Yet more Shed Electrics!

[Stefek Zaba]

I get what you and Lurch are driving at; the "contradiction in terms" was inaccurately forceful. But the OP did seem confused about whether their installation was TT or TN-C-S, or some deliberate or accidental hybrid inbetween.

I can quite see that PME in an urban setting, with genuinely Multiple places where the N and REC-supplied-E are bonded, and relatively short distances between local substation and consumer with several points along the way where the bond to physical earth is made, is a Good And Safe arrangement which provides a low-impedance fault path in all but the most bizarre multiple-fault circumstances. And by far the greater contribution to a low phase-to-installation-E fault impedance is the (reliable) bonding of the installation-E to the supplier's N. (OK so far? Pull me up if I'm all washed up here!) And the significance of the bonding to "local" E is (it further seems to me) to reduce any dangerous disparity between the potential attained by the installation-E-bonded metalwork (both Extraneous and Exposed ;-) and locally-referenced E, be that a damp cellar floor, water/gas pipes, etc, right? (Hence the more rigid main equipotential bonding requirments in a PME install.)

Where that reasoning seems to me to be shakier, though, is a supplier making PME available on an overhead supply. There, I'm imagining more risk of (selective) conductor breakage; greater distances to substations; higher impedance paths back to said substation/nearly-local-pole-mounted transformer; all conspiring to make the likelihood of the supplier's N, and hence their provided installation-E in a PME system, floating some significant distance away from "local" E. For such an installation, the relative simplicity of TT as a concept seems to win in my under-informed brain over the supplier's provision of a TN-C-S earth.

I really would like to understand this, not just theoretically but because it's just the situation in a scout house I feel partly morally (luckily not legally) responsible for. There we have an overhead supply, treated at the time of a thorough and competent upgrade about 20? years ago as TT (oh, it was a joy last summer when the rain and rodents got into a couple of the circuits and the 100mA whole-installation breaker kept tripping, expecially in the small hours so setting off the fire-alarm's mains-loss warning... but I digress). The transformer up the pole, which supplies that house's LV, was replaced by the REC more recently, and the pole now sports a splendid PME label. So it's possible that at the next overhaul the idea of taking the supplier's E might be seriously considered, and I'd like to know what that would mean in practice in this intermittently-occupied, hired-out-to-various-groups, occupied by groups of kids as young as 7, with dippy leaders who don't know one end of a CU from the other (fantastic as they are with the kids, mind), electric-showers-in-outbuilding, etc. etc. house!

Stefek

[Stefek Zaba]

Not all that reassuring. The old name for TN-C-S is P*M*E, and as Andy W's just been pointing out, the 'M' stands for Multiple in the phrase 'Protective Multiple Earthing'. A single fault - 'the N to E connection at the pole' - is firmly supposed *not* to give rise to any such touch voltages...

[Tim (Remove NOSPAM.]

But it did...

The particular house in question has no earth rod, but relies on a rod at the pole connected to the N. At the house the N and E connections are bonded together

The connection between house N and REC N failed. The Earth rod at the pole remained connected to the REC N.

At the house there is no earth bond as the water pipe is plastic. This meant that everything earth bonded at the house floated upto 240v with respect to the pole earth rod- so touching the sink which is supported on an insulating wooden cupboard completed the circuit through the human to ground.

Hope that makes sence?

tim..

[Mike]

'Just' isn't quite the word as the neutral terminal has been used as an earth 'star' point by the previous owner. I would propose to use one of those solid metal earth choccy blocks though is there something about it not being allowed to be mounted on the meter board but must go on the CU mounting board (in my case below) ?

I assume they are (standard Wylex) but I think you'd better enlighten me more before I do something I shouldn't.

Thanks

Mike

[Lurch]

On Sun, 9 Jan 2005 18:33:59 -0000, "Mike" strung together this:

I wouldn't worry about that. Fit an earth block in the meter cupboard and you'll be fine.

All it is is that if they're metal the tails shouldn't be allowed to come into contact with the metal case unless they are RCD protected.

2 ways around it, briefly, are;

Ensure the tails are encased in PVC when in the metal enclosure.

