Discuss Disconnection Time Requirements? in the Electrical Forum area at ElectriciansForums.net

Welcome to ElectriciansForums.net - The American Electrical Advice Forum
Head straight to the main forums to chat by click here:   American Electrical Advice Forum

Cookie

-
Reaction score
117
Why do circuits 32 amps and under require a 0.4 second disconnection time at 230 volts, but circuits over 32 amps are legally allowed 5 seconds? 115 volts for 5 seconds is enough to cause injury to the human body.
 
These disconnection times are for over-current and related Zs, not for RCD protection which is a lot faster (300ms/40ms at 30mA/150mA I think for personnel protection RCD). I'm not sure of the specific reasoning behind the choices of time, etc, but here is my guess:

In the UK most final circuits that would have anything plugged in to them would be 32A or less at the DB and they are at a much higher risk of fault/shock than fixed equipment (less likely to have an open earth, etc, as little or no movement).

Also for high current circuits you find the Zs needed to achieve 0.4s or faster disconnection can become onerous, maybe impossible depending on the supply Ze, and again they are less likely to be at fault so my assumption is the trade off in safety risk by allowing 5s was seen as acceptable.
 
Last edited:
Thats probably it as it makes the most sense.

I have a theory, (I could be wrong) that the a fault closer to the source will "pull down" the voltage more. Thus the actual voltage at the service (or transformer spades) will drop below 230 volts, and the 50/50 split of voltage drop between the live and earth conductor in TN-C-S will be less than 115 volts.
 
Another thing that used to be a more prominent requirement on the earth impedance, but is now covered under 415.2.2 for supplementary bonding, was to have no more than 50V AC fault potential at that fault current.

So at the current leading to 5s disconnection you would be exposed to no more than 50V, and if the current was higher (so bigger voltage) the disconnection time would be dropping very quickly. For example, if you look up the figures for a BS88-2 fuse rated at 80A your 5s time occurs at 400A, so if you had your 0.125 ohm bonding but 800A to get around 100V fault voltage then your disconnect time has dropped significantly below 0.4s.
[automerge]1589375712[/automerge]
But now it is just the overall disconnection time that has to be met (i.e. R2 earth is not specifically requested to be met, just Zs = Ze + R1 + R2) under 411.4.4
 
Last edited:
Also a lot of circuits over 32 amps tend to be distribution circuits to other boards, the 5 second disconnection helps with discrimination between devices so under fault the final circuit protective device is taken out rather than the whole submain.
 
Another thing that used to be a more prominent requirement on the earth impedance, but is now covered under 415.2.2 for supplementary bonding, was to have no more than 50V AC fault potential at that fault current.

Can you further elucidate on this? Did this mean that the system CPC between boards or the DNO PME was required to be a lower Z than the live phases?

If I have it right this would dictate a Z of at least half the live conductors...

I can still picture a remote 100 amp distribution board feed to an outbuilding, caravan or farm coop where the source is very strong but the run such that 4 seconds would be realistic. The caravan and barn are the most dangerous in that you have remote earth everywhere around the metal object.
 
Can you further elucidate on this? Did this mean that the system CPC between boards or the DNO PME was required to be a lower Z than the live phases?
The requirement is for supplementary bonding of conductive parts in an installation, so supply/DB earth to steelwork, pipes, etc. It is not a requirement on the total resistance to "true Earth".

So it won't guarantee you don't exceed 50V AC absolute, only that you would be unable to touch parts within an installation that had more than 50V between them. Essentially it is establishing the degree of equipotential in an equipotential region!

I can still picture a remote 100 amp distribution board feed to an outbuilding, caravan or farm coop where the source is very strong but the run such that 4 seconds would be realistic. The caravan and barn are the most dangerous in that you have remote earth everywhere around the metal object.
There are additional requirements for both of those areas!

For the caravan case there are restrictions on the earth type so you don't try and extend the equipotential region of a TN-S-C supply outdoors to a site with a high risk of exposed PME-earth to true-Earth being dangerous in the event of a supply fault. In addition, the caravan supplies must be all-pole RCD protected, so your disconnect times are going to be sub-300ms.

For more info search for "Electrical installations in caravan/camping parks, caravans and motor caravans" and you should find an IET wiring matters article on it.

