IET logo
 
IET
Decrease font size
Increase font size
Topic Title: 2-core instead of 3
Topic Summary:
Created On: 14 November 2017 04:06 AM
Status: Read Only
Linear : Threading : Single : Branch
1 2 3 Next Last unread
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 14 November 2017 04:06 AM
User is offline View Users Profile Print this message



yobby1

Posts: 4
Joined: 30 October 2017

I have proposed the use of 25mm^2 3-core SWA for a 100a sub-main installation for a workshop. The run itself is about 28 metres, and in twin wall duct for addition protection/expandability.

The cost is significantly higher than desired, so I am exploring the possibility of reducing to 2-core, and use the SWA for CPC (will be glanded w/ the usage of earth nuts). Should I compromise on this issue?
 14 November 2017 08:06 AM
User is offline View Users Profile Print this message



alanblaby

Posts: 754
Joined: 09 March 2012

Less than £50 difference doing a quick google.

Are you exporting the earthing from the source to the workshop?
Has it got any extraneous parts that will need bonding?

My thoughts - don't skimp if you are exporting the earth. It is a very small increase in price, so keep it at 3 core.
If you are putting the workshop on its own earth (TT'ing it) then 2 core is fine, as the armour will only be connected at the source end, and is only there for protection of the cable.
 14 November 2017 08:53 AM
User is offline View Users Profile Print this message



daveparry1

Posts: 7745
Joined: 04 July 2007

I agree with Alan, if you are exporting the earth use 3 core. The last similar one I done a couple of years ago was about 120m of 70mm swa but as I was TT'ing the workshop I did use 2 core on that one but I think 3 core is more reliable if using the supply earth.
 14 November 2017 09:57 AM
User is offline View Users Profile Print this message



davezawadi

Posts: 3905
Joined: 26 June 2002

Technically it is completely unnecessary to have an earth core in 25mm BS SWA. The armour cross section is entirely adequate whether the load is TT or not, and for a workshop it is not wise to use TT as you will suffer tripping problems with machine tool starting current. 13A sockets will still need RCD protection, but BS4343s (Blue plugs) for machines will not if you choose the rating carefully. 25mm SWA copper equivalent is also adequate for the main bonding of any metalwork, but it is unlikely you will have any steel gas or water supplies. Remember a domestic 100A supply which is TNC-S only needs a 16mm main bond under all circumstances. 2 Core SWA tends to be comparatively expensive because much less is used, but still you can save some money. You should find 25mm 2 core for less than £5.00 per metre if you look carefully. Try your local wholesaler.

The duct is also completely unnecessary, so you can use direct burial as the install method and save a fair bit of money. If you need telephone and network cables it might be convenient, but as these can also be buried you can save a lot, at the inconvenience of needing to dig to replace them. Do not ignore the force requited to pull a 25mm cable, particularly if the run is not straight, it could be very difficult unless you have the right kit available. It would be wise to install an earth electrode local to the workshop whatever the supply type, to keep the local ground somewhere near the earth conductor potential.

-------------------------
David
BSc CEng MIET
david@ZawadiSoundAndLighting.co.uk
 14 November 2017 10:55 AM
User is online View Users Profile Print this message



mapj1

Posts: 9568
Joined: 22 July 2004

Is it possible to do cost reduction by transfer of effort - that is can you get the customer to provide a ready dug trench, put holes in the right walls and so on, leaving you to do the clever bit.

-------------------------
regards Mike
 14 November 2017 10:56 AM
User is online View Users Profile Print this message



AJJewsbury

Posts: 16007
Joined: 13 August 2003

25mm SWA copper equivalent is also adequate for the main bonding of any metalwork

Are you sure about that David? According to (my outdated copy of) GN 8, 25mm2 2-core SWA has an armour c.s.a. of 42mm2 - for PME bonding you need equivalent conductance of say 10mm2 copper - which would be around 90mm2 in steel.
- Andy.
 14 November 2017 05:02 PM
User is offline View Users Profile Print this message



yobby1

Posts: 4
Joined: 30 October 2017

Thanks for the suggestions guys, very valuable input.

There are no extraneous parts that require bonding.

Supply is TNCS, and I am exporting the earth in current design. I could go down the TT route, will discuss options I guess!

Whilst we're on this topic, has anyone used a 100mA time delayed RCD at source to protect the run from a L-E fault? Atm I am only using a switched fuse which is fine for short circuits, but there is no L-E protection. My reasoning for omitting the RCD is the shear nuisance of double trips. Also, considering the fact that the run is SWA, and in twin duct, any damage to the cable is very unlikely. However, parts of the run do go through some joists and are clipped onto the exterior wall.
 14 November 2017 05:19 PM
User is online View Users Profile Print this message



mapj1

Posts: 9568
Joined: 22 July 2004

When you say no LE protection, is there a reason you expect the switched fuse not to do its fusey thing ? Is Zs very high ?

