I've just discovered that the kitchen sink in the house I've just moved into has no earth bonding (plastic hot & cold supplies).
I'm going to fit a waste-disposal unit, so that would provide an earth through it's metalwork to the sink.
Is this enough, or should I run earth cable from the sink back to the consumer unit (& if so, how thick would it need to be - about 25' from the sink to the CU)
thanks,
Jim
Kitchen sinks do not require 'earth bonding'. Supplementary bonding of such extraneous-conductive-parts is only necessary in bath and shower rooms.
It's OK.
That's not necessary, but make sure that the main bonding from the main earth terminal to the incoming gas and water service pipes (& oil, if applicable) is in place, close to the point of entry of each service. The minimum size for this is 10 mm^2.
What if incoming water pipe is plastic? (no oil or gas) The Earth rod is then sufficient??
Tim..
No - if all or most of the house pipework is in copper (or other metal, naturally), then that should be bonded to your main earth terminal close to the transition from plastic-incomer to copper-internal - usually this'll be at your internal stopcock. The idea is that the bulk of the metal pipework in your house should provide a low-resistance path for any fault current: that way, if some fault (e.g. in a hot water boiler, immersion heater, or less likely misplaced nail touching-but-not-severing L-in-cable and going on to a water pipe - I said less likely, OK!?) brings your supply Live into contact with that pipework, (a) a big enough current will flow to make any fuse/MCB pop nice and quickly, (b) even while the fuse/MCB is deciding how quickly to pop, the pipework doesn't rise to a voltage far above earth (cos the low resistance of its connection to earth means most of the 240V is dropped on the L side of the fault).
But if your pipework's all or nearly-all plastic (e.g. just visible copper tails going to radiators), there's no need to bond the short segments of copper pipework.
HTH - Stefek
If the house is plumbed in copper the need for main bonding is unaffected. If the house is plumbed mostly in plastic then no main bonding to the water service is required.
Err, what earth rod? Are you talking about a TT installation?
snip
I have difficulty with the standard answers to the question of bonding in a kitchen
(a) a big enough current will flow to make any fuse/MCB pop
hopefully yes, but a fuse/mcb won't trip on the current/voltage that gives a fatal shock - you need an RCD - even then that takes 0.4s - or
40 half cycles each reaching a peak 340volts.(b) even while the fuse/MCB is deciding how quickly to
Something like that was the subject of the recent inquest. The metal shelf went to 240v & the earthed casing of the (?)dishwasher stayed resolutely at zero volts.
I wonder whether over-earthing/bonding in a kitchen can compromise safety: perhaps more regard should be paid to the merits of relying on natural protection from clothes, shoes etc & minimising exposed metal surfaces in a kitchen.
One might argue that the MCBs purpose is to protect the installation from overload, not to prevent electrocution. That is what the RCD (if fitted) is for.
Had adequate RCD protection been in place then this would probably not have not ended in a fatality though.
Can work both ways I guess. Typically there is no requirement for supplementary bonding on a kitchen anyway, and there are pretty strong arguments for earthing appliances in these circumstances. A faulty earth on an appliance could also result in a severe shock hazard when combined with other faults.
One might, but one would be in error. The little slogan is "protection by EEBADS" - that's protection of people, against electric shock, by Earthed Equipotential Bonding and Automatic Disconnection of Supply. The
*point* of the low-resistance earthing we commonly practice (not just the supplementary bonding in bathrooms, but earthed cases on metal appliances) is to produce a big enough flow so that the MCB or fuse will operate fast enough. RCD's are a useful top-up, esp. where good earthing can't be assured (portable appliances outdoors being the clearest case), but good old EEBADS is the design and implementation which has kept death-by-electrocution down to very low levels in the UK for, yay, 40 yearses or more.Stefek
No it doesn't. That may be the spec, but the RCD will probably trip at the next zero crossing point (or the next one in a particular direction, anyway).
The spec for a 30mA RCD was designed to ensure adequete shock protection without leading to unnecessarily complex design or excessive nuisance tripping.
Indeed. The IEE recommend even in bathrooms that any metal surfaces should be isolated if possible, rather than earthed. It is only if they will be earthed anyway (perhaps at a different potential to other parts) that supplementary bonding is required.
Christian.