For agriculture again there are various extra rules, off the top of my head I don't remember all, but essentially they also require RCD protection against fault/fire risk, and the "safe" limit is 25V AC (lower than the 50V AC used elsewhere). I think there are more earthing/bonding requirements as a fault has a high risk to people and livestock (a lot of farm animals are more sensitive than humans, and don't have the luxury of largely-insulating shoes).
 
Last edited:
Why do circuits 32 amps and under require a 0.4 second disconnection time at 230 volts, but circuits over 32 amps are legally allowed 5 seconds?
Hi - I think the short disconnection times are intended for (most) final circuits with current using equipment or outlets. The longer disconnection times are permitted for distribution circuits as per @Strima :) .
See wording of Reg 411.3.2
 
Sounds reasonable for caravans and agricultural buildings. In regards to this:

The requirement is for supplementary bonding of conductive parts in an installation, so supply/DB earth to steelwork, pipes, etc. It is not a requirement on the total resistance to "true Earth". So it won't guarantee you don't exceed 50V AC absolute, only that you would be unable to touch parts within an installation that had more than 50V between them. Essentially it is establishing the degree of equipotential in an equipotential region!

For PE or PME voltage drop between the DNO transformer and MET, bonding all extraneous conductive parts within a building will reduce voltage potentials.

However, there is still the issue of voltage drop between the main distribution board and sub-mains, especially for strong sources (11kv in the basement) with long SWA runs in large buildings. It is possible to legally exceed 50 volts, correct?
 
Hi - I think the short disconnection times are intended for (most) final circuits with current using equipment or outlets. The longer disconnection times are permitted for distribution circuits as per @Strima :) .
See wording of Reg 411.3.2


What is driving the 32 amp limit? In the Indian Electrical Code for example 0.4 seconds is required for all final circuits, even those over 32 amps. An excerpt from India's code:

View: https://Upload the image directly to the thread.com/BFUNhZE
 
For PE or PME voltage drop between the DNO transformer and MET, bonding all extraneous conductive parts within a building will reduce voltage potentials.
That is the idea. Also with TN-C-S (i.e. the PME case) the bonding has to be at least 10mm copper equivalent (10mm Cu, 16mm Al, 80mm Fe, etc) as it could carry high currents under fault conditions (tens or hundreds of amps, depending on how low the Ra of the extraneous conductive parts are) so you have a fire risk if thin wire is used.

However, there is still the issue of voltage drop between the main distribution board and sub-mains, especially for strong sources (11kv in the basement) with long SWA runs in large buildings. It is possible to legally exceed 50 volts, correct?
The 50V limit is planned for under the fault-disconnection currents, and it is for an "installation" which of course varies depending on the context. Could be a single floor flat, up to a large factory building, etc.

If you have a lot of current (e.g. your own HV/LV transformer) you would be looking at conductors to match. For large SWA the armour is not always sufficient so you might have an internal copper core for the CPC, or an additional earth cable run along side the SWA.

I have no experience of the DNO side (or private equivalent) from transformer to DBs, but maybe others on here can comment on any particular requirements as that is moving outside of the remit of the usual wiring regulations in to those covering the supply industry.
[automerge]1589388546[/automerge]
What is driving the 32 amp limit? In the Indian Electrical Code for example 0.4 seconds is required for all final circuits, even those over 32 amps. An excerpt from India's code:
I guess traditionally most ring final circuits feeding our 13A sockets were fed from 30A fuse / 32A MCB. But checking the 18th regulations it now has 411.3.2.2 stating the 0.4s TN case applies to circuits not exceeding a rated current of:

(i) 63A with one or more socket-outlets
(ii) 32A supplying only fixed connected current-using equipment

Then 411.3.2.3 has the allowed time of 5s for distribution circuits (i.e. sub-main) and circuits not covered by 411.3.2.2 (above).
[automerge]1589389296[/automerge]
If you have a lot of current (e.g. your own HV/LV transformer) you would be looking at conductors to match. For large SWA the armour is not always sufficient so you might have an internal copper core for the CPC, or an additional earth cable run along side the SWA.
Also remember you probably would not run the a single feed up a building, probably you would spilt it at a panel near the transformer and each area in the building would get a smaller feed via a 100A/200A fuse/MCCB or whatever at that panel. So the disconnection current for that area would be based on its feed fuse, not an 1600A ACB or whatever on the transformer secondary.
 
Last edited:
All true. But I'm still thinking of a piece of equipment (like a pizza oven with metal handles) connected to a circuit in excess of 32 amps. Also the voltage drop along a feed between a 1,600 ACB and and a 200amp sub-board with the fault inside the sub-board.
 
