lets pretend that the generator was laid on a glass slab, or floating in the air - no contact with the ground at all, and the earth reference point (The earth stake hammered into the ground and connected to the Generator earth point for example) is effectively an infinite value as an earth fault path back to the Generator, but it does provide the earth part of the generator with a reference point to what the earth value should be. I understand that the earth fault relay RCD "viggi" unit would function perfectly well.

So to my question - why should we try to achieve a 20 ohm value for this earth reference point - its just an earth reference and not an earth fault path - surely? 

(Obviously I'm not talking about a TT system here - I understand the need for that earth stake to be less than 200 ohms and all its difficulties, I think)

If the generator earth reference point was 200 ohms or 2000 ohms, would it make a difference? I don't think so.....

I feel an ohms Law coming on as the explanation here.........(V=IR maybe?)

But it's not a fault path, just a reference point........Education required please.

If i stuck a good old metal shaft screw driver in the ground it'd (The RCD part of the Genny) probably function perfectly well surely?

Your understanding is sound, but there are a couple of extra situations to consider,

1)

Actually even without the earth rod there is merit in the RCD for 'glass slab' systems, or vehicles on tyres as we more commonly find the nearest approximation to that - the CPC may bond solidly to the chassis but a human in between an exposed  live and the vehicle body at potential of the system CPC is a few k ohms and will not trip an MCB.

But an RCD sees the current coming back via the CPC as 'missing' from the expected neutral current, and trips on the imbalance between L and N current in the normal way.

2) 20 ohms is a bit arbitrary but it cannot be too high - now our genset is on the ground and the earth rod is in the dry and not a great contact.  Any fault to that connects live to terra-firma (wire pinched on the rigging of a lighting tower, socket box falls into a ditch full of water...) will cause current into the Terra-firma at the fault, and back out of terra firma back to the genset star point at the genset earth electrode. This current raises the genset case and the rig CPC relative to terra-firma, and we do not want the rise to be too painful for anyone leaning against any 'earthed' metalwork connected to the genset CPC loop.

With a 300mA and delay set on the 'Viggy'  unit, 0,2 amps (plus some) could be flowing all day, so a hundred ohms would give a rise of 20V plus and 200 ohms would be 40V or more and some rise. 2000 ohms would never operate that RCD.


A safety factor of ten from a 50V rise with the highest common RCD setting is a very safe starter for ten with a lot of allowance for seasonal variation and corrosion.. but  yes, 20 ohms is to be seen as an aspiration, not an essential.

Mike

Your understanding is sound, but there are a couple of extra situations to consider,

1)

Actually even without the earth rod there is merit in the RCD for 'glass slab' systems, or vehicles on tyres as we more commonly find the nearest approximation to that - the CPC may bond solidly to the chassis but a human in between an exposed  live and the vehicle body at potential of the system CPC is a few k ohms and will not trip an MCB.

But an RCD sees the current coming back via the CPC as 'missing' from the expected neutral current, and trips on the imbalance between L and N current in the normal way.

2) 20 ohms is a bit arbitrary but it cannot be too high - now our genset is on the ground and the earth rod is in the dry and not a great contact.  Any fault to that connects live to terra-firma (wire pinched on the rigging of a lighting tower, socket box falls into a ditch full of water...) will cause current into the Terra-firma at the fault, and back out of terra firma back to the genset star point at the genset earth electrode. This current raises the genset case and the rig CPC relative to terra-firma, and we do not want the rise to be too painful for anyone leaning against any 'earthed' metalwork connected to the genset CPC loop.

With a 300mA and delay set on the 'Viggy'  unit, 0,2 amps (plus some) could be flowing all day, so a hundred ohms would give a rise of 20V plus and 200 ohms would be 40V or more and some rise. 2000 ohms would never operate that RCD.


A safety factor of ten from a 50V rise with the highest common RCD setting is a very safe starter for ten with a lot of allowance for seasonal variation and corrosion.. but  yes, 20 ohms is to be seen as an aspiration, not an essential.

Mike

Ah - so the value of 200 ohms keeps the touch voltage to 40 odd Volts - R<50V/Ia - has been used I assume?

Less than 200 ohms then ever less touch voltage.....

Now - this is where my uderstanding might be falling over - if that Genny is on wheels, a glass slab or up in the air or something similar and an earth stake is installed - then there wont be a path back to the generator, via the earth stake - clearly there's a fault path via the TNS path.........and perhaps a division of fault current (depending on the resistance) -down to terra firma via the earth stake. - Like a Zs would be infinite - or as near to - if you did a live or dead test - from the Genny and tested from the Genny Live to earth stake (disconnected from the Genny MET)

In old fashion parlance - an ELCB would work. before even my time, but obviously they're still around.

However our modern viggi units also work, (don't they? won't they?) because the earth stake gives the RCD an earth reference to work from, and holds the Genny frame to a local earth potential - that is different to the neutral value - which is also held, now at a local earth value - and allows a current loss too - to somewhere other than the TNS fault path, and the RCD functions....

So in this case the 20 ohms, 200 ohms, or 2000 ohms path back to the genny is irrelevant surely? only a good contact with earth? it'd give an earth reference, its not a fault path.........

I'm still tripping over the same mis understanding.........

Kind Regards

Chris

So in this case the 20 ohms, 200 ohms, or 2000 ohms path back to the genny is irrelevant surely? only a good contact with earth? it'd give an earth reference, its not a fault path.....

It's not a fault path for a conventional ADS style L-PE fault, correct. In an additional protection kind of situation (e.g. picking up the end of a severed flex while stood outdoors) it does though. Likewise for the (rare) situations where a line conductor come into contact with true Earth without an PE conductor intervening. On larger systems you start to see stray capacitance and general leakage currents due to imperfect insulation having similar effects.

    - Andy.

Likewise for the (rare) situations where a line conductor come into contact with true Earth without an PE conductor intervening.

Not so rare. That is your expensive Ho7 style  rubber 3 phase 100A extension lead  being thrown over a metal fence and snagging on a sharp edge, and striking a phase core first, or the free socket on the far end going 'splosh' as it lands on the other side .

While in a perfect world this would not occur, it can and does happen when things are being deployed in a hurry and / or  in bad light.

Something like SY or CY cable that actually had a low enough resistance braid to operate a large MCB would help with the snagging, but the flooding has no easy fix. If your genset has sockets for leads for unknown loads to plug into, it needs an electrode, an  NE bond and an RCD.