My first EICR and I'm confused! I'm industrial this is domestic

[Inteificio]

Hi,

I'm an industrial spark and taking my 2391 next week (wish me luck, I need it).

I decided to get a bit of practice and started an inspection on my own house.

I got as far as, b32 socket (ring?!) breaker looks slightly browned, better check it.

3 cables in the breaker two identical and one that appears to be a spur to an adjacent socket.

Q1) Let's start with the adjacent socket as that one should be easy:

It's a double socket.

There is no way a load can be put on that will stress the 2.5mm without blowing the bs1361s...however, the 2.5 is undersized to the b32.

clearly dodgy, but just curious how some experienced guys would treat this.


Q2) Ok now for the confusing bit.

The 'ring' is only a ring on the N. Live and E are open.
All sockets pass inspection (well small arcing on one N, which is fixed)
IR passes (on this circuit).

The board is split RCD so a borrowed N would take out the RCD.

All sockets have an in/out feed apart from one.

r1/r2 unsurprisingly no reading.

r1/rn would have been a bit more interesting, but haven't done yet.


Does anyone have any advice, this seems a strange one?

Thanks

 

[westward10]

The fact you have open end to ends on the lines and cpcs is an issue which would attract in most people's eyes a Code 2. The fact you have continuity across the neutral conductors suggests it was a ring final circuit.

 

[Julie.]

Hi,

I'm an industrial spark and taking my 2391 next week (wish me luck, I need it).

I decided to get a bit of practice and started an inspection on my own house.

I got as far as, b32 socket (ring?!) breaker looks slightly browned, better check it.

3 cables in the breaker two identical and one that appears to be a spur to an adjacent socket.

Q1) Let's start with the adjacent socket as that one should be easy:

It's a double socket.

There is no way a load can be put on that will stress the 2.5mm without blowing the bs1361s...however, the 2.5 is undersized to the b32.

clearly dodgy, but just curious how some experienced guys would treat this.


Q2) Ok now for the confusing bit.

The 'ring' is only a ring on the N. Live and E are open.
All sockets pass inspection (well small arcing on one N, which is fixed)
IR passes (on this circuit).

The board is split RCD so a borrowed N would take out the RCD.

All sockets have an in/out feed apart from one.

r1/r2 unsurprisingly no reading.

r1/rn would have been a bit more interesting, but haven't done yet.


Does anyone have any advice, this seems a strange one?

Thanks

Why is it dodgy to not have overload protection on a cable (spur) which cannot be overloaded?

433.3.1 would clearly be applicable

If you have a break in the rfc then it would automatically require a C2.

It needs fixing, either split into two at the break in which case both would need 20A mcb, or fix the fault.

 

[Julie.]

The fact you have open end to ends on the lines and cpcs is an issue which would attract in most people's eyes a Code 2. The fact you have continuity across the neutral conductors suggests it was a ring final circuit.

I think it's the line which is the ring, both n & CPC are broken.

 

[Julie.]

To the op, have you looked at gn3 - do you have a copy? You need to be familiar with this (and the regs) before you consider periodic inspections.

This explains how to test rfcs

 

[westward10]

I think it's the line which is the ring, both n & CPC are broken.

He says Live and E are broken?

 

[Julie.]

He says Live and E are broken?

Sorry, yes he does, but I got confused by him thinking that the n was made via borrowed neutral, albeit discounted.

 

[Inteificio]

wow, thanks for the fast responses!

I've been doing my best to memorise GN3.

It's doing the ring tests that I've come across the faults.

Why is it dodgy to not have overload protection on a cable (spur) which cannot be overloaded?

433.3.1 would clearly be applicable

If you have a break in the rfc then it would automatically require a C2.

It needs fixing, either split into two at the break in which case both would need 20A mcb, or fix the fault.

That's why I'm asking, I'm checking what the skilled inspectors think.

If there was a short on the back of that socket the cable would potentially (not done the calc) overheat before the b32 trips. So there is a small chance of an unprotected fault.

433.3.1 nails it, thanks.

The fact you have open end to ends on the lines and cpcs is an issue which would attract in most people's eyes a Code 2. The fact you have continuity across the neutral conductors suggests it was a ring final circuit.

I would have thought C1, can you explain why C2?

My thinking is that the circuit can overheat in non-fault conditions hence C1. Functionally I've got a 2.5mm spur that can happily maintain 32A without a trip.


