Hob & oven circuit

[wrightsonm]

For a 5s disconnection, the loop impedance required for a B40 is way less than that.

The rough calc was to consider that the EFLI would be unlikely to be anywhere near that value considering the csa of the cable. Without performing any measurements it would likely be many times smaller, so the fault current would be many times higher. I also was skipping over the time/current charts to establish the minimum EFLI to achieve the 5s disconnection time. At this stage i hadn't even looked up what the max permitted EFLI should be (0.88R) , was merely using engineering judgement and putting a smaller cable in didnt feel right. I have a background in safety critical design, so i can be overly pessimistic with margins as a consequence! Maybe i could have done a better job of formulating some of my thoughts and replies....

Indeed - in fact, one fifth of the figure mentioned (before correcting for Cmin!) - and that because he doesn't seem to understand that the current required to produce an assured magnetic trip of B40 is 200A, not 40A.

Yes you require 200A fault current to achieve the disconnection in a time between 0.1 and 5s. And an overcurrent condition would trip somewhere under 60A, albeit slowly (2000s ish). These presumably being worst case figures since trip current is stated elsewhere as 3 to 5x nominal rating. Again I was too quick in replying with too many assumptions...

You still seem unclear on the difference between overloads and faults.

Whilst i do understand the difference between the two, i think i may have been interchanging them between replies, and over thinking a few things. With a low enough ELFI, the 2.5mm cable will be protected against faults. It was the theoretical and improbable over current condition i was getting worried about, needlessly since a single point would have a single fused plug connected to it.

Which indicates that you are neither clueless nor uninformed.

Can anybody here recommend a few decent guides to the regulations, basic theory etc?

I picked up the OSG, 7671 and the testing guidance note from the IET as I am a member and they offer a decent discount. Ive even got a copy of the regs from 1965 that describe the rules from which this house was wired by! I havent read them cover to cover, more using them as reference guide to check things as i go (like this one). Yes any guides would be appreciated, thank you.

Ultimately im keen to:
a) get it right
b) fully understand the justification
c) have 0 issues arise at the point of inspection (still 1st fixing at the moment)

Maybe i should start a separate thread for your enjoyment to share the wide range of attrocities that i have been uncovering recently.

This is one of the better ones - sellotape!

Many thanks
Mark

 

[wrightsonm]

@ban-all-sheds - I was reading another post via google as I have forgotten to provision for my cooker hood, and spotted a rather dated post (2005) from you that fits with this same topic:

https://www.diynot.com/diy/threads/powering-a-cooker-hood.25547/

When I said "discuss", it wasn't a rhetorical, or sarcastic remark. A cooker circuit is of course a radial, so a circuit coming off of it is not a spur, it is a leg of the radial, and there are no rules explicitly limiting the number of sockets on it - you're into the realm of designing the circuit properly. There is no simple "only 1 socket" rule that "prevents" more from being added, which is why you do have to be very careful, and which is why I agree that you should not attach a 2.5mm² leg to a 32A 6mm² radial, except via an FCU to provide proper protection of the cable. A spur on a ring is a special exemption to the Ib <= In <= Iz rule, and it does not apply to other circuits.​

 

[JohnW2]

Whilst i do understand the difference between the two, i think i may have been interchanging them between replies, and over thinking a few things. With a low enough ELFI, the 2.5mm cable will be protected against faults.

I think you are probably making the same mistake again. As I have said, fault protection is nothing to do with 'protecting cables' - it's simply to do with achieving the required disconnection times in the event of an L-E fault (of 'negligible impedance'), under 'worst case' conditions.

If the EFLI of an entire circuit (i.e. measured at the 'furthest point' of that circuit) is low enough that a fault (deemed to be of 'negligible impedance) will result in a current of at least 5x (the 'worst case') the In of the protective device (hence guaranteed to provide the required disconnection time, even in the 'worst case'scenario), then the fault protection is considered adequate - and that remains true whether he cable CSA is 1mm² or 100mm² (or any other size that would result in a low enough EFLI) .

Kind Regards, John

 

[EFLImpudence]

A cooker circuit is of course a radial, so a circuit coming off of it is not a spur, it is a leg of the radial,

The definition in Part 2 for a spur is: A branch from a ring or radial final circuit.

and there are no rules explicitly limiting the number of sockets on it

There are if you use a reduced csa cable.

you're into the realm of designing the circuit properly.

All circuits should be designed properly.

There is no simple "only 1 socket" rule that "prevents" more from being added,

The one socket 'rule' is only informative in Appendix 15 for the circumstance shown. If you design the circuit such that it need not apply, then it doesn't.
It applies to the spurs we have been discussing because of the nature of the circuit.

which is why you do have to be very careful, and which is why I agree that you should not attach a 2.5mm² leg to a 32A 6mm² radial,

What do you mean agree? You are the only one who believies it.

except via an FCU to provide proper protection of the cable.

Please explain how the FCU does anything the plug fuse doesn't.

A spur on a ring is a special exemption to the Ib <= In <= Iz rule, and it does not apply to other circuits

No it isn't and yes it does.

Single socket, 13 ≤ 13 ≤ 27; Double socket, 26 ≤ 26 ≤ 27.



