Sub-mains and more

[westie101]

I strongly we are likely, if not already, be running around a hamster reel with this discussion.

Electrically I think you are in a most ways correct that it is hard to justify the need for another piece of protective equipment - in fact in some ways it could be detrimental.

However for a number of non-electrical reasons there is seen to be a benefit in installing the device. For these reasons the ESI decided over 30 plus years ago that the length of tails (face it an unprotected submain is in effect a set of tails) without addition fuse protection should be 3m.
The point of supply being the meter outgoing terminals is the starting point.

This physical point also marks the changeover from the Supply Industry to the IEE regulations which may also have some relevance to the discussion.


In the real world, as we know, life is not that simple, this whole argument being negated in some ways by the huge amount of supply cables (both armoured and not) that run through properties with their only protection being the fuse at the local substation.

 

[JohnW2]

I strongly we are likely, if not already, be running around a hamster reel with this discussion.

Indeed, particularly now we appear to have achieved almost complete agreement.

Electrically I think you are in a most ways correct that it is hard to justify the need for another piece of protective equipment - in fact in some ways it could be detrimental.
However for a number of non-electrical reasons there is seen to be a benefit in installing the device.

Exactly, and that benefit deriving from those non-electric reasons (legal, political or whatever) is primarily a benefit to the DNO, not the consumer, and not safety.

If I'm allowed to be philosphical at the weekend, it does seem a great shame that we live in a world in which political, legal or whatever considerations can overturn decisions based on sound scientific or technological principles. Ah well - time to open a bottle, I guess

Kind Regards, John

 

[westie101]

If I'm allowed to be philosphical at the weekend, it does seem a great shame that we live in a world in which political, legal or whatever considerations can overturn decisions based on sound scientific or technological principles. Ah well - time to open a bottle, I guess

Though considering that this situation has been in operation for the 37 years I have been industry, it seems unfair to blame it on today's society!

As for the bottle, I've a 3l box of something in the fridge!

Cheers

 

[westie101]

If I'm allowed to be philosphical at the weekend, it does seem a great shame that we live in a world in which political, legal or whatever considerations can overturn decisions based on sound scientific or technological principles. Ah well - time to open a bottle, I guess

Though considering that this situation has been in operation for the 37 years I have been industry, it seems unfair to blame it on today's society!

As for the bottle, I've a 3l box of something in the fridge!

Cheers

 

[JohnW2]

If I'm allowed to be philosphical at the weekend, it does seem a great shame that we live in a world in which political, legal or whatever considerations can overturn decisions based on sound scientific or technological principles. Ah well - time to open a bottle, I guess

Though considering that this situation has been in operation for the 37 years I have been industry, it seems unfair to blame it on today's society!

Oh, I'm not! The "world we live in" to which I referred has been around for a very very long time. Politics and religion, in particular, have over-ridden logic and fact/science/whatever since almost the year dot. If anything, it's got a bit better over time - at least we now give some credence to scientists etc, rather than burn them at the stake but, more recently (decades), the increasingly intrusive and tight regulation of everything ('nanny state' some would say) may (IMO) have turned that clock back a bit.

I assume from what you write that you are an engineer in the electricity industry - and I guess what I'm saying is that, in that position, I would be frustrated by feeling obliged to try to defend a situation which I ultimately had to admit could not be justified in terms of engineering. It's a phenomenon I am only two well aware of - my primary profession (far divorced from electricity, but I have several different professional hats, plus a few non-professional ones!), in which I co-incidentally have also been working for 37 years, is a prize example, in which I very often find it professionally frustrating and difficult to do/explain/defend many of the things which go on, usually for political reasons!

As for the bottle, I've a 3l box of something in the fridge!
Cheers

Cheers to you, too, but if its in the fridge, it's either the wrong colour for me or is too cold!

Kind Regards, John.

 

[JohnW2]

Have a browse through Chapter 43 and all will become clear !

OK, I've browsed! It looks to me as if 433.2.2 is probably saying that the 10 mm² cable to my immersion CU is OK in terms of overload protection (that being provided by the MCBs, about 1 metre downstream), and 435.2 allows for overload and fault protection to be separate. What I am having more difficulty in working out is whether the 80A fuse is acceptable as fault (short circuit) protection for that ~1 metre of 10 mm² cable.

OK, although we spun off into some interesting discussions, my question didn’t really get answered and, as I wrote, I couldn’t really do the sums without knowing what PFC to design for. So, since my philosophy has always been “if in doubt, do the experiment”, I did – albeit fairly crudely.