What is driving the 32 amp limit? In the Indian Electrical Code for example 0.4 seconds is required for all final circuits, even those over 32 amps.
Hi - I don’t know why the UK has put the 63A and 32A limit in place for the shorter disconnect times in final circuits. My guess it’s it’s a matter of practicality. Keep in mind this is the minimum standard and we can always exceed the requirements of the reg and provide shorter disconnection times.
 
All true. But I'm still thinking of a piece of equipment (like a pizza oven with metal handles) connected to a circuit in excess of 32 amps.
But you would only see the local potential differences, and that would be kept reasonably low. Yes, you might be very unlucky to see more than 50V but it would be for less than 5s and you would have to be in contact with the oven and related ground potential at the same time. Also remember the 5s limit is not a goal, it is the upper case, and when designing this you are using worst-case assumptions on supply voltage, etc.

Also the voltage drop along a feed between a 1,600 ACB and and a 200amp sub-board with the fault inside the sub-board.
Well the sub-board or bus-bar chamber in that case would be at least 1600A internally, only the sub-main leaving it would be 200A rated! That would be the sort of situation where fault currents in the 50kA region would be likely, and a good case for HRC fuses to limit the peak fault current stress.

Maybe it has changed, but I think HRC fuses were always better than MCCB at limiting fault current. So if you don't need a particularily fancy trip profile, remote control, etc, and don't expect faults to be frequent, then having a set of fused-switches off a bus bar chamber might look old fashioned, but it is reliable and effective at that one job that matters.
[automerge]1589392869[/automerge]
Also remember if your working volt drop is, say 5% of 230V = 11.5V max for both phase & neutral (so 5.75V on each, assuming identical conductors), and you are looking to contain a fault to 50V then your disconnect current would be 8.7 times the working current. In the 1600A ACB case that is just under 14kA!

Now I don't have an ACB curve to hand, but the 1250A BS88-2 fuse has a 5s trip at 10950A so 8.76 times rated.

Basically your impedance to operate scales down as current goes up, so your fault voltage control is scaled in practically the same way automatically by that.
[automerge]1589393311[/automerge]
Realistically the volt drop from transformer to distribution panel would be a lot lower, probably under 1%, so the fault clearing will be very fast (and spectacular).
 
Last edited:
Makes sense. However- can anyone confirm or deny my theory that when short circuit occurs, the voltage at the transformer or generator is pulled down- actually resulting in a voltage lower than 230 volts at the spades? Thus, if for example, the output is 160 volts during the fault, there would only be only 80 volts to remote earth? This has been making me think.

A 32 amp circuit would make the source look strong, and as such the source can be considered infinite- providing all the current needed during a fault without voltage drop. However a fault on a 400 amp circuit would make the source look weak- dropping in voltage as it supplies the current.
 
What happens to the phase voltage would depend very much on the relative impedances of the supply transformer and wires, etc:

If you have a hard three-phase fault (L1-L2-L3) and a symmetric system you get essentially 0V = true earth (at hypothetical fault point centre), though in reality you would see some volts on the conductors if imbalance, or tens/hundred volts you are sustaining an arc-fault (even if the centre of the arc is hypothetically at ground potential).

If you are at the end of a long supply cable with similar L & N/E conductors then your phase impedance (Zxfmr + Zdist + R1 of installation) and earth impedance (Zpen + R2) are very similar so under a hard single phase fault (L-N or L-E) you would get roughly Uo/2 w.r.t. true earth at the fault point.

Close to the transformer, and assuming the fault was really hard (not an arc, a "bolted short"), then you would probably see tens of volts as the transformer impedance might dominate the short run of L & N/E cables to the distribution panel. But it is going to be a serious amount of current flowing so the volts that happen to be there are the least of your worries if you have the misfortune to be present...

However, if you have a system with a higher earth return impedance, worst case a TT earth rod, then you would see practically the whole supply as the VD on the phase cable would be a fraction of that on the circuit to rod and back via the Earth.
 
What happens to the phase voltage would depend very much on the relative impedances of the supply transformer and wires, etc:

If you have a hard three-phase fault (L1-L2-L3) and a symmetric system you get essentially 0V = true earth (at hypothetical fault point centre), though in reality you would see some volts on the conductors if imbalance, or tens/hundred volts you are sustaining an arc-fault (even if the centre of the arc is hypothetically at ground potential).