It would be better if you posted saying "I don't understand so and so, could you please explain so that I might learn?" rather than post statements that are wrong and need correcting lest someone else should read them and think you know what you are talking about.
​

 

[wrightsonm]

@eflmpudence - your comments above are not about something I wrote, but something @ban-all-sheds wrote several years ago. Since it was directly related I added a link, and copied the relevant post hoping @ban-all-sheds could clarify.

 

[wrightsonm]

I think you are probably making the same mistake again. As I have said, fault protection is nothing to do with 'protecting cables' - it's simply to do with achieving the required disconnection times in the event of an L-E fault (of 'negligible impedance'), under 'worst case' conditions.

Maybe i am mis understanding something. If you have a L-E fault, a significant increase in current occurs. The temperature of the cable will begin to rise, and will overheat and become damaged assuming no adequate protection mechanism is in place. But we have a B40 MCB and tested that the EFLI is less that stated maximum of 0.88R so the MCB disconnects within 5s limiting the temperature rise in the cable and preventing the cable from being damaged. Forgive me, but is this not protecting the cable?

 

[JohnW2]

Maybe i am mis understanding something. If you have a L-E fault, a significant increase in current occurs. The temperature of the cable will begin to rise, and will overheat and become damaged assuming no adequate protection mechanism is in place. But we have a B40 MCB and tested that the EFLI is less that stated maximum of 0.88R so the MCB disconnects within 5s limiting the temperature rise in the cable and preventing the cable from being damaged. Forgive me, but is this not protecting the cable?

Well, yes, if the circuit is properly designed, the cable is protected by the MCB (or other over-current protective device) from (thermal) damage due to over-current at any current above its CCC, whether that current be 1.2 times its CCC or 50 times its CCC. It has been determined, and is hence deemed to be the case, that if the CCC of a cable is no greater than the In of the protective device, then the nature of the disconnection time/current curve of the device is such that no particular current will be allowed to flow for long enough to thermally damage the cable (whether that current be 1.5 x In for an hour, or 10 x In for 0.2 seconds, or whatever). That is 'overload protection' and you will not be surprised to hear that its purpose is to protect the cable - and the adequacy of a particular over-current device (e.g. MCB) to protect a given cable is, not surprisingly, crucially dependent upon the size (CSA), and hence CCC of the cable.

Fault protection is totally different and, as I said, nothing to do with protecting cables. Its sole purpose is to minimise the period of time for which exposed 'earthed' parts of equipment/appliances will remain 'live' (thus presenting a risk of electric shock) in the case of an L-E fault - a clue to it's purpose is that the requirements for fault protection exist in the chapter on "Protection Against Electric Shock" in the regs. I'm not sure where this talk about '5 seconds' came from. For nearly all circuits in an installation the requirement is for a maximum disconnection time of 0.4 seconds with TN installations or 0.2 seconds in TT installations (see Table 41.1 of regs).

As I keep saying, the fact that fault protection has nothing to do with protecting cables is clear from the fact that determination of the adequacy of fault protection does not even consider the size/CCC of the cable. Perhaps a hypothetical (and fairly extreme) example will help you....

.... one circuit comprising a short length of 1mm² cable and another circuit comprising a much longer length of 16mm² cable could both have an EFLI of, say, 0.8Ω. That means that under ('negligible impedance') fault conditions at the furthest part of the circuit, the fault current would, in both cases, be 230/0.8 = 287.5A at 230V, or 287.5 x 0.95 = 273.1A when one applies a 'Cmin correction' (effectively considering the case in which the supply voltage was only 218.5V). That means that fault protection of the circuit (i.e. disconnection times under fault conditions) would, in either case, be considered adequate if 5 x In of the (Type B) MCB was no greater than 273.1A. A B50 MCB (5 x 50 = 250A) would therefore provide adequate fault protection in either case.

It is apparent that, whilst, in the case of 16mm² cable, a B50 would give adequate ('overload') "protection to the cable" with most common installation methods (for which CCC >50A), it is equally apparent that the (overload) "protection of the cable" would be woefully inadequate if the circuit were wired in 1mm² cable.

The bottom line is therefore that the fact that a circuit has adequate fault protection does not in any way mean that the cable is necessarily adequately "protected" (against 'overload' over-current).

Does that help?

Kind Regards, John

 

[ban-all-sheds]

@eflmpudence - your comments above are not about something I wrote, but something @ban-all-sheds wrote several years ago. Since it was directly related I added a link, and copied the relevant post hoping @ban-all-sheds could clarify.

The clarification is that a circuit (or a branch of one) with an indeterminate number of socket outlets is not protected against overload currents by either downstream protective devices or the characteristics of the load.

 

[ban-all-sheds]

View attachment 164185

You are not alone. A past discovery of mine in a kitchen:

Also, I can't remember for certain about what accessories had or had not been where, but I'm pretty sure it was a case of "Zones? Wassat then?"

 

[wrightsonm]

I'm not sure where this talk about '5 seconds' came from.