Increasing demand of my installation by 38.2A resulted in the line-neutral supply voltage (measured close to meter) falling by 2.8V, which gives a supply impedance of 0.073 Ω . Assuming a supply voltage of 240V (on low load, it’s usually around 243V), that would imply a PFC of 3288 A.

Taking a ‘k’ value of 115 (PVC, 70 degrees) for my infamous 10mm² cable, that calculates as a maximum disconnection time requirement of 0.12 seconds. An 80A BS1361 fuse (which is what I would have if I changed to a MEM KMF800) gives a 0.12 second disconnection time with just over 1000A (and the time/current graphs in the regs stop just after that) – so, per what I ‘knew’ anyway, the fuse appears to be very adequate for protecting the cable against short-circuit fault. Coupled with the downstream overload protection (2 x 16A MCBs, well able to protect 10mm² cable), it therefore looks to me if 10mm² cable (or the 80A fuse, whichever way you care to look at it) is more than compliant.

Any comments?

Kind Regards, John.

 

[NotHimAgain]

I have not paid much attention to most of this discussion so you may have covered these points.

1) Length of 'tails' - the DNO may allow the use of their protective device to provide overload and short circuit protection provided they are satisfied that the protection is effective. Each DNO has their own view on the requirements so these tend to vary a bit around the country (2 to 4 metres being quite common). However, BS 7671 only allows the use of this device if the DNO agrees that it affords protection - 433.2.2 (iii) & 434.3 (iv).

So what protection do we need:
a) Overload;
b) fault current;
c) shock protection.

a) Is generally ok provided the cable size is adequate to carry the rated current of the device (In) - given the particular installation conditions.

b) Depends on which type of fault we are considering, and also on circuit length. In a domestic installation the circuit we are considering will generally meet the requirements section 412 (Double or Reinforced Insulation) - so faults to earth are not really the issue. This also rules out c), and this is just as well as I doubt that any DNO would confirm shock protection disconnection times - so your on your own if you have any extraneous-conductive-parts associated with the circuit i.e you must make your own assessment.

So its phase to neutral (or phase to phase if three phase). The longer the circuit - the higher its impedance - the lower the fault current - the longer the disconnection time - the greater the thermal stress.

If you want to check this look at the Appendix 3 graph for a BS 1361 60A fuse. Take a reading for a fault current of 1000A (about 0.15 seconds), and then one at 200A (about 5 seconds).

The lower current would occur in the longer circuit. The let-through-energy at 1000A is 150000 A²s and at 200A it is 200000A²s. So low values of fault current can cause more thermal damage than high ones .

So my comments on your design are - have you considered the lowest values of fault current that might reasonably be expected .

 

[JohnW2]

I have not paid much attention to most of this discussion so you may have covered these points.

1) Length of 'tails' - the DNO may allow the use of their protective device to provide overload and short circuit protection provided they are satisfied that the protection is effective. Each DNO has their own view on the requirements so these tend to vary a bit around the country (2 to 4 metres being quite common). However, BS 7671 only allows the use of this device if the DNO agrees that it affords protection - 433.2.2 (iii) & 434.3 (iv).

My point/question about this has been .... if, for whatever reason, the DNO do not agree that their fuse provides adequate protection for long tails, the standard solution is for the consumer to install a second fuse, similar or sometimes even identical to the DNO one, a metre or so downstream of the DNO's one, at the start of the long tails. I am at a complete loss to understand how this would improve protection - and even westie101 (who appears to have been speaking for DNOs) eventually agreed that there is no 'electrical' reason for requiring this.

So its phase to neutral (or phase to phase if three phase). The longer the circuit - the higher its impedance - the lower the fault current - the longer the disconnection time - the greater the thermal stress. ..... So low values of fault current can cause more thermal damage than high ones .
So my comments on your design are - have you considered the lowest values of fault current that might reasonably be expected .

That's a very interesting point which I have often wondered about - and it is a totally general question, nothing specific to 'my design' (which is inherited, so the design is not 'mine'!).

Nor is it just a question of cable impedence. One also has to consider the nature of the fault. The concept of designing for PFC assumes that the fault will be a 'dead short' (i.e. approximately zero impedence), but no-one has told cookers and other things about this! In terms of the provision of protection (but obvioulsy not the cause), realisation of this seems to totally blur the distinction between 'overload' and 'fault' protection - since an appropriate fault could result in any degree of 'fault current', from negligible up to very high.

The regulations appear to deem a cable to be able to carry a current of 1.45Iz for fairly short periods of time without damage, effectively requiring overload protection to operate within an hour for currents greater than this.

