Steel Bath how to earth.

[JohnW2]

at some point I will undoubtedly decide that the time has come for me to get on with the rest of my life proper.

Oh dear, I guess you didn't read the small print when you signed up
DIYNOT rules: - Pressing accept indicates you agree to making lifelong contributions to the forum. Forum members may only leave after producing a signed copy of their enlistment into the foreign legion.

We'll see. You will have seen what I've written to others.

Kind Regards, John.

 

[JohnW2]

Absolutely, I only joined a matter of weeks ago and wish I had joined much earlier. It has been a significant learning experience for me too. I am also inspired by the fact that there are electricians/electrical engineers, DIYers and interested professionals from other professions with so much knowledge of electrical science, electrical regulations and full of challenging questions.

I think one small value of people like me is that of an outsider's 'new pair of eyes'. Within any discipline, ingrained ways of thinking, and sometimes even 'myths', can arise - so it's sometimes useful for an outsider to barge in and rock the boat a little with points/questions/views which might not occur to the 'in crowd'.

Kind Regards, John

 

[ebee]

Lets say you have a 16th compliant bathroom.
No RCD for existing circuits.

Lets say now , under the 17th, you are adding a heated towel rail.

Lets say you do not intend to RCD the existing circuits (one person might disagree with you on that!).

Not ALL of the conditions are satisfied therefore supplementary bonding must be present even though your new circuit on its own complies.

 

[JohnW2]

Lets say you have a 16th compliant bathroom.
No RCD for existing circuits.
Lets say now , under the 17th, you are adding a heated towel rail.
Lets say you do not intend to RCD the existing circuits (one person might disagree with you on that!).
Not ALL of the conditions are satisfied therefore supplementary bonding must be present even though your new circuit on its own complies.

Fair enough.

What if (as was probably common) the existing bathroom was not 16th ed compliant, in that it had no supplementary bonding. If you added the towel rail, with RCD protection, but without providing RCD protection to the existing circuits in the bathroom, would you then argue that it would be compliant (even if not sensible) to leave the room without supplementary bonding, on the basis that the new work satisfied the requirements for ommiting it?

Kind Regards, John.

 

[sparkticus]

That is surely the classic misconception about RCDs? .... it would only be true that the RCD will disconnect "'before the voltage reaches 50v" if the touch voltage increased slowly, which is not what usually happens. When, as is the case in practice, faults usually appear suddenly and completely, all one can say is that, with a EFLI <1667&#937;, the 30mA RCD will disconnect if the touch voltage (at the point where the EFLI was measured) is >50V. If, as is the common situation with faults, there is sudden low impedance (regs talk about zero impedance) L-E fault, then the touch voltage will surely rise the the full supply voltage (230v or whatever, within a maximum of 0.005 sec {quarter of a cycle}) until the RCD operates.

Exactly and since we do not know the exact phase angle at the instant of the fault then the initial PD could be anywhere between 0v and 230v (RMS) In other words the full impact of 230v could be near-instantaneous.
Other confounding physical phenomena such as Power Factor (I & V out of phase with one another), any stored energy in equipment capacitors and any back EMF from inductive loads does not make the situation any safer or predictable (an under statement)

 

[sparkticus]

Lets say you have a 16th compliant bathroom.
No RCD for existing circuits.
Lets say now , under the 17th, you are adding a heated towel rail.
Lets say you do not intend to RCD the existing circuits (one person might disagree with you on that!).
Not ALL of the conditions are satisfied therefore supplementary bonding must be present even though your new circuit on its own complies.

Fair enough.

What if (as was probably common) the existing bathroom was not 16th ed compliant, in that it had no supplementary bonding. If you added the towel rail, with RCD protection, but without providing RCD protection to the existing circuits in the bathroom, would you then argue that it would be compliant (even if not sensible) to leave the room without supplementary bonding, on the basis that the new work satisfied the requirements for ommiting it?

