# rationale for main bonding

Discussion in 'Electrics UK' started by radweld, 21 Aug 2010.

Joined:
10 Aug 2009
Messages:
73
3
Location:
Sussex
Country:
The wiki on this site (and everything else I've found) says something along the lines of: "The purpose of main protective bonding is to create an earthed equipotential zone"

I'm fine with that, but why does the cable have to be so thick: my current house is undersized as it is "only" 6mm CSA.

Secondly, my water pipe is effectively an earth electrode, so while it may not be identical to to my DNO-supplied earth, what likely scenarios can cause these two "earths" to be so different as to represent a hazard?

2. ### Spark123

Joined:
2 Jun 2005
Messages:
14,829
474
Location:
Cumbria
Country:
The idea is that two exposed or extraneous conductive parts do not become more than 50v apart under fault conditions. If they are not bonded together and someone touches the two parts whilst a fault occurs they may get an electric shock.

The sizing of main PEBs depends on the supply earthing type, for TN-S 6mm may suffice (by way of the adiabatic equation) however for simplicity tables are used.
The earthing conductor for a 100A service with 25mm tails needs to be 16mm (17th edn regs table 54.7).
The PEBs should not be less than half the size of the earthing conductor, the only common size down is 10mm.
A maximum resistance of 0.05ohms for the PEBs is considered satisfactory, this figure is given by the amount of current required to blow the 100A service fuse in the required time with a max touch voltage of 50v.

For TN-CS (PME) the minimum sizes for main protective bonding are governed by the tables and enquiry to the distribution network operator, for TT a smaller cable may be used (min 6mm).

• Thanks x 1
3. ### ricicle

Joined:
14 Sep 2006
Messages:
6,344
382
Location:
Gloucestershire
Country:
Larger sizes of PEB are required on TNC-S systems due to the fact that there is the possibility of circulating network currents through them. Also there is the possibility of them carrying a proportion of load current in the event of a lost supply Neutral

• Thanks x 1

Joined:
10 Aug 2009
Messages:
73
3
Location:
Sussex
Country:
Thanks, though I'm still a bit confused. From the above, I can see one reason for the PEB is actually to provide an alternative path the earth if the neutral in a PME system fails in order to prevent earthed metal work being mains potential - though this would not be sufficient to cause the service fuse to blow.

ricicle mentions 'circulating network currents' - where is this current coming from? If the PEB must be <50mohm to keep <50V, this implies a current of 1000A.

5. ### Spark123

Joined:
2 Jun 2005
Messages:
14,829
474
Location:
Cumbria
Country:
It isn't to provide another path, it is to keep all exposed and extraneous conductive parts (E/ECPs) at or about the same potential during a fault. For example if you have a washing machine near a water pipe and the washing machine develops a fault, if they were not bonded together and someone touches both at the same time they could recieve an electric shock for the duration of the fault.
Large currents can flow when there is a fault, 1000A will clear a 100A BS1361 II fuse somewhere in the region of 0.4s which is good as time is important factor. If less than 1000A flows the fuse will take longer to clear the touch voltage between E/ECPs will be limited to less than 50v which is good as limiting touch voltages is the important factor, for example it may take up to 5s for the fuse to disconnect a fault current of 630A.
TN-CS (PME) has other hazards i.e. if the combined neutral and earth fails upstream all E/ECPs may become live at mains potential. The DNO have to apply additional earthing to this type of system in an attempt to keep this voltage down.
You can get circulating currents in TN-CS (PME) if any of your metal services are common to other premises as the service will in effect be a parallel path for neutral current to flow which is why you need to consult the DNO when sizing the bonds.

Joined:
10 Aug 2009
Messages:
73
3
Location:
Sussex
Country:
With the washing machine example, the fault current will typically flow through several metres of the earth wire in 2.5mm TE cable, and only a proportion of this will flow through the PEB. This sort of fault wouldn't usually cause the service fuse to blow. So it don't think this example shows why the PEB needs to be so chunky. I suppose you could argue a fault on a shower or cooker circuit could flow through a 10mm CPC. However, many installations do not have larger than 6mm cable and I have never seen it suggested that this would permit a smaller PEB. Also, RCD protection would surely remove the need for the PEB in the same way it does for supplementary bonding.