Hi John, Thank you for such a detailed response. 5s was first mentioned by @ban-all-sheds earlier on. I believe it was in reference to the disconnection times in 411.3.2.2, 411.3.2.3
Table B6 of the OSG also refers to 0.1 to 5s disconnection times which I think are related to the MCB current curve graphs pg 370 Fig3A4
And 3.5.1 OSG is quite hard to read. Think there are a couple of words missing "0.4 s is required for final circuits not exceeding":

So as I understand it in a TN system:
32A ring final shall have a 0.4s disconnection time
40A cooker circuit shall have a 0.4s disconnection time (i think)
40A circuit to Garage CU shall have a 5s disconnection time

Or should the cooker be 5s?
Maximum disconnection times stated in Table 41.1 shall be applied to final circuits with a rated current not exceeding:
(i) 63 A with one or more socket-outlets, and
it is less than 63A but is fixed equipment + 1 socket. or the oven is considered fixed?
(ii) 32 A supplying only fixed connected current-using equipment.
oven and hob combined is > 32A.

Thank you for all of your time & explanations.
Regards
Mark

 

[JohnW2]

So as I understand it in a TN system:
32A ring final shall have a 0.4s disconnection time
40A cooker circuit shall have a 0.4s disconnection time (i think)
40A circuit to Garage CU shall have a 5s disconnection time

Yes, I suppose that a circuit feeding a garage CU is a distribution circuit, and therefore (per 411.3.2.3) only requires a 5s disconnection time (1s for TT).

Or should the cooker be 5s?

As you go on to imply, in terms of the regs, that is potentially debatable (particularly given that the 'after diversity' current of a cooker circuit will be well under 32A).

However, in practice, certainly for MCBs, this is all hypothetical and totally unimportant/irrelevant. If you look at the characteristic curves of MCBs (e.g. Fig 3A4 for Type B's), you will see that, at least as far as the eye can see, the curve is 'vertical' from 10s downwards. In other words, there is no discernible difference between the currents (hence EFLI) required to result in 10s, 5s, 0.4s, 0.2s or even 0.01sec (and probably even shorter) disconnection times.

Kind Regards, John

 

[EFLImpudence]

Table B6 of the OSG also refers to 0.1 to 5s disconnection times which I think are related to the MCB current curve graphs pg 370 Fig3A4

Yes B6 only refers to MCBs and RCBOs.
Don't forget the regulations also apply to commercial and industrial installations where fuses might still be used.

And 3.5.1 OSG is quite hard to read. Think there are a couple of words missing "0.4 s is required for final circuits not exceeding":

Agreed. BS7671 does include the words as you say below.
(An example where the OSG is not ideal. It frequently has mistakes and states 'cover all' figures and methods when, in fact other figures and methods can be acceptable.)

So as I understand it in a TN system:
32A ring final shall have a 0.4s disconnection time
40A cooker circuit shall have a 0.4s disconnection time (i think)
40A circuit to Garage CU shall have a 5s disconnection time

32A ring - a maximum of 0.4s.
40A cooker - it is permitted (not shall) to have up to 5s (unless a socket on the circuit, e.g. in the cooker switch).
40A circuit to garage - it is permitted (not shall) to have up to 5s

Or should the cooker be 5s?

As above - it is permitted but, don't forget, this is not possible with MCBs and RCBOs.

Maximum disconnection times stated in Table 41.1 shall be applied to final circuits with a rated current not exceeding:
(i) 63 A with one or more socket-outlets, and
it is less than 63A but is fixed equipment + 1 socket. or the oven is considered fixed?

Depends whether it is fixed or not.
Ovens usually are but stand alone cookers not.

I think it relates to whether or not one can pick up or grab hold of the item.

(ii) 32 A supplying only fixed connected current-using equipment.
oven and hob combined is > 32A.

No, It depends on the OPD - "circuits not exceeding 32A" - 32A circuit is not exceeding 32A.



Stick to BS7671. Hang the OSG in the bathroom where it belongs.

 

[JohnW2]

... 40A cooker - it is permitted (not shall) to have up to 5s (unless a socket on the circuit, e.g. in the cooker switch).

Are you sure? As I recently wrote, I think it is debatable given ....

Maximum disconnection times stated in Table 41.1 shall be applied to final circuits with a rated current not exceeding: .... (ii) 32 A supplying only fixed connected current-using equipment.

'Rated current' presumably relates to the In of the OPD, so how could one argue that a 32A cooker circuit (even without a socket) did not have to have the Table 41.1 (0.2s or 0.4s) disconnection times?

However, as I recently wrote, and you have effectively repeated, the reality is that, with an MCB (or RCBO), one can't really have a disconnection time which is <5s which isn't actually <0.2s.

Kind Regards, John

 

[EFLImpudence]

The example was a 40A circuit.

Isn't that what the regulations say when exceeding 32A and no sockets?

 

[JohnW2]

The example was a 40A circuit. Isn't that what the regulations say when exceeding 32A and no sockets?

Oh, fair enough - I was obviously thinking of a 'standard' (32A) cooker circuit, which appears to require a 0.4s/0.2s disconnection time (which, as we've both said, it will have, with an MCB or RCBO, if the disconnection time is <5s).

Kind Regards, John