So the simple solution is a break in the L/E, it just seems such an unlikely fault to me. Is there anything else that I could be overlooking, or is it time to start ripping up floorboards before my house burns down?

Downstairs
spur
Zs - 0.25 - Zn - 0.19

Near socket
Zs - 0.54 - Zn - 0.31

Far socket
Zs - 0.58 - Zn - 0.35

Upstairs
Landing
Zs - 0.38 - Zn - 0.26

Vlad's bedroom - Spur
Zs - 0.43 - Zn - 0.31

Mordecai's bedroom
near 0.46 - 0.3
far 0.39 - 0.26

My room
0.33 0.29

I'll check for patterns shortly, but pizza has arrived.

Thanks for the help, understanding how other people tackle problems is good training.

 

[westward10]

We cannot assume the breaks have occurred in the same place but you will have to carry out contuity tests. Have you tried an insulation resistance test across the open ends to see if there is a tenuous connection.

 

[Inteificio]

We cannot assume the breaks have occurred in the same place but you will have to carry out contuity tests. Have you tried an insulation resistance test across the open ends to see if there is a tenuous connection.

I have not, good idea, thanks

 

[Julie.]

wow, thanks for the fast responses!

I've been doing my best to memorise GN3.

It's doing the ring tests that I've come across the faults.


That's why I'm asking, I'm checking what the skilled inspectors think.

If there was a short on the back of that socket the cable would potentially (not done the calc) overheat before the b32 trips. So there is a small chance of an unprotected fault.

433.3.1 nails it, thanks.

I would have thought C1, can you explain why C2?

My thinking is that the circuit can overheat in non-fault conditions hence C1. Functionally I've got a 2.5mm spur that can happily maintain 32A without a trip.


So the simple solution is a break in the L/E, it just seems such an unlikely fault to me. Is there anything else that I could be overlooking, or is it time to start ripping up floorboards before my house burns down?

Downstairs
spur
Zs - 0.25 - Zn - 0.19

Near socket
Zs - 0.54 - Zn - 0.31

Far socket
Zs - 0.58 - Zn - 0.35

Upstairs
Landing
Zs - 0.38 - Zn - 0.26

Vlad's bedroom - Spur
Zs - 0.43 - Zn - 0.31

Mordecai's bedroom
near 0.46 - 0.3
far 0.39 - 0.26

My room
0.33 0.29

I'll check for patterns shortly, but pizza has arrived.

Thanks for the help, understanding how other people tackle problems is good training.

If there is a short then it would be a fault, normally this is assumed to be zero resistance and therefore your Zs check would show it would disconnect within .4s - if it is a high resistance arc, then yes it could pass just enough not to trip due to fault, but more than the cable rating (remembering the arc itself is more of a likely fire issue than being able to overheat the cable over several hours).

Why C1 - C1 means it is an immediate danger (for example live contacts exposed)
Why C2 - C2 means that there is a danger if something else occurs to cause the danger.

In this case it doesn't present an immediate danger, there has to be a previous problem (overload etc).

As to where is the fault, you can usually narrow it down without ripping up too much, split the rfc at the cu, short one end l-n then go round the sockets looking for l-n continuity (just use the standard plug from the mft - no need to remove sockets), then repeat for the other end, this generally narrows it down to between two sockets (then it's a bad connection in either socket, or the cable between).

Then repeat for the cpc

There is no guarantee that you have just the one point of fail - you may find a couple of breaks in the same conductor and/or different points of break between conductors.

 

[Inteificio]

If there is a short then it would be a fault, normally this is assumed to be zero resistance and therefore your Zs check would show it would disconnect within .4s - if it is a high resistance arc, then yes it could pass just enough not to trip due to fault, but more than the cable rating (remembering the arc itself is more of a likely fire issue than being able to overheat the cable over several hours).

Why C1 - C1 means it is an immediate danger (for example live contacts exposed)
Why C2 - C2 means that there is a danger if something else occurs to cause the danger.

In this case it doesn't present an immediate danger, there has to be a previous problem (overload etc).

As to where is the fault, you can usually narrow it down without ripping up too much, split the rfc at the cu, short one end l-n then go round the sockets looking for l-n continuity (just use the standard plug from the mft - no need to remove sockets), then repeat for the other end, this generally narrows it down to between two sockets (then it's a bad connection in either socket, or the cable between).