Correct, but an arc is not a linear sine wave current draw. It will produce some neutral/ground current.

If you are at the end of a long supply cable with similar L & N/E conductors then your phase impedance (Zxfmr + Zdist + R1 of installation) and earth impedance (Zpen + R2) are very similar so under a hard single phase fault (L-N or L-E) you would get roughly Uo/2 w.r.t. true earth at the fault point.

Agreed.

Close to the transformer, and assuming the fault was really hard (not an arc, a "bolted short"), then you would probably see tens of volts as the transformer impedance might dominate the short run of L & N/E cables to the distribution panel. But it is going to be a serious amount of current flowing so the volts that happen to be there are the least of your worries if you have the misfortune to be present...

Good point, agreed.

However, if you have a system with a higher earth return impedance, worst case a TT earth rod, then you would see practically the whole supply as the VD on the phase cable would be a fraction of that on the circuit to rod and back via the Earth.

True- hence the shorter times required for TT?

The thing about TT, is that is behaves like a high impedance earthed (IT) system from my understanding. Everything in the building that is bonded (has a CPC or bond wire) will be near zero relative to each other and the presence of limited fault current will actually reduce the voltage drop across the CPC while the fault is happening. The only real danger is being say outside, holding a portable device, where one is referenced to remote earth. The voltage will, IMO, be small between a faulted BUNN coffee maker and bonded plumbing while the fault is happening.
 
Well I was told, having asked that question that over 32a is for high resistive loads. on fixed equipment i.e. ovens or pizza dish washers where maybe the thermostat fails and there is no temperature control. We do have rcd for additional protection where there is a fault to earth. I thought at the time that was a satisfactory explanation. But looking at the above maybe there is more to it. I like the idea of selectivity being better served on sub mains. Also it is a more robust circuit that can take more abuse before it fails, but not so much that damage could occur to property person or livestock. A dead short would no doubt achieve a much faster disconnection time.
 
Correct, but an arc is not a linear sine wave current draw. It will produce some neutral/ground current.
Very non-linear for voltage, less so for current waveform.

The thing about TT, is that is behaves like a high impedance earthed (IT) system from my understanding. Everything in the building that is bonded (has a CPC or bond wire) will be near zero relative to each other and the presence of limited fault current will actually
In all cases bonding is used to maintain equipotential regions, and so no matter what you should be safe if CPC integrity is maintained.

But I would say no, TT is much lower impedance than an IT system.

Really TT is more like TN in the sense they both assume a fault to earth/CPC will result in a protective device tripping. The problem is with TT return impedances typically in the tens of ohms or higher region it would only trip an OCPD rated at a couple of amps max, so in that sense TT only became a safe and reliable method with the widespread use of RCD (the VOELCB were always dodgy).

The IT approach to faults is very different and for different reasons. There the earthing impedance is used to control leakage (capacitive, etc) and to make sure that in the event of a single fault nothing much happens. So here you don't expect the earth path to trip anything, only to generate a warning that you are now under fault conditions and that a second fault is going to be bad.

The need for complex monitoring, and the tendency for unskilled folk to do nothing on that until it actually goes bang! are the reasons I think it is prohibited for normal installations, and is only used in special cases like ships and medical environments where you have to be tolerant of a single fault (because there are high risks from power failing) and you have skilled staff on hand at all times to respond to fault warnings.
 

Reply to Disconnection Time Requirements? in the Electrical Forum area at ElectriciansForums.net

OFFICIAL SPONSORS

Electrical Goods - Electrical Tools - Brand Names Electrician Courses Green Electrical Goods PCB Way Electrical Goods - Electrical Tools - Brand Names Pushfit Wire Connectors Electric Underfloor Heating Electrician Courses
These Official Forum Sponsors May Provide Discounts to Regular Forum Members - If you would like to sponsor us then CLICK HERE and post a thread with who you are, and we'll send you some stats etc

Electrical Forum

Welcome to the Electrical Forum at ElectriciansForums.net. The friendliest electrical forum online. General electrical questions and answers can be found in the electrical forum.
This website was designed, optimised and is hosted by Untold Media. Operating under the name Untold Media since 2001.
Back
Top
AdBlock Detected

We get it, advertisements are annoying!

Sure, ad-blocking software does a great job at blocking ads, but it also blocks useful features of our website. For the best site experience please disable your AdBlocker.

I've Disabled AdBlock