Yup agreed - perhaps my statement was too hard on EEBADS! it is indeed a major life saver. However there are still catagories of faults for where no amount of earthing is going to give sufficent protection though, like the incident refered to above, or the classic mower/hedge trimmer through the cable scenario.
Leaving aside the merits or otherwise of RCD protecting and/or bonding everything in sight, I thought the specification for the "life saving"
30mA trip RCD was 40ms, not 400ms (i.e. 0.04s, not 0.4s) though admittedly this 40ms is at 5x30mA. Having said that, even 150mA isn't a lot of current - even a really manky earth should allow that amount through (V/I=1600R).Hwyl!
M.
"Martin Angove" wrote | (jim_in_sussex) wrote: | > you need an RCD - even then that takes 0.4s - | Leaving aside the merits or otherwise of RCD protecting | and/or bonding everything in sight, I thought the specification | for the "life saving" 30mA trip RCD was 40ms, not 400ms
The requirement is for a 0.4s disconnection time (through EEBADS) on socket circuits, whether that is achieved by fuse, MCB or RCD, and is in the wiring regs. The BS for the RCD is a different matter.
Owain
Yes, I realise that the requirement for the circuit is 0.4s (or 5s depending on the circuit), but Jim wrote something which implied to me that he thought that an RCD took 0.4s too; the bit which went:
In testing RCDs I have to admit that even at 1x current I've only once seen a (30mA, non-delay) RCD take longer than 40ms to trip, and that was because it was mechanically knackered. At 5x current most seem to manage around 10ms. Again though, I realise that this is a feature of the RCD, not a specific requirement of the BS disconnection times.
Hwyl!
M.
It's probably quite difficult to design an RCD that doesn't instantaneously trip at the next crossing point (or the next crossing point of a particular polarity). It would need some sort of additional time delay mechanism, like those intended for TT system incomers.
Christian.
I see what you mean - that if there is sufficient current flowing (about
1x the operating current) the thing should trip. Does it trip at the crossing point (when the current is effectively zero), or at the peak voltage, hence peak current? The tests I carry out (and which usually involve a lot of running up and down ladders) are quite interesting:1: at half rated current the device should not trip, tested at 0 degrees, 180 degrees and then 0 degrees again.
2: at rated current it should trip within a specified time, test and discard reading at 0 degrees and then take reading at 180 degrees and 0 degrees again.3: at 5x rated current it should trip within the specified time (e.g.
40mS), three tests as in 2:That's usually nine tests in all, and the nearest convenient socket at which to plug the tester in is rarely within arm's reach of the RCD :-)
Hwyl!
M.
It trips when the current is zero. Anything else would torch the contacts. Even a manual switch will "let go" at a crossing point if there are any. This is why a normal switch is rated higher for currents at AC than DC.
Note that the voltage being zero will only coincide with current being zero if the circuit is power factor corrected. But switching voltage is easy. It's current that hurts.
I fitted a double socket next to the consumer unit for one ring circuit. The other ring circuit is for the adjacent room anyway...
Christian.
There's a difference between "letting go" and "actuating". MCBs have arc extinguishing chambers precisely because they are highly likely to trip just about anywhere *except* at a zero point - they are largely mechanical devices after all with absolutely no electronics. Are RCDs any different? My gut feeling was that the coil would be actuated at a certain current - which is likely to be other than at the zero voltage point - and that at that point the thing is mechanically "tripped". Quite how long it takes for the arc to extinguish is another matter. I have never opened up an RCD but would be surprised if there is any electronics in there to "hold off" until the zero point, merely in the interests of extending contact life.
Obviously, but for domestic use it's usually good enough to say that the pf is reasonably good unless there's a lot of fluorescent lighting about.
Hwyl!
M.
Well, the mechanism notwithstanding, the current will flow until the next zero crossing point. I didn't mean to imply that there was any magic or electronics in this, just that it is what AC switches do.
However, RCDs are often much more complicated than MCBs and do often have internal analogue electronics, rather than being simple mechanical/inductive devices. This means that they could be designed to have the ability to sense an impending zero crossing point before attempting to move the contacts. I don't know if they actually do that, though, but I'd guess not.
The effect on the system as a "black box" though is that the current will cease at a zero crossing point whether the RCD attempts to nurse its contacts or not.
Christian.
HomeOwnersHub website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.