Run the tails through a 100mA time delayed RCD in a PVC enclosure first. You should already have an isolator anyway as there should be a common point of isolation for the entire installation.

[Mike]

Tell me about it. I've lived rurally all my life and hadn't met PME until my latest place. And the first time I realised it was when I got a tingle (42v) shock from an 'earth-bonded' item to the real damp soil underneath the house.

And this despite the previous occupant having 'hybridised' the PME with an earth rod at the other end of the house. The four farms further down the road seem to screw the supply up something rotten.

[Mike]

Totally agree with you Tim. When this site works well it works really well.

I just wish some people wouldn't try to shoot down others when those others express opinions on something that they don't agree with. That would help make it even better.

[Stefek Zaba]

All too much sense, yes. And the failure of the house-N to REC-N connection isn't really a PME matter - it could happen on the consumer's side (e.g. connection from N busbar to N side of meter works loose, or to get even more fanciful, a whole-house RCD fails open on the N side only). The extra culprit here, then, is the absence (a) of any whole-installation RCD - but you don't use one in a PME setup; (b) of any locally-referenced earth rod on the customer's premiseses - but again, you don't usually have one on a PME setup; (c) all incoming services (water, I presume no mains gas round your (rural) way) supplied through non-conducting pipework, but internal pipework being copper, hence (i presume) either deliberately bonded where the incoming pipework transitions to copper, or through something 'innocent' like the earth continuity conductor in the electrickle supply to the CH boiler, causing all the metalic pipework in contact with the boiler to be at 'house-E' potential.

Intriguing, and seems to strengthen my current prejudice against the combination of PME and overhead supply. (Incidentally, was the break in the house-N to REC-N in a final overhead section to the house? or some house-internal fault?)

Thanks for sharing the details of this one - Stefek

[Mike]

Ah - you mean that thing with the fuse in :-)

Seal is intact - if you don't look for the Araldite holding the two ends together ;-)

Seriously though, United Utilities would probably prefer to disconnect us once and for all. We get at least six power cuts a year, the voltage sits at 250V continuously with lots of glitches up and down (and we're the second farm on the chain - I assume first is over the limit) and recently I talked to somebody on here who said that we were only wired up in the 1970s to start with because he remembered wiring up one of my neighbours. Apparently before that there was no supply at all. I do wonder if PME was the 'latest thing' in the 70s and was installed without thinking about it.

[Andy Wade]

Lurch is arguing differently I think, but I was saying that "contradiction in terms" is not necessarily true.

Yes, except that there is no hybrid. It must be either TN or TT.

Lack of clarity or inadvertent conflation there: "multiple places" refers to the earthing of the supplier's combined neutral and earth (CNE) conductor. I'm not sure that the term REC has any meaning any more - let's use DNO (for distribution network operator) from now on. For consumers taking a PME supply the consumers' earthing conductors are bonded to the DNO's CNE conductor at the supply head at each point of supply - but this is a different "multiple".

Well, good and safe, in conjunction with the consumer's equipotential bonding, *unless* the DNO's CNE happens to go o/c. That's a single fault, but one they go to great lengths to guard against, supposedly.

That bit's OK. The worst case external earth fault loop impedance (Ze) is lower for PME than for old TN-S supplies because the (aluminium) CNE conductor can be assumed to have a lower impedance than an old (lead) cable sheath. The consumer thus benefits from higher earth fault current, giving reduced fault disconnection time and lower fault energy let thro' (I^2 * t) where the protection is provided by a fuse or MCB.

Yes, the local earthing will help a little in keeping the CNE and consumer's protective conductors closer to the nearby ground. It will also reduce the touch voltage if the service cable's neutral is severed, as explained here:

Where that reasoning seems to me to be shakier, though, is a supplier

Be that as it may, PME is becoming standard on overhead supplies. Under the old Electricity Supply Regulations suppliers had no obligation to provide their consumers with earth terminals. That changed in 2003, when the ESQC regulations came into force, requiring:

"24(4) Unless he can reasonably conclude that it is inappropriate for reasons of safety, a distributor shall, when providing a new connection at low voltage, make available his supply neutral conductor or, if appropriate, the protective conductor of his network for connection to the protective conductor of the consumer's installation.