In such a situation (not adding an RCD) I would say not compliant and would want to provide supplementary bonding to ensure that extraneous conductive parts (parts that were near or at MET potential by chance or at some risk of becoming so by chance) where all fixed at the same potential (neglecting any volt drop on bonding conductors etc)

The reason for me doing that would be to reduce the risk (during a phase-earth fault on the towel rail) of the rail outer conductive parts creating a significant PD with respect to that of the other conductive parts in the room. In other words the entire conductive elements of the room would rise (for the disconnection response time - about 0.4 seconds) to roughly the same potential with minimal PD between them - therefore reducing the intensity of any potential shock hazard.

 

[JohnW2]

Exactly and since we do not know the exact phase angle at the instant of the fault then the initial PD could be anywhere between 0v and 230v (RMS) In other words the full impact of 230v could be near-instantaneous.

I wasn't really quibbling about those insignificant fractions of a cycle one might have to wait for the instaneous voltage to rise to its peak (and, of course, we should be talking about 325V, not 230V). I only added that comment, rather than just saying 'immediately' in case some smart-**** tried to pick me up on the fact that the a.c. supply probably wouldn't be at it's peak at the moment a fault occured!

My main point, of course, was that it's a common misconception, even amongst electricians, that an RCD will somehow limit the touch voltage to some value (e.g. 50V) dependent upon the EFLI. As I hope I explained satisfactorily, that simply isn't true - all an RCD can do is limit the time for which a high touch voltage is present - but that voltage may well be close to the supply voltage until that disconnection occurs.

Kind Regards, John.

 

[JohnW2]

In such a situation (not adding an RCD) I would say not compliant and would want to provide supplementary bonding to ensure that ...

So would I, for the reasons you state. It's interesting, however, that (essentially by the application of common sense) we sometimes choose to consider the (17th ed.) compliance of parts of the installation which have not been altered, but sometimes not.

Kind Regards, John.

 

[sparkticus]

Exactly and since we do not know the exact phase angle at the instant of the fault then the initial PD could be anywhere between 0v and 230v (RMS) In other words the full impact of 230v could be near-instantaneous.

I wasn't really quibbling about those insignificant fractions of a cycle one might have to wait for the instaneous voltage to rise to its peak (and, of course, we should be talking about 325V, not 230V). I only added that comment, rather than just saying 'immediately' in case some smart-**** tried to pick me up on the fact that the a.c. supply probably wouldn't be at it's peak at the moment a fault occured!

My main point, of course, was that it's a common misconception, even amongst electricians, that an RCD will somehow limit the touch voltage to some value (e.g. 50V) dependent upon the EFLI. As I hope I explained satisfactorily, that simply isn't true - all an RCD can do is limit the time for which a high touch voltage is present - but that voltage may well be close to the supply voltage until that disconnection occurs.

I fully agree and I only touched on the phase angle at the instant of the fault for the same reason (an attempt to head off "smart" comments) but also for the reason that you mention (some people's & some electricians misconceptions about RCDs) that they somehow limit current and/or voltage which we know is absolutely not the case. All they do is start to disconnect when the current reaches or approaches the trip threshold.
In the time between initiating disconnection and actual disconnections the current will continue to rise and the voltage can be at full amplitude instantaneously or at least within quarter of a cycle (5 milliseconds)

In fact for that reason and for the reason that I see many RCDs fail to disconnect at all, I put little faith in them at the moment.

 

[sparkticus]

In such a situation (not adding an RCD) I would say not compliant and would want to provide supplementary bonding to ensure that ...

So would I, for the reasons you state. It's interesting, however, that (essentially by the application of common sense) we sometimes choose to consider the (17th ed.) compliance of parts of the installation which have not been altered, but sometimes not.

Absolutely. I am a firm believer of "one size fits all only where it touches"
I want to make my choice based upon my experience/knowledge despite the fact that both need improving

 

[Paul_C]

... in essence, 'what I was missing' was the impedance of the CPC. Even if all the extraneous-conductive parts are effectively connected together (and connected to MPB/MET), then the p.d. which can arise between those e-p-cs and an exposed-conductive-part which becomes live is primarily determined by the impedance of the CPC which is earthing that exposed-conductive part.