Then repeat for the cpc

There is no guarantee that you have just the one point of fail - you may find a couple of breaks in the same conductor and/or different points of break between conductors.

Good to know, thanks.

 

[Inteificio]

urgh, just thought about this more.

OK I now get why it's not a C1 for the reason I gave (someone would need to plug something in to make it overload).

However, you mentioned arcing and fire...

Is that not an immediate problem?

Or is the arcing a cause that could create a future problem (fire) which is why C2?

 

[Julie.]

urgh, just thought about this more.

OK I now get why it's not a C1 for the reason I gave (someone would need to plug something in to make it overload).

However, you mentioned arcing and fire...

Is that not an immediate problem?

Or is the arcing a cause that could create a future problem (fire) which is why C2?

You mentioned a short at the back of a socket, how could you get that unless it was either a full short , or an arc causing just the correct amount of current?

This fault could occur anywhere.

The likelihood of an arc occurring with just the right amount of current to not cause a fault protective device to trip, but cause a fire is the same irrespective of what you have found to date.

If there is a reason to believe a fault such as this would occur then yes that particular issue may draw a code.

 

[littlespark]

The spur wouldn’t overload because as a double socket, the most that can be plugged in is 2 x 13A… 2.5 cable can take 26A, (although the double socket itself might struggle!)
If a dead short, then the 32 breaker would trip.

Looking for the fault could be the case of taking off every socket and testing every cable between. There may be hidden joints you yourself are unaware of, even if it is your own house.

And… unfortunate positioning of dodgy advert on previous message……

 

[Julie.]

The spur wouldn’t overload because as a double socket, the most that can be plugged in is 2 x 13A… 2.5 cable can take 26A, (although the double socket itself might struggle!)
If a dead short, then the 32 breaker would trip.

Looking for the fault could be the case of taking off every socket and testing every cable between. There may be hidden joints you yourself are unaware of, even if it is your own house.

And… unfortunate positioning of dodgy advert on previous message……

View attachment 93674

Wtf...

Great advert yet again

 

[pc1966]

If you see an open RFC conductor then you know there is a fault. Unless you know exactly what it is you have to assume the worst and that potentially the ring's cable could be overloaded (if open at one end so really a 2.5mm radial on 32A MCB) and/or there is a loose terminal at one or more location (back of socket, junction box) that could result in arcing/overheating. So in this case it is C2 as not the extreme danger that exposed live parts or reversed polarity causes, but it is far worse than C3 "improvement recommended".

That is not unique to the RFC, the same risk applies to a radial. But with a radial an open on a live conductor results in one or more socket not working at all, so it is more likely to be fixed. On open CPC on a radial of course is likely to be missed outside of testing, where as the RFC ought to have some redundancy in providing a path for fault currents to hopefully clear safely.

The issue of a spur on the RFC is more complicated, as 2 * 13A could just overload 2.5mm though it is unlikely to do so in practice and the test current for a double socket is 20A which is usually OK for 2.5mm and in most cases a more realistic maximum.

As said above, test the RFC systematically by some means such as linking L/E at one end of the ring at the CU and checking socket by socket for L-E continuity/resistance until you see a fault (open or suspiciously high R), then remove that socket and see what you find. If good, go back check the last good one as it might be open on the cable going 'out' in the direction of test.

Of course, in most houses you will have no idea of just how the ring was routed, so it is a bit of a guess, but you can generally get some sense after a couple of sockets in order are found and from the general practically of how it could be done.

 

[Inteificio]

Well I've had a good root around.

The zs/rn readings are so low that they are useless for analysis.

Excluding the spur at the board the circuit is:

Downstairs 2 ring sockets on leg 1.
Upstairs 2 ring sockets and a spur on leg 1.
Upstairs 1 ring and 1 spur on leg 2.

Cpc and live split at same point.

I can't get any further without ripping up carpets and floor boards!

Is it safe enough to just swap to a b20 and sign it off as a split ring?

Or should I fit a new fangled arc detection doohikey (yeah I know nothing about them)?

Or do I have to start pulling floor boards as the risk is too high?

Industrial is so much easier, no nice carpets to worry about =)

Thanks

 

[nicebutdim]

Complile the report and pass remedial work on to another spark

 

[brianmoooore]

Have you removed the sockets either side of the fault? Chances are that it's wires pulled out of one of those sockets. Far more common than an actual cable fault.
Even if it were a cable fault, and you decide to split the circuit into two radials, you would still have to disconnect the faulty leg completely at the two sockets at either end.