"24(5) In this regulation the expression "new connection" means the first electric line, or the replacement of an existing electric line, to one or more consumer's installations."

For such an installation, the relative simplicity of TT as a concept

A properly designed, installed and maintained TT installation is arguably safer; certainly it's safer where the equipotential zone concept is difficult or impossible to apply. OTOH TT has some disadvantages: it's less idiot-proof then PME, and relies more on good maintenance - i.e. regular testing of the earthing and RCD operation. RCDs are not fail-safe devices; some which rely on internal electronics can fail to operate under conditions of reduced supply voltage, which can occur under fault conditions. The 'official' national committee viewpoint therefore seems to be to prefer TN over TT, unless there's a really good reason otherwise.

Does the place have dry floors? If so I don't see any reason not to use PME - with all the usual caveats of course: main and supplementary bonding in good order; RCD protection for socket circuits feeding outdoor equipment; possible separate TT earthing for any exported outdoor or outbuilding supplies - especially if Class 1 equipment is going to be used outdoors. Also important is to establish, and not to skip, an appropriate periodic inspection and testing frequency - with the certificates issued by a properly-insured entity.

[Lurch]

On Sun, 9 Jan 2005 19:54:08 -0000, "Mike" strung together this:

Er, no. Isolation covers both poles. A seperate switch in an enclosure would be a good idea.

Oh dear!

They don't usually provide PME on overheads, I don't know why they sometimes do. I haven't seen any recent PME\overhead supply combos recently so maybe they don't anymore.

[Lurch]

On Sun, 09 Jan 2005 19:53:00 +0000, Stefek Zaba strung together this:

If the neutral break occured on the customers side it wouldn't matter as the reason the metal becomes live is because the neutral and earth are bonded together on the service head. If the live doesn't get back as far as this then there won't be an instance of all the metalwork becoming live.

[Ian White]

Andy Wade - did you receive my attempt at a direct e-mail?

[Stefek Zaba]

Andy Wade wrote, after a most useful piece of instruction on PME:

Essentially, yes - GF is nearly all suspended floors, solid 'crete in one area but proper DPC, so not notably damp; one sort-of outside lav with a dampish floor, no sockets there but a light with plateswitch. So no massive confusion between "local" earth and the putative PME earth.

Yeah well, "mostly" in place. Main and supp bonding are in place, but not to PME specs. For example, the supply to the showers I mentioned in the outbuildings (it's only just an outbuilding - a narrow passageway separates the loos-n-shower-block from the main house) have a (by eye)

6mmsq single-core earth (lacking extra mechanical protection) running alongside the 6mmsq? T&E (both suspended from a catenary, couldn't hand on heart assure you the catenary's properly earth bonded but it's likely to be). They're not instantaneous electric showers - there's an immersion heater as the only means of heating in each of two hotwater tanks. The pipework throughout is all copper internally, there's a metallic main incoming water feed in a water cupboard close to the showerblock which these days feeds through some 6m of MDPE to the main house. AFAIR the main bonding to the internal pipework is close to the point of entry, where there's an MDPE-copper junction; and from memory it'd be a 6mmsq bond rather than the 16mmsq? most typically associated with PME. The earthing block is made off to an earth rod just by the outside wall, and a 100mA time-delay RCD controls all final circuits.

The main disboard is a 36-way MEM 'industrial' jobbie - not all ways are populated, but the majority are. There's enormous application of diversity - the supply incomer is 100A, the real heavy loads are for electric water heating, our abstemious lifestyle when the place is in use (and the absence of portable electric heaters, the big woodburning stove in the main-house kitchen/dining room, and bottled gas for catering-quantity cooking) justify the large number of final circuits - mainly 20A radials - going off to multiple dormitories and smaller bedrooms. There's also an Ingeneous arrangment involving a contactor, this disboard, and the fire alarm control panel, which forces the occupier to enable the fire alarm panel before turning on the power - the same single key fits the alarm panel and the contactor control switch, and can be withdrawn from the fire alarm panel for use with the main contactor only once the panel's keyswitch is in the 'on' position.

