Bit quick advice please

[hightower]

Thanks for your help everyone. On phone at minute so will respond properly when i get on a computer.

 

[Richard Burns]

If you were going to install a rod it would have to be a very good rod connected to the MET of the installation and the cost would be prohibitive unless you have very good soil as the resistance would likely need to be in the region of one ohm. The container should be bonded to the MET as well.
Placing a rod at the containers will reduce potential differences at the container but the rod may take the installation current on an incoming broken neutral.

 

[hightower]

Okay, on a computer so I'll try and put all the relevant details in a single post.

Two containers that are insulated and have refrigeration units, to be installed about 20m away from a factory unit in the factory car park. Supply is TN-C-S so my initial thoughts were stop the supply earth at the isolators and have an RCD to protect with a rodded earth. But then my thoughts were we need to avoid putting it on an RCD so there is less chance of the units tripping out the feed and the stock within going bad.

However, the units being big metal things got me thinking what happens if the DNO loses a neutral. That's why I thought about rodding it alongside the supply earth, but I'm not really sure what that'll achieve. I think the electrical supply is totally separate from the unit, as in Class II, however, is the unit still classed as an extraneous conductive part and therefore needs bonding?

 

[Wilko]

Happy to be shown the error of my ways, but if it's to be an extended TNCS then I'd be inclined to bond the metal container back to main building met.

 

[hightower]

^^^ That's exactly what I'm thinking as well @Wilko

 

[hightower]

But TN-C-S means it's gonna be a 10mm. Don't suppose anyone has a table of SWA and their armour CSAs (in copper equiv.) to see if the armour can be used? Otherwise guess it's a case of getting the calculator out.

 

[hightower]

According to this table:

http://www.askthetrades.co.uk/hosted_images/Armour CSA.pdf

6mm 4-core has a copper equiv. CSA of 16.8mm so my thoughts are I can take a 10mm cable straight from the feed terminated in the isolator and bolt it on to the metal frame, and use the armour as my bonding return to the MET?

 

[Leesparkykent]

According to this table:

http://www.askthetrades.co.uk/hosted_images/Armour CSA.pdf

6mm 4-core has a copper equiv. CSA of 16.8mm so my thoughts are I can take a 10mm cable straight from the feed terminated in the isolator and bolt it on to the metal frame, and use the armour as my bonding return to the MET?

no if another metal is used for bonding then it needs to offer at least equivilant conductance. The carbon content of steel varies so you have to use the guidance in GN8 which I believe is in the region of 8.5. So if 10mm copper is deemed adequate then the steel armour would need to have a cross sectional area of at least 85mm. It would be better to use 5core SWA or run in a seperate bonding conductor. Remember if the seperate bonding conductor is to be buried you would have to use 16mm.

 

[hightower]

So what are the copper equivalent columns in that link referring to then? Reason we've used 4-core is because it's what the factory owner had lying around and he's on a cost saving exercise. Will have to have a look at pulling in a separate conductor.

 

[Leesparkykent]

So what are the copper equivalent columns in that link referring to then? Reason we've used 4-core is because it's what the factory owner had lying around and he's on a cost saving exercise. Will have to have a look at pulling in a separate conductor.

They are for earthing not bonding.

 

[hightower]

They are for earthing not bonding.

Why should they be different though?

 

[Ian1981]

tables will give you the minimum CSA of steel deemed to comply with table 54.7 for armour to be used as a circuit protective conductor which is not the same as a protective bonding conductor.
Like lee says you would need the armour to be at least 85mm steel to offer the same conductivity of 10mm copper using the 8.5 ratio.

 

[hightower]

Thanks but why are they different. Why is a steel to copper equivalency that's good for a CPC not good enough for bonding?

 

[Ian1981]

Thanks but why are they different. Why is a steel to copper equivalency that's good for a CPC not good enough for bonding?

Because for a cpc the Adiabatic equation can also be used.
You often find table 54.7 gives an over sized cpc than what is needed.
Bonding requirements have no wiggle room I'm afraid

 

[Leesparkykent]

Why should they be different though?

544.1.1 and the * under table 54.8 both say about the bonding conductor being copper or a cross-sectional area affording equivilant conductance in other metals. GN8 has a section on it but I don't have it around me ATM. Probably to do with limiting touch voltages under fault conditions etc.

 

[Wilko]

Why should they be different though?

Man did I struggle with this one ... The SWA table gives k numbers suitable for use in the adiabatic equation. I haven't got it open, but I recall the factor is about 2.2. For bonding it's about voltage rise so it's the resistivity that's important and as LSK has said the factor is about 8.5. So armour is normally ok for CPC but rarely enough for bonding.

 

[Flanders]

There was a site i was on the other day with metal porta cabin's, but the earth had been exported from the pme , i was looking at using the structural steel as the earth sorce to the cabins , i spoke to nic tech , and said they could not see a problem doing that way ,not sure if that somthing you can do with your job.

 

[hightower]

Cheers guys. I've just had a read of the relevant section of GN8 but it didn't offer any reasoning behind the 8.5 factor and the difference between a CPC and bonding conductor in terms of copper equivalency.

@Wilko thanks, that answer makes a lot of sense (I think). Appreciate you all taking the effort to explain it, not being one to accept 'cos that's he way it's always been'

 

[westward10]

For the armour to be adequate for a 10.0 bonding conductor the minimum four core cable size would be 50.0. Are 10.0 bonds the required size for the factory supply.

 

[davesparks]

So what are the copper equivalent columns in that link referring to then? Reason we've used 4-core is because it's what the factory owner had lying around and he's on a cost saving exercise. Will have to have a look at pulling in a separate conductor.

They are for situations where you need to know the copper equivalent CSA, this is not used for bonding.

For bonding you need equivalent conductance, this is a different physical property. A factor of 8.5 is generally accepted as being the ratio of conductance of copper versus steel. So for a 10mm copper bond you need 85mm steel.