 

[Inteificio]

All the sockets are fine.
When separating the legs I've identified which sockets are on which side.

Fortunately as only one ring socket is on one leg I'm pretty certain the fault lies between that socket and another. Finding the 'another' will be a faf.

My next step is to remove all the ring socket covers and separate the cables. A bit of long leading the N will help find the dodgy leg.

I can then remove that one cable from each end leaving (hopefully) two spurs.

What a faf.

I hope the actual exam isn't this bad or I'm going to run out of time massively!

 

[brianmoooore]

If the fault is near to one end of the RFC, you are going to end up with a spur that loops back on itself, so are going to end up with a spur that uses much more cable than one that was designed as a spur from the start.
You will need to check that loop impedance and voltage drop are all acceptable, and access the likely maximum load of the circuit, to ensure it doesn't exceed the rating of the smaller MCB.

 

[UNG]

All the sockets are fine.
When separating the legs I've identified which sockets are on which side.

Fortunately as only one ring socket is on one leg I'm pretty certain the fault lies between that socket and another. Finding the 'another' will be a faf.

My next step is to remove all the ring socket covers and separate the cables. A bit of long leading the N will help find the dodgy leg.

I can then remove that one cable from each end leaving (hopefully) two spurs.

What a faf.

I hope the actual exam isn't this bad or I'm going to run out of time massively!

Rather confused by the fact that you have identified what sockets are on each side of the fault but then you say you haven't and that finding the other socket on the ring will be a faf

I would suggest buying a cheap continuity tester https://cpc.farnell.com/tenma/ten01049/continuity-tester/dp/IN05723 rather than spending a lot of time removing all the sockets

With the continuity tester you can work along one leg of the ring quite quickly and identify where the break is and then do the same on the other leg which should bring you to a point where you can then diagnose the cause of the fault and repair the ring

 

[pc1966]

The zs/rn readings are so low that they are useless for analysis.

What are you using to measure things?

Generally speaking Zs measurements are not that precise, not when you get down to under an ohm and more so if "no trip" due to RCD action so testing at very small L-E currents, so you are unlikely to get far finding a RFC break on that alone.

However, DC low ohm measurements using a typical MFT or similar usually are good enough but you have to allow for dirty switch contacts and socket pins so sometimes you have to insert/remove the test plug adaptor (or lead) a couple of times and operated the switch a could of times to get a low stable reading.

Even then you are likely to only reliably see differences in cable distance of several meters as, for example 2.5m T&E has R1+R2 of about 0.02 ohm/meter and (I guess) most low ohm meters likely to be at hand resolve around 0.01 ohm.

 

[pc1966]

I would suggest buying a cheap continuity tester https://cpc.farnell.com/tenma/ten01049/continuity-tester/dp/IN05723 rather than spending a lot of time removing all the sockets

I don't think that is CAT rated, so I would rather not use it on anything that might be live!

Maybe a cheap voltage tester that has a continuity buzzer?

 

[UNG]

I don't think that is CAT rated, so I would rather not use it on anything that might be live!

Maybe a cheap voltage tester that has a continuity buzzer?

So is he is going to disconnect all the sockets while they are live, I think that would be a no

Given the OP is an industrial electrician I would like to think he would be able to use a simple continuity tester to locate a simple ring circuit fault

 

[Inteificio]

I'm using a Megger mft1721.
I've also got my fancy multimeter for precise testing.

You're right, the Zs is useless for this, I'll do a dc test as that may have more accuracy. Good tip, thanks.

What I'm hoping is that the failure is directly between two ring sockets. All I need to do is identify the cable between and disconnect it at both ends.
If that is the case then I know which socket is on one side. I don't know where the next one in the supposed ring is (there are 3 probables).
Unfortunately I don't know where the spurs cut in, which adds a wild card.



Hopefully the accuracy of the r1rn dc test will give clues.

As for splitting the ring, my whole house was using 14A in total. This circuit is just skirting board sockets. Unless I suddenly start loving electric room heaters the load will be limited.