From some of the comments above, I think what is being said is that the PEB is dimensioned to ensure the service fuse blows. That being the case, what is/are the scenario(s) where >100A flows from the phase conductor through the PEB?

7. ### Spark123

Joined:
2 Jun 2005
Messages:
14,829
474
Location:
Cumbria
Country:
The PEB sizing is NOT to ensure the service fuse blows.
The PEB sizing needs to be adequate to ensure that it does not become damaged should fault current flow through it and also that the resistance of the conductor keeps E/ECPs at or about the same potential during a fault.
It is the job of the Earthing Conductor to provide a path for fault current to flow to ensure the service fuse blows should a fault of sufficient magnitude occur. Some current may still flow though the PEBs should this happen.
RCDs or no RCDs, you still need main PEBs.
Take the washing machine scenario again - a washing machine connects to the cold water pipe by means of a flexiable hose.
If there was no bonding in place and a fault occurs (somewhere in the system) causing the washing machine body to become live at 240v wrt earth. The cold water pipe is nicely in touch with mother earth. The potential between the two is now 240v, somebody touching both at the same time is going to get a belt.
Now add main PEBs. The fault can cause the washing machine body to become live at 240v. As it is bonded to the water pipe via the circuit protective conductor, the MET and the main PEBs the potential between the two is now limited so anyone touching the two now is only going to get a tingle <50v.

Joined:
10 Aug 2009
Messages:
73
3
Location:
Sussex
Country:
Hi Spark123,
Thanks for reply - as I said I do understand the need to bond to avoid potential differences, but the bit I am struggling with is the reason this cable is dimensioned for such high current.

Taking the washing machine fault again, could you provide an example of what sort of fault on a TN-S supply would cause a large current to flow through the PEB.

9. ### ban-all-sheds

Joined:
27 Aug 2003
Messages:
58,308
2,381
Location:
London
Country:
A fault from L to an extraneous-conductive-part?

10. ### plugwash

Joined:
28 Mar 2004
Messages:
8,861
266
Country:
Appliances are fed by thin flexes and the protective devices are deemed adequate to protect those thin flexes so appliance faults clearly aren't the reason main bonding needs to be so big.

The real reason main bonding needs to be so big is that for various reasons (most of them outside of your control). the suppliers earth can end up at a different voltage from the local earth potential. Therefore when you connect the two together and force thier potentials to be equal current will flow. How much current will depend on the voltage and impedances but it could easilly get into tens of amps and if there are faults in the network could well get over 100A and stay that way for quite some time.

Consider for example the situation where the lead sheath of a cable has corroded through and then the sheath after the break has shorted to live!

11. ### Spark123

Joined:
2 Jun 2005
Messages:
14,829
474
Location:
Cumbria
Country:
In reality the only time you need to depend on the service fuse to blow is if there is a L-E fault on the service head>meter>CU tails. Faults on downstream circuits shouldn't depend on the service fuse to disconnect.

As the meter tails etc are on the same system as the washing machine and the extraneous conductive parts such as the water main, the potential difference between exposed conductive parts (washing machine body) and extraneous conductive parts need to be kept close together should a fault like this occur.

In this example even though the fault is not on the washing machine itself the potential of bodywork of the washing will increase (with respect to mother earth) as it is directly connected to the earthing of the same system.

12. ### NotHimAgain

Joined:
8 Feb 2008
Messages:
435
20
Country:
I despair, I really do

Some basic theory for a shock received when touching a faulty appliance.

In an installation with Protective “equipotential” bonding

Ut = If * R2 volts

If the bonding is removed

Ut = If * (R2 + Se) volts

Where (for simple small installations):
Ut = the touch voltage due to the steady state fault current;

If = the steady state fault current;

R2 = the resistance of the circuit-protective-conductor(s) (cpc) within the installation – so for this example - R2 would typically include the resistance of the cpc in the final circuit conductor and the resistance of the cpc in appliance flex;

Se = the impedance (here treated simply as a resistance) of the supply earth path e.g:
1) PEN conductor in TN-C-S;
2) “Cable sheath” or similar in TN-S;
3) Earth electrode resistances of the installation electrode and the supply transformer electrode in TT.