I really need to to plot some graphs and see what is going on, since there presumably must be a 'worst case scenarion' (probably at a fault current much less that PFC) for which one should be designing. Whilst, as you illustrate by your example (and becaue of the characteristics of the protective device), the thermal damage may be greater with a 200A fault current than a 1000A current, this thermal damage (or 'let through current', if you must!) presumably does not carry on increasing as fault current decreases all the way down to 1.45Iz - so, there must be a 'peak' to this curve somewhere - which, to my mind, is really the fault current (not PFC) for which one should be designing. ... or am I missing something?

Kind Regards, John.

 

[NotHimAgain]

BS 7671, and all other standards I am aware of, always assume a 'bolted fault' i.e. a short circuit of zero impedance. This is because it is often the worst case and, in any event, it is generally too difficult to consider impedances higher than zero when drafting regulations.

If the DNO require that you fit your own protective device it is generally because they are unable or unwilling to confirm that their device would suffice. If you end up fitting an identical device to theirs that is fine, but it is now your responsibility not theirs.

Their is no requirement for selectivity (discrimination) between the devices as the situation is no worse than if you just had the DNO device - and there is no inherent danger. It is in fact very difficult to engineer selectivity with a BS 1361 100A fuse over the whole range of fault currents - but that is another story .

I see that you have noticed that the graphs stop at 0.1 second in Appendix 3. I am afraid that the 'real life' situation is far more complex than Appendix 3 appears to indicate. First the current plotted is the sinusoidal rms value - in fact the fault current will be asymmetrical and around twice the peak value for about the first two cycles (40 ms) and it will then decay to the sinusoidal value in around 100ms or five cycles. This is why the graphs ignore the first 100 ms of the fault. If disconnection occurs within this period (and it often does) the let-through-energy cannot be calculated it has to be found by experiment. This is the reason for BS 7671 434.5.2 paragraph 2.

The adiabatic equation is only valid between t >= 0.1 to t <= 5 seconds. However, it does produce reasonably good 'engineering' solutions within its range.

At the end of the day a 100A BS 1361 fuse generally saves your bacon regardless

 

[JohnW2]

BS 7671, and all other standards I am aware of, always assume a 'bolted fault' i.e. a short circuit of zero impedance. This is because it is often the worst case and, in any event, it is generally too difficult to consider impedances higher than zero when drafting regulations.

Indeed, but as you chose to point out, zero impedance may well not be (probably rarely is, with real-world protective devices) the 'worst case', and that rekindled doubts that I've always had

If the DNO require that you fit your own protective device it is generally because they are unable or unwilling to confirm that their device would suffice. If you end up fitting an identical device to theirs that is fine, but it is now your responsibility not theirs.

Exactly. As westie101 eventually agreed, any such requirements are for 'non-electrical' reasons!

.....I see that you have noticed that the graphs stop at 0.1 second in Appendix 3. I am afraid that the 'real life' situation is far more complex than Appendix 3 appears to indicate. This is why the graphs ignore the first 100 ms of the fault. If disconnection occurs within this period (and it often does) the let-through-energy cannot be calculated it has to be found by experiment. This is the reason for BS 7671 434.5.2 paragraph 2.

That's all very well, but that para relies upon knowledge of the I²t value 'as quoted by the manufacturer' for disconnection times under 0.1 sec - and, like Appendix 3, they generally don't quote it! As for 'by experiment', I don't think that would be a practical approach to designing an installation.

The adiabatic equation is only valid between t >= 0.1 to t <= 5 seconds. However, it does produce reasonably good 'engineering' solutions within its range.

I realise that, but it's perhaps unfortunate that it doesn't encompass the disconnection times (<0.1 sec) which, as you imply, we probably should be allowing for in our designs.[/quote]

At the end of the day a 100A BS 1361 fuse generally saves your bacon regardless

That, of course, is what I've always believed - but then people start throwing regs at me, so I have no choice but to throw back reg-based arguments - and often find (as here) that the regs are actually a little lacking

So, where does your pragmatism start? You have quoted an example of a situation in which if you were working with a PFC of 1000A, and designed by calculation the system to just be adequate to protect against that, the protection might theoretically be inadequate for a fault current of 200A. Would you be happy to go ahead with the PFC-based design, hence assuming that lower fault currents than that would not occur?

Kind Regards, John

 

[NotHimAgain]

That's all very well, but that para relies upon knowledge of the I²t value 'as quoted by the manufacturer' for disconnection times under 0.1 sec - and, like Appendix 3, they generally don't quote it! As for 'by experiment', I don't think that would be a practical approach to designing an installation.

The 'experiments' are carried out by the manufacturer when they design a device - they have the data - often available on their web sites or in technical literature.

How much effort you put into any given design depends upon the potential risks. In a domestic installation the risk are relatively low.