Only time I'd normally require a type S or an earth fault relay (same idea, more options) at the start of armoured cable is when source impedance cannot guarantee to operate the ADS in time.

(so perhaps on a TT supply or certain genset configurations.)

edited to change " iz Ss" to "is Zs" hopefully clearer

-------------------------
regards Mike


Edited: 14 November 2017 at 05:26 PM by mapj1
 14 November 2017 05:22 PM
User is offline View Users Profile Print this message



daveparry1

Posts: 7745
Joined: 04 July 2007

I used a 100m/a time delay rcd at the supply end on the job I mentioned above.
 14 November 2017 05:24 PM
User is offline View Users Profile Print this message



davezawadi

Posts: 3905
Joined: 26 June 2002

Yobby, the run is protected from a L-E short by the CPD (presumably a 100A BS88 fuse). As it is TNC-S you only need to ensure the submain Zs is < ~0.4 Ohms for a 5 second disconnection time which is easily met by the cable as specified. (Isc >= 570A). Note that there will be no discrimination with the REC fuse, and it will fail first unless it is rated at 100A or more, even then a close run thing and random. However as it is only protecting the SWA against faults the chance of this is minute. You should use Type C breakers in the CU in the workshop for your machine and welder supplies, of suitable ratings.
Regards

-------------------------
David
BSc CEng MIET
david@ZawadiSoundAndLighting.co.uk
 14 November 2017 05:25 PM
User is offline View Users Profile Print this message



ArduinoXR

Posts: 48
Joined: 16 August 2017

Originally posted by: davezawadi

for a workshop it is not wise to use TT as you will suffer tripping problems with machine tool starting current.


Hey Dave, could you expand on that a little please? Why would inrush current cause more nuisance tripping in a TT system?
 14 November 2017 05:41 PM
User is offline View Users Profile Print this message



davezawadi

Posts: 3905
Joined: 26 June 2002

Hi
Yes this can be a problem with RCDs, as the transformer does not have exact cancellation of currents for practical production reasons, and during current spikes is also rather non-linear. This leads to output from the sense winding which will cause trips (sometimes!). The effect can be reduced with time delayed types, but in this kind of install there is no need for RCDs on machines anyway as they are not portable appliances, and do not use BS1363 plugs and sockets. You may refer to the BYB for more info on higher rated sockets to BS4343 (EN30609-2). RCDs are very useful for many things but are not ideal devices and really can never be at a sensible price. Correctly wired machines do not pose any significant electrical danger, although may easily kill or injure you for other reasons. Sudden loss of the mains supply could also present danger with circular saws or spindle moulders or large grinding machines where excessive load could be presented to cutters or wheels. Many machine tools also have electronic motor drives and controls which may present more than 10mA of leakage current, so you are better off without the RCDs, and perhaps with extra earth bonding wires to give additional protection against cable damage. I assume that you intend to get some largish machines although single phase supplies give additional sourcing problems or need phase converters. We can provide more answers on these if needed.

-------------------------
David
BSc CEng MIET
david@ZawadiSoundAndLighting.co.uk
 14 November 2017 08:38 PM
User is offline View Users Profile Print this message



sparkingchip

Posts: 10104
Joined: 18 January 2003

Originally posted by: yobby1

I have proposed the use of 25mm^2 3-core SWA for a 100a sub-main installation for a workshop. The run itself is about 28 metres, and in twin wall duct for addition protection/expandability.



The cost is significantly higher than desired, so I am exploring the possibility of reducing to 2-core, and use the SWA for CPC (will be glanded w/ the usage of earth nuts). Should I compromise on this issue?


What is at the supply end?

A house with a suppliers 100 amp fuse holder that actually has a 60 or 80 amp fuse in it?

If it is a house or other building with a 100 amp fused supply, what proportion of the available supply is needed in that house or building?

Andy B.
 16 November 2017 10:37 AM
User is offline View Users Profile Print this message



ArduinoXR

Posts: 48
Joined: 16 August 2017

Originally posted by: davezawadi

Hi

Yes this can be a problem with RCDs, as the transformer does not have exact cancellation of currents for practical production reasons, and during current spikes is also rather non-linear. This leads to output from the sense winding which will cause trips (sometimes!).


Hey Dave. I don't understand why taking a PE from a local electrode results in the transformer not having 'exact cancellation of currents'. Can somebody please explain?
 16 November 2017 12:07 PM
User is online View Users Profile Print this message



AJJewsbury

Posts: 16007
Joined: 13 August 2003

I don't understand why taking a PE from a local electrode results in the transformer not having 'exact cancellation of currents'. Can somebody please explain?

