What on Earth?

Talk Electrician Forum

Help Support Talk Electrician Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Only for circuits intended to supply one or more items of equipment, such that the total protective conductor current is likely to exceed 10mA, where the requirement is met by having two individual protective conductors, each complying with the requirements of Section 543.

Otherwise: do what you like. :)

 
I was fitting a nexus 45A cooker switch last week which had 2 earth connections. The metal strip between the two connections was 6mm wide by 1mm thick, and looked like galvanized steel. The instructions said to put an earth in each. But steel is 9 times less conductive than copper. So the CSA should have been 9x2.5=22.5mm squared, not 6mm squared. Which equates to reducing the CSA of the CPC by two thirds.

I used the same terminal for both cpcs.

 
Further to my post above, just spoke with bg nexus tech. They say that it is not necessarily the CSA which is important but more the surface area of the outside surface, as electrons prefer to travel along the surface of a conductor, one of the reasons why stranded wires can carry more current, the csa might be the same but the surface area is much higher. In this case the csa of the steel is 6mm sq but the surface is 14mm. By comparison a 6mm sq copper cable has a circumference of 8.7mm.

I know that the required csa calculated by the adiabatic equation would normally be less than the actual csa of the cpc.

But I still think that there is not much wiggle room in the design. BG nexus seem happy with it though and have not had any problems in the 12 years they have been making them.

 
Further to my post above, just spoke with bg nexus tech. They say that it is not necessarily the CSA which is important but more the surface area of the outside surface, as electrons prefer to travel along the surface of a conductor, one of the reasons why stranded wires can carry more current, the csa might be the same but the surface area is much higher. In this case the csa of the steel is 6mm sq but the surface is 14mm. By comparison a 6mm sq copper cable has a circumference of 8.7mm.

I know that the required csa calculated by the adiabatic equation would normally be less than the actual csa of the cpc.

But I still think that there is not much wiggle room in the design. BG nexus seem happy with it though and have not had any problems in the 12 years they have been making them.
That will be the 'skin effect' if I remember from my college days many many years ago in a land far away.

However I would be in the same camp as you Rich in thinking that earth is not man enough.

 
That will be the 'skin effect' if I remember from my college days many many years ago in a land far away.

However I would be in the same camp as you Rich in thinking that earth is not man enough.


You're correct, it is the skin effect. 

That's why for instance 2 x 2.5mm conductors will carry more current than 1 x 5mm conductor. It's why parallel feeds are so useful. 

It's also the reason that harmonics are a problem. 

 
looking at Wikipedia page there is a graph showing skin depth against frequency. Steel at 50Hz has a depth of 2mm, so the 6mm x 1mm strip would not be affected by the skin effect. So bg nexus tech is wrong. Good job too in this instance as the worry is overheating of the steel under earth fault conditions, and anything increasing the overall resistance would increase the heating effect

Copper at 50Hz has a skin depth of 9mm so no need to worry too much in a domestic environment.

 
The skin effect is an phenomena that applies to RF frequencies of many MHz and above. It is not relevant to 50Hz mains.


Not true, it's just more prevalent at higher frequencies. 

Hence why when we design MCC's that are rated to take 4KA+ the copper bars usually end up being in a triple I extrusion rather than flat bar. 

Skin effect on copper is around 9mm for 50Hz, so in a domestic situation it's not worth thinking about. 

 
Last edited by a moderator:
From https://en.wikipedia.org/wiki/Skin_effect

I really don't think the skin effect has any relevance to domestic or even industrial 50Hz mains unless you are dealing with cables more than 17mm diameter.


Which I do on a fairly regular basis. Add harmonics onto the back of that an it becomes something you need to pay attention to. 

Like I mentioned, for most electricians, not something they're going to ever have to think about. An totally irrelevant to this thread.

 
I am not sure that we should be looking simply at the relative conductance of steel vs Cu , I would expect the thermal withstand as per the adiabatic formula to be more relevant here, as that shows how its going to bear up under a fault, whereas the conductance ratio simply shows that it will increase the R2 value by the same amount as about an extra 140mm of 2.5mm copper cpc... and being a cpc under normal conditions, you'd have no, or very little current, so increased power dispatation under loads isn't a consideration.

Comparing K²S².

K= 143 for copper: 143x143x2.5x2.5 = 127806.25

K=51 for steel: 51*51*9*9 = 210681

Working it backwards, 9mm of steel can stand the same I²t as:

210681 / (143*143) = 10.328

Root of 10.328 = 3.1mm of copper

The parrallel feeds thing, as others have said for smaller cables its pretty much irrelevant, however the surface area is important still.... it affects how easily the heat can be disapated into the surrounding air

 
Top