The maximum prospective fault current will be quoted by the DNO - this could be as high as 16kA at the point of connection between your service line and the distribution main (in the road). A value of 16kA used to be quoted at the cut-out in parts of London due to the nature of the distribution network but this is usually OTT. The quoted figure would allow you to estimate the maximum required breaking capacity of your protective devices in the worst case. The fault level reduces with cable length and so by the time you get to your device it would rarely exceed about 5kA. People often quote higher measured values but unless they are next door to a sub-station they are probably inaccurate.

In any event you don't need to worry in the UK because we have a magic fix called the 'conditional' rating - its 16kA provided certain conditions are met - even for a 1kA rated rewireable fuse - nice . I am not going to explain all of that as I would be here all night - see BS EN 60439-3 1991 Annex ZA if you want proof.

So for a domestic the first part of the design - selecting breaking capacity - is not relevant thanks to the conditional rating.

Next I said you should check the minimum fault current situation. If you know the supply characteristics (and you really should if you are trying to design a circuit) you can work this out. It will often be the case that a fault to earth will produce a lower current flow than a line to neutral fault.

You could just take the worst case figures for your type of supply - the DNO will quote an earth fault loop impedance of 0.35 ohms for a TN-C-S supply and 0.8 ohms for TN-S. Taking this approach will, in many cases, 'over engineer' the installation.

In practice this sort of effort is not generally called for - however, you are supposed to check these things out - particularly for any long cable runs.

So, where does your pragmatism start? You have quoted an example of a situation in which if you were working with a PFC of 1000A, and designed by calculation the system to just be adequate to protect against that, the protection might theoretically be inadequate for a fault current of 200A. Would you be happy to go ahead with the PFC-based design, hence assuming that lower fault currents than that would not occur?

That is what most of us do - as I said faults with impedance are not easy to accommodate.

BTW I do have £1M Professional Indemnity Cover - it costs me more than £1k per annum .

 

[JohnW2]

In any event you don't need to worry in the UK because we have a magic fix called the 'conditional' rating - its 16kA provided certain conditions are met - even for a 1kA rated rewireable fuse - nice . ... So for a domestic the first part of the design - selecting breaking capacity - is not relevant thanks to the conditional rating.

Well, whatever one feels about the legitimacy of that 'conditional rating', as you say it's 'nice' (at least conventient!) for us. However, it's not really breaking capacity that we (at least I) have been discussing (i.e. we've been assuming that was adequate) but, rather, the disconnection time - since that (together with the fault current) is what determines whether or not the cable is adequately protected.

....Next I said you should check the minimum fault current situation. If you know the supply characteristics (and you really should if you are trying to design a circuit) you can work this out.

You did indeed say that. However, it subsequently transpired that you were talking of the minimum fault current given a 'bolted short' - which, as you demonstrated with an example, is not necessarily the worst case

Would you be happy to go ahead with the PFC-based design, hence assuming that lower fault currents than that would not occur?

That is what most of us do - as I said faults with impedance are not easy to accommodate.

Maybe not easy to accomodate, but perhaps a little surprising that regulations do not insist that one accommodates them, given your illustration that they could (contrary to initial intuition) represent a greater threat to cables.

I think this whole thread illustrates why regulations often get a bad name (should I mention the scapegoat 'Health & Safety'?!). If you look back, you will see that in addition to the general question about the protection of long tails/submains, I mentioned the essentially trivial matter of a CU I have serving just two immersion heaters (2 x 16A Type B MCBs) which is fed by a 'sub-main' <1 metre in length (actually about 0.6m), which appears to be 10mm² and whose 'only' upstream protection is an 80A fuse and the DNO's fuse. This has generated considerable discussion, during which I have repeatedly been referred to the regulations, despite the fact that it is 'obvious' to me that this situation is fine.

One thing that this whole discussion about the sizing/protection of those tiny lengths of 10 mm² cable has ignored is the fact that any fault which arose would almost certainly be in the final circuit, wired in much smaller cable. If the MCB is considered OK for protecting that smaller cable (and I'm sure that everyone agrees that it is), then the larger (10mm²) feed cable is surely also thereby adequately protected (against faults arising in the final circuit) - since 'disconnection' (anywhere in the current path) is disconnection! Quite apart from all the discussions about numbers, that leaves reliance on the fuse only for protecting 0.6 metres of cable from a short arising prior to the MCB - I would dare to suggest that a reference to Mr Jobsworth might be appropriate here

Kind Regards, John

 

[JohnW2]

Edit: not sure what happened here; I must have pressed some wrong button, since I managed to re-post my last posting - now deleted. Apologies!