Bonding extraneous parts

[Jako13!]

Do extraneous conductive parts such as hand railing have to be bonded back to the MET or can they be linked to a known earth ie uni strut or cable tray?

 

[sheppertonspark]

To be an extraneous conductive part the metalwork has to be capable of introducing a potential into the Building. Unless the hand rail is connected in some way to structural steelwork of similar this is unlikely. We then call it isolated unearthed metalwork. Check the regs. there is a specific ohmic value to exceed for it to require bonding.

 

[Jako13!]

To be an extraneous conductive part the metalwork has to be capable of introducing a potential into the Building. Unless the hand rail is connected in some way to structural steelwork of similar this is unlikely. We then call it isolated unearthed metalwork. Check the regs. there is a specific ohmic value to exceed for it to require bonding.

What about unistrut?

 

[Julie.]

Extraneous conductive part is something that has the potential to bring external ground voltages into the property, it has to be in contact with the ground (like a metalic water pipe from outside).

Something that can become live due to a fault in the item, or if a cable fails to its supports etc is not Extraneous, its an exposed conductive part and may require supplementary bonding.

Check section 415.

Basically, you check the resistance, then you need or don't need it

 

[pc1966]

From memory if its above 22k then you are unlikely to get any serious risk (even if you+MET go to 230V you would see no more than 10mA assuming human body as 1k), also if below 0.05 ohms (and an allowance for X meters of 10mm cable) it is already well-bonded.

Small stuff and support stuff like Unistrut, etc, that measures such that bonding is reasonably required you should also use your judgement: Is it likely to be touchable and the person simultaneously in contact with anything else? Is it used for something that might fail to make it live? Is there any part that someone might grab and, due to shock current, be unable to release?

In terms of bonding to nearby metal (instead of all the way back to the MET) then it certainly serves the purpose of maintaining equipotential regions, but you should check how well bonded that metalwork is back to the MET and what sort of CSA is used to see if any likely fault current can be handled.

Usually for external extraneous parts like pipes, structural steel, etc, you have the requirement for large conductors of the 10mm and above size (especially for TN-C-S) as you could see sustained fault currents in the many tens of amps range, but for internal hand rails, etc, that is unlikely so the sort of size used for CPCs would be reasonable.

Again the resistance test will tell you something about this. For example, if the handrail is showing 1k then you are unlikely to see high currents no matter what so, say, 4mm bonding of related metal via CPC or similar is fine, but if it is a couple of ohms then you might see hundred amp sort of faults and should treat it as you would for extraneous pipes, etc.

But don't believe me - get a copy of Guidance Notes 8 and read that!

 

[Jako13!]

From memory if its above 22k then you are unlikely to get any serious risk (even if you+MET go to 230V you would see no more than 10mA assuming human body as 1k), also if below 0.05 ohms (and an allowance for X meters of 10mm cable) it is already well-bonded.

Small stuff and support stuff like Unistrut, etc, that measures such that bonding is reasonably required you should also use your judgement: Is it likely to be touchable and the person simultaneously in contact with anything else? Is it used for something that might fail to make it live? Is there any part that someone might grab and, due to shock current, be unable to release?

In terms of bonding to nearby metal (instead of all the way back to the MET) then it certainly serves the purpose of maintaining equipotential regions, but you should check how well bonded that metalwork is back to the MET and what sort of CSA is used to see if any likely fault current can be handled.

Usually for external extraneous parts like pipes, structural steel, etc, you have the requirement for large conductors of the 10mm and above size (especially for TN-C-S) as you could see sustained fault currents in the many tens of amps range, but for internal hand rails, etc, that is unlikely so the sort of size used for CPCs would be reasonable.

Again the resistance test will tell you something about this. For example, if the handrail is showing 1k then you are unlikely to see high currents no matter what so, say, 4mm bonding of related metal via CPC or similar is fine, but if it is a couple of ohms then you might see hundred amp sort of faults and should treat it as you would for extraneous pipes, etc.

But don't believe me - get a copy of Guidance Notes 8 and read that!

Thanks for you replies it’s hard to refer to the regs as it’s on machinery and plant. I know there’s a separate regs but not too sure on what they state. As far as I know BS7671 covers us to the isolater. Just wasn’t sure if I got a low reading it I could bond straight if the cable try to the extraneous part as the cable tray will be bonded back to the panel and then back through the supply to the MCC.

 

[pc1966]

I don't know what the other regs might be or say, but electricity is basically the same before and after the isolator switch!

Check the resistance to earth of the item(s) and look to find out the CPC cable size all the way back to the MET. If it appears that all of the CPC is easily able to handle at least double the fault current the resistance would suggest, and it is physically robust (another reason the wiring regs has minimum sizes for bonding), then doing it that way ought to be OK.

 

[Jako13!]

I don't know what the other regs might be or say, but electricity is basically the same before and after the isolator switch!

Check the resistance to earth of the item(s) and look to find out the CPC cable size all the way back to the MET. If it appears that all of the CPC is easily able to handle at least double the fault current the resistance would suggest, and it is physically robust (another reason the wiring regs has minimum sizes for bonding), then doing it that way ought to be OK.

Why’s it double the fault current?

 

[pc1966]

The x2 factor was just my suggestion for a safety margin for measured earth impedance against what it might be following a wet spell, or if the resistance at 230V is much lower than few volts of multimeter, etc, used to measure it.

In most cases something like 2.5mm (physically protected) or more generally 4mm is going to be fine (in line with supplementary bonding 544.2.1) but if you have something internal such as a hand railing that is big enough to have a resistance to Earth below, say, 10 ohms (but not near-zero implying it is already bonded to the MET) you are looking at a possible fault current above 23A, so that sort of cable size is maybe not going to be acceptable.

You could go straight for the main protective conductor sizing (e.g. based 544.1.1 or PME case from Table 54.8), or consider it from a realistic fault current point of view given internal stuff is not likely to find itself bonded (or not) to other structures outside of your control (as service pipes could be).

So if bonding to other metalwork that has an existing CPC back to the MET the obvious question is what will happen under fault cases and would it overheat the CPC? Hence my suggestion of measuring it and seeing what might happen before deciding if an existing CPC is going to be OK or if it really ought to have substantial bonding all the way back to the MET.