Now the question is do I fit the b16 from my spares, or go buy a b20


P.s. That's a good tip about turning the sockets off before splitting them too. I wish I'd known that last night as the burns on my hands are really smarting this morning

 

[UNG]

I'm using a Megger mft1721.
I've also got my fancy multimeter for precise testing.

You're right, the Zs is useless for this, I'll do a dc test as that may have more accuracy. Good tip, thanks.

What I'm hoping is that the failure is directly between two ring sockets. All I need to do is identify the cable between and disconnect it at both ends.
If that is the case then I know which socket is on one side. I don't know where the next one in the supposed ring is (there are 3 probables).
Unfortunately I don't know where the spurs cut in, which adds a wild card.



Hopefully the accuracy of the r1rn dc test will give clues.

As for splitting the ring, my whole house was using 14A in total. This circuit is just skirting board sockets. Unless I suddenly start loving electric room heaters the load will be limited.

Now the question is do I fit the b16 from my spares, or go buy a b20


P.s. That's a good tip about turning the sockets off before splitting them too. I wish I'd known that last night as the burns on my hands are really smarting this morning

While splitting the ring is an easy fix and while it is very rare that a cable is at fault I think actually locating / repairing the fault would be more beneficial from a learning point of view for you

 

[pc1966]

So is he is going to disconnect all the sockets while they are live, I think that would be a no

Given the OP is an industrial electrician I would like to think he would be able to use a simple continuity tester to locate a simple ring circuit fault

Here, no it should be fine to use that kit.

It is more of a general dislike on my behalf of having test equipment that is not mains-fault rated in use where mains is about. Sooner or later we all make a mistake and probe the wrong circuit (or discover they are mislabelled somewhere) that is live, and then having something that does not explode on you is nice.

 

[Inteificio]

Here, no it should be fine to use that kit.

It is more of a general dislike on my behalf of having test equipment that is not mains-fault rated in use where mains is about. Sooner or later we all make a mistake and probe the wrong circuit (or discover they are mislabelled somewhere) that is live, and then having something that does not explode on you is nice.

I think we all knew what you meant.

Exploding kit aka audiovisual alert.

 

[Inteificio]

OK we have progress, but more questions.

So please humour the newbie =)

The dc measurement is more accurate so I pinned the end of the other ring to the Pc socket.
I disconnected the wires at the back of the relevant sockets. Long lead the n to confirm the link. Then r1r2 the other legs to check they were fine.

Sorted, b16 is in! I'll get the b20 if I nuisance trip. (can you guess which county I live in).

So eicr question.

What would you code the discolouration of the (original) mcb. I'd go c3 as the problem is identified and repaired.

I've also run the calcs on the single spur originally for cpc.

Now the earth fault current is 800a, which is too much for a 1.5mm mcb. However as there is rcd protection am I allowed to use the 40ms or actual trip time to calculate cable overheat?

However a l-n short will give a 1.2kA fault current. The circuit will trip in 0.1s

Unless I've got the formula wrong (or totally misunderstood) minimum CSA needs to be over 3.2....

S>{root(i²t)} /k

Have I got this wrong?

Thanks

 

[Julie.]

OK we have progress, but more questions.

So please humour the newbie =)

The dc measurement is more accurate so I pinned the end of the other ring to the Pc socket.
I disconnected the wires at the back of the relevant sockets. Long lead the n to confirm the link. Then r1r2 the other legs to check they were fine.

Sorted, b16 is in! I'll get the b20 if I nuisance trip. (can you guess which county I live in).

So eicr question.

What would you code the discolouration of the (original) mcb. I'd go c3 as the problem is identified and repaired.

I've also run the calcs on the single spur originally for cpc.

Now the earth fault current is 800a, which is too much for a 1.5mm mcb. However as there is rcd protection am I allowed to use the 40ms or actual trip time to calculate cable overheat?

However a l-n short will give a 1.2kA fault current. The circuit will trip in 0.1s

Unless I've got the formula wrong (or totally misunderstood) minimum CSA needs to be over 3.2....

S>{root(i²t)} /k

Have I got this wrong?

Thanks

What do you mean by 1.5mm^2 mcb?

Do you have a 1.5mm^2 cable in a 20A mcb, and a perspective fault current of 800A?

If so, the trip time will be in the current limiting area of the characteristic less than 0.01 sec, this should be well within the adiabatic range of 1.5mm^2

You should not code something that doesn't exist, therefore you cannot put C3 against a mcb which is no longer there. Just put a note that the damaged mcb and ring has been repaired.