Two different issues - RCDs are imperfect and can sometimes (often?) not quite cancel properly when very large currents flow through them - say the L and N windings aren't quite perfectly identical and given say a current of 100A one registers the equivalent of say 100.000A correctly while the other is slightly out and registers the equivalent of 99.995A - in effect seeing 5mA leakage "in error" - normally that's all within acceptable tolerances and matters not. But during say a motor startup and transformer switch-on when there's a large inrush current you might see say 500A flowing for a very short time - during that time the RCD's "error" will also be 5x as large - so increase to around 25mA and so likely cause the RCD to trip even though there's no fault anywhere.

Similar issues can happen with slightly uneven magnetic saturation of the RCD's core - meaning the error can suddenly increase above a certain current level.

That's all the same regardless of what means of earthing you have. The point about local electrodes is that Zs will be much higher and so you'll almost certainly need RCDs in situations where you could manage without them on a TN installation.

If you're forced into a TT situation (like you'd be in much of southern Europe) then higher rated RCDs and time delays can help to reduce the problem.

- Andy.
 16 November 2017 12:31 PM
User is online View Users Profile Print this message



geoffsd

Posts: 1760
Joined: 15 June 2010

Originally posted by: AJJewsbury

25mm SWA copper equivalent is also adequate for the main bonding of any metalwork


Are you sure about that David? According to (my outdated copy of) GN 8, 25mm2 2-core SWA has an armour c.s.a. of 42mm2 - for PME bonding you need equivalent conductance of say 10mm2 copper - which would be around 90mm2 in steel.


May I query that?

Whilst the regulation does indeed state that it shall be of equal conductance; what does that mean in relation to the installation?
Whilst the armour does have an resistance 8.5 times that of copper, surely for the cpc/earth conductor, as long as Zs is within limits then it will be satisfactory.

As for use as a main protective bond, in the regulations there is no mention of any value of the impedance of the bonding conductor so with what value does it have to have equal conductance?

For example, a 10m. 10sq.mm. copper conductor has an impedance of 0.018 ohms.
Why would you need an armour with this same resistance?
Even at 8.5 times this, it would still be only 0.15 ohms. Would this be unacceptable?




Edit - removed calculation as I think I went wrong.

Edited: 16 November 2017 at 12:43 PM by geoffsd
 16 November 2017 01:57 PM
User is online View Users Profile Print this message



AJJewsbury

Posts: 16007
Joined: 13 August 2003

As for use as a main protective bond, in the regulations there is no mention of any value of the impedance of the bonding conductor so with what value does it have to have equal conductance?

My guess would be that similar conductance is a way of asking for similar current carrying capacity - so not related to length or overall impedance. Diverted N currents in PME systems can sometimes be quite substantial.

Some of the GNs caution against using cable armour and similar as a bond in PME systems as it could then be carrying some current pretty much all the time - warming the cable and so reducing the current carrying capacity. Same argument for using the 3rd core of 3-core SWA as a main bonding conductor of course. Some think that using the 3-phase rating (3 cores loaded) of the cable instead of the 2-core rating would normally compensate adequately for that though.

- Andy.
 16 November 2017 04:13 PM
User is online View Users Profile Print this message



geoffsd

Posts: 1760
Joined: 15 June 2010

Is the 'conductance' (for this situation) not determined by the k values, rather than the impedance?

The heating and reducing of ccc may be a reason for NOT using the armour as main bonding at all, but doesn't the fact remain that when used it does not have to be 8.5 or 9 times the csa of a copper conductor, merely k2/k1xs.
 16 November 2017 05:07 PM
User is online View Users Profile Print this message



AJJewsbury

Posts: 16007
Joined: 13 August 2003

Is the 'conductance' (for this situation) not determined by the k values, rather than the impedance?

I don't believe so - the "k" values represent how much heat the conductor can absorb before it hits its thermal limits - two conductors with the same electrical resistance but of different materials with differing heat capacities (e.g. how many Joules you shove in for a 1 degree temp rise) would have different "k" values. K values only really make sense when you ignore heat loss from the cable (i.e. you assume all the resistive heating goes to increase the temperature of the conductor) - as we do when using the adiabatic for fault conditions - it makes a lot less sense for conductors carrying current for a 'long' time - e.g. bonding conductors (especially PME ones).

- Andy.
 16 November 2017 05:42 PM
User is online View Users Profile Print this message



geoffsd

Posts: 1760
Joined: 15 June 2010

Originally posted by: AJJewsbury

I don't believe so - the "k" values represent how much heat the conductor can absorb before it hits its thermal limits

But isn't that what is required? - the reason for increasing mpbs to 10sq.mm. for pme; it wasn't to reduce the impedance.
IET » Wiring and the regulations » 2-core instead of 3

1 2 3 Next Last unread
Topic Tools Topic Tools
Statistics

New here?


See Also:



FuseTalk Standard Edition v3.2 - © 1999-2017 FuseTalk Inc. All rights reserved.

..