RCD provision for external powertools is "under way" - I've fitted captive RCD plugs onto the two extension reels most often used for external power, provision's recently been made for an externally-mounted "weatherproof" 13A socket with integral RCD, but it's not ideal. Separating off the GF radials through an add-on CU with built-in 30mA RCD would be better.

Supply to the outbuildings which are at a slightly greater distance is (from memory) done differently for the two outbuildings. There's a kitchen block which has the main house earth exported to it - a 63A MCB feeds a 'submain' (another catenary-flown T&E with supp earth conductor, this time the main T&E looking like 10mmsq, the supp earth is our 6mmsq friend). That block's maybe 10m from the main building, and there's bonding in place from its incoming water supply to the CU in the kitchen.

The other outbuilding (the original farmhouse) has yet another catenary-flown supply, AFAIK a 6mmsq T&E, but this one has no supp E conductor but has instead a 100mA RCD by its CU and a local earth rod. This outbuilding's a little further from the main house, but not by much (say 20m rather than the 10 for the kitchen block). It's an "independent" TT installation, though I'd have to check (a) the original wiring diagram, with subsequent hand-drawn alteration, and (b) the actual connections, to know which side of the main-house RCD this building's feed is from. Externally-used equipment is in practice most often powered from sockets in this building and its associated cellars and nearly-attached garages. And thinking about it, this outbuilding's floors in the semi-cellars are dampish - so even if the main house were overhauled for PME, I think we'd need to keep this one as a "separate" TT installation the way it is now.

All of which rambling leads me to conclude that your final words:

are definitely appropriate; and I'll raise it at our yearly planning meeting which happens to be a couple of weeks away. As a voluntary organisation we won't necessarily like the expenditure which any "not to current regs" comments would imply - but we have a sympathetic but independent local sparks who won't (I hope!) go over the top with "ALL YOUR KIDS ARE AT TERRIBLE RISK" where it's not true. (The risks my amateur eye sees are for the volunteers who help maintain the place, rather'n the kids and other groups who use it).

Thanks again for your detailed reply, Andy - Stefek

[Dave Stanton]

Which site is that then ?

Dave

[Tim (Remove NOSPAM.]

Well I can use an isolator at the house end OR move the T/D 100ma RCD from the CU to the house and combine it either with a HRC fuse or keep the 63a MCB.

Which do u reckon guys??

tim..

[John Rumm]

Not quite sure I follow what you are proposing there...

What you need to arrive at is a situation where the SWA is protected from P-E and P-N faults before it reaches the outbuilding consumer unit(s). Assuming your 100A isolator is downstream of your TD RCD, then that will ensure the SWA is protected it from P-E shorts. Your MCB will hopefully[1] protect it from P-N shorts. Since the upstream RCD is a TD type, the outbuilding CUs can have their own standard 30mA RCDs for socket circuits and not loose discrimination.

The difficulty I discovered when I did the sums for my setup, was firstly: making sure that I got discrimination in the event of a current overload (i.e. the MCB in the outbuilding should trip and not the one in the house). Secondly avoiding cable damage in the event of a fault in the SWA (in my case it was complicated by poor choice of pre-existing cable sizes).

When you study the response curves of the MCBs you realise that just because the upstream one is a higher trip current, it does not guarentee it will not trip in the event of a large overload (i.e. a short). In your case a 63A upstream of a 32A one in the outbuilding CU may well be ok. In my case cable sizes dictated more closely spaced cascaded current limiting devices, and hence a HRC fuse worked out better choice for the house end.

[1] As Andy Wade highlighted on the CU Wiring Scheme thread I started a couple of weeks back, you need to compute the prospective short circuit current that could occur in the SWA. The fault current will be dictated by the impeadance of the cable and supply. The lower it is, the greater the potential fault current. A high fault current will result in the cable being able to susstain it for less time without damage, but will also result in the protective device breaking the circuit faster. You need to check that the protective device gets there first!

[Tim (Remove NOSPAM.]

Many thanks for this John :)

Tim..