The overall loop impedance external to the installation will have a big effect on that though. Your 10m of 1 sq. mm earth conductor will have a resistance of approx. 0.2 ohm. If that's connected to the main earthing terminal of, say, a TN-C-S installation, then in the event of a short it will form an appreciable proportion of the total loop impedance. But if it's a TT installation, the resistance of the earth electrode will be the main current limiting factor, since even with a resistance as low as 10 ohms the fault current will be limited to 24A or so. That 0.2-ohm earth conductor will then show very little voltage drop, but the main earthing terminal of the installation will rise to a much higher potential, bringing any other metalwork bonded to it to a similar potential. In other words, the higher the Ze, the less difference you'll see between those pieces of exposed metalwork in the absence of supplementary bonding between them in the event of a fault.

If the EFLI measured at the exposed-conductive part is low, and hence fault current large, then this p.d. (essentially the voltage drop along the CPC) can be very high. For example, with just 10 metres of 1mm² CPC, at a fault current of 1000A, the voltage drop, hence p.d. of interest in the bathroom, would by my reckoning be around 220v.

Don't forget that you need to allow for the resistance of the live supply conductor though. For example, if the supply cable is 1 sq. mm T&E, then the resistance of the live line will be the same as the earth, so 0.2 ohm. So even before accounting for the external loop impedance, the fault current is limited to 600A maximum by the resistance of the cable. Where the resistance from the main earthing terminal of the installation back to the supply transformer is no higher than the resistance of the live side of the supply, as with a typical TN-C-S arrangement, that immediately limits the voltage which can be developed across that 10m run of earth conductor to no more than half the supply voltage (higher if that 1 sq. mm earth is used in conjunction with a 1.5 sq. mm live conductor, of course).

 

[JohnW2]

The overall loop impedance external to the installation will have a big effect on that though. ... In other words, the higher the Ze, the less difference you'll see between those pieces of exposed metalwork in the absence of supplementary bonding between them in the event of a fault.

Sure. I was just talking in general terms about worst case scenarios.

Don't forget that you need to allow for the resistance of the live supply conductor though. For example, if the supply cable is 1 sq. mm T&E ...

Yes, that's another factor which will restrict touch voltages. However, once one gets above 1mm T&E, the live conductor becomes appreciably fatter than the CSA, thereby reducing this effect.

Kind Regards, John.

 

[JohnW2]

In fact for that reason and for the reason that I see many RCDs fail to disconnect at all, I put little faith in them at the moment.

Indeed, in addition to the 'misconceptions', there is the issue of reliability, particularly when RCDs are not regularly tested/'exercised'. The more I know/think about RCDs, the more I worry about the general public's (and even a few electrician's) apparent perception that they provide guaranteed immortality!

The reliability issue is at least one reason why I decided to retain 100mA Type S RCDs as the 'main isolators' in my (TT) installation even when all the circuits came to be protected by their own RCDs/RCBOs. Admittedly the time delay appreciably increases the hazard but with such 'belt and braces' I would hope that at least one of the RCDs would operate when it should!

Kind Regards, John.

 

[Paul_C]

The more I know/think about RCDs, the more I worry about the general public's (and even a few electrician's) apparent perception that they provide guaranteed immortality!

Indeed, as I've mentioned before, there seems to be an alarmingly widespread belief that the RCD will provide protection against all known electrical ills. I do not believe in throwing an RCD at everything, and while acknowledging the extra safety they can bring, I'm certainly not worried about not having RCD protection in all the places it's now specified (by BS7671).

But even where an RCD is present, I would never take that as an excuse for not taking some other precaution which I would take if that RCD were not present. In other words, if I wouldn't omit some particular bonding between two pieces of equipment if the RCD were not there, I'm not going to omit it just because an RCD has been added to the design.

 

[ebee]

Well said that man