So removing the main protective bonding conductors increases the touch voltage from:

Ut = If * R2 volts to Ut = If * R2 + If * Se volts

So this gives the mind blowing conclusion that, for the situation discussed here, main bonding removes an effect that occurs outside of the installation – good init!

Now there is more – Ut is not limited to 50 volts by this method of protection.

The only way you can guarantee Ut <= 50 volts for the whole duration of the fault (in an installation with totally effective bonding) is to arrange that the resistance of R2 is around 1/4 that of R1 (the line conductor resistance). This means you should use a 10mm2 cpc with a 2.5 mm2 line conductor.

If you use PVCTWE cables you should note the following:-

Under the conditions described above, and assuming 240 volts for Uo (the system voltage between line and earth):
1.0 / 1.0 gives Ut = 120 volts
2.5 / 1.5 gives Ut = 150 volts
4.0 / 1.5 gives Ut = 175 volts

Ever wondered why our EU buddies don’t like PVCTWE – any cable where line and cpc are the same size gives Ut = 120 volts for Uo = 240 volts.

Now don’t tell me you use this cable in your home .

The sizing of these protective bonding conductors is not related directly to the above. It has more to do with faults occurring on the supply network than those occurring within an installation. It is generally related to mechanical strength and, for TN-C-S, the ability to handle network faults, in particular a broken supply neutral – its resistance is not generally a factor.

When I was a boy main bonds were generally 2.5mm2 [D13 14th Edition] in small installations and, in terms of reducing touch voltage, they worked just as well as the larger ones used now.

For the OP
Supply network conditions have changed since I was a boy - so you should install the bonding conductors sized as recommended in BS 7671 for the type of earthing system you have. Many people just use the worst case conditions (TN-C-S) and install a minimum of 10 mm2 Protective Bonding Conductors and a 16 mm2 Earthing Conductor.

A water pipe is no longer considered to be a reliable connection to earth as most services eventually get upgraded to plastic pipework.

Now there is far more to this subject than given above so if you seek understanding you will need to do a lot of reading.

13. ### dingbat

Joined:
29 Sep 2003
Messages:
2,846
68
Location:
West Midlands
Country:
I got a bit bored half way through, but your explanation doesn't really explain the purpose of bonding, mostly because you were talking about the sizing of conductors involved in earthing. (cpc & earthing conductor)

Oh and perhaps misunderstanding the notion of 'touch voltage', as referred to in a bonded installation.

I absolutely see where you're going with your argument, by the way, but I'm not sure most would follow you. Far easier to just explain how reducing the resistance between exposed and extraneous conductive parts means that, in the event of a fault, no dangerous difference in potential will exist, don't you think?

14. ### NotHimAgain

Joined:
8 Feb 2008
Messages:
435
20
Country:
I added a bit on the end for the OP concerning conductor sizing etc.

The reasons for looking at touch voltage is that others mentioned it, and that bonding has a role in touch voltage considerations.

Taken as a whole this subject can be quite complex and I am afraid that many only have a limited grasp of it. That said, much of it can be understood with average mathematical ability.

IMO the best approach is to look carefully at each particular situation and not to get too excited by terminology such as 'equipotential'. Diagrams and basic calculations should be used for each case considered - and always remember that 'real life' is usually more complex than any model we make of it .

15. ### dingbat

Joined:
29 Sep 2003
Messages:
2,846
68
Location:
West Midlands
Country:
Just 'a role'? I think you actually ignored the role of bonding, pretty much altogether, but, as you say...
Ah. Maybe you are thinking of 'average' by reference to your own ability. It's easier to overestimate general educational standards if you are a product (like me) of a bygone age of 'actually learning things'.

For the general population, 'average mathematical ability' is very close to 'no mathematical ability'. A large proportion of electricians are barely numerate.

DIYnot News
1. ### rewiring socket ring main

andydyer8, in forum: Electrics UK
Replies:
5
Views:
219
2. ### Ring main continuity broken(?) - some sockets no longer work

Gaz42, in forum: Electrics UK
Replies:
3
Views:
220
3. ### Unconnected bonding

camerart, in forum: Electrics UK
Replies:
1
Views:
219

Replies:
29
Views:
919
5. ### Cabling procedure for adding a FCU to ring main?

alan333, in forum: Electrics UK
Replies:
5
Views:
269