 

[brianmoooore]

Are you saying that the original 32A MCB is showing signs of overheating? Why? If it is because the load of the RFC exceeds 32A for long enough periods to cause problems, then reducing the MCB to 16A or 20A is only going to cause more problems.

 

[Inteificio]

I'm not exactly sure why the MCB looked browned. My loose guess was that as the ring was split, but still with a 32A breaker on it, the wires could be getting very hot. e.g taken to the limit, one leg has 30A of kit on it, but is functionally a 20A, then that wire would get hot without tripping.

I've replaced overheating, with potentially nuisance tripping. I'm more than happy with that exchange.

Am I understanding what you are saying?

 

[brianmoooore]

You'll be replacing a C2 under one reg. with a C2 under another.

 

[Inteificio]

What do you mean by 1.5mm^2 mcb?

Do you have a 1.5mm^2 cable in a 20A mcb, and a perspective fault current of 800A?

If so, the trip time will be in the current limiting area of the characteristic less than 0.01 sec, this should be well within the adiabatic range of 1.5mm^2

You should not code something that doesn't exist, therefore you cannot put C3 against a mcb which is no longer there. Just put a note that the damaged mcb and ring has been repaired.

Yeah re-reading what I wrote is as clear as mud!

What I meant with the "1.5mm MCB":

The offending spur is 2.5 T+E on a B20, backed by an RCD.

My understanding is an MCB with sufficient fault current will disconnect in 0.1s. regs p370 (though you suggest 0.01? why?).

With my understanding, an 800A earth fault current will overheat a 1.5mm CPC in 0.1s.

However, the RCD would trip much faster than the 0.1s (can I use the actual of 17ms or the 40ms from minimum)?

So for the CPC overheating, I'm OK (please confirm).

What I'm wondering about is the potential for a short circuit at that socket.

The short circuit would be 2.5/2.5 and 1.2kA.

That would trip in the 0.1s and still cook.



The C3 was more of a theoretical. If I had repaired the ring and left the original breaker, as we have a probable cause for the browning that had been repaired, would it be a C3 for brown?
As that breaker is in the bin the question is moot, it was just a wondering?

 

[Inteificio]

question removed due to muppetness

 

[Inteificio]

You'll be replacing a C2 under one reg. with a C2 under another.

Care to expand?

 

[Julie.]

Yeah re-reading what I wrote is as clear as mud!

What I meant with the "1.5mm MCB":

The offending spur is 2.5 T+E on a B20, backed by an RCD.

My understanding is an MCB with sufficient fault current will disconnect in 0.1s. regs p370 (though you suggest 0.01? why?).

With my understanding, an 800A earth fault current will overheat a 1.5mm CPC in 0.1s.

However, the RCD would trip much faster than the 0.1s (can I use the actual of 17ms or the 40ms from minimum)?

So for the CPC overheating, I'm OK (please confirm).

What I'm wondering about is the potential for a short circuit at that socket.

The short circuit would be 2.5/2.5 and 1.2kA.

That would trip in the 0.1s and still cook.



The C3 was more of a theoretical. If I had repaired the ring and left the original breaker, as we have a probable cause for the browning that had been repaired, would it be a C3 for brown?
As that breaker is in the bin the question is moot, it was just a wondering?

Check the characteristics from a manufacturer, the 0.1s trip time is the maximum once you start to get into the "instantaneous " portion of the curve.

However 800A is around 40x mcb rating actually within the current limiting portion of the characteristic. To limit the current it has to operate within a half cycle of the ac wave - so must be less than 10ms

 

[Inteificio]

Check the characteristics from a manufacturer, the 0.1s trip time is the maximum once you start to get into the "instantaneous " portion of the curve.

However 800A is around 40x mcb rating actually within the current limiting portion of the characteristic. To limit the current it has to operate within a half cycle of the ac wave - so must be less than 10ms

Hi,

I've never actually come across this energy limiting class being used before. I only knew it existed as I wondered what the 3 on the front of the mcb was =)

Google was not kind with this. Its been discussed quite rarely and most of the time it's been abusing the guy asking the question, not helping! We've all got to learn sometime!

So Schneider puts class 3, 6k b20 as 45kA

Is it is simple as 4500=1200²xs.

S being the disconnection time?

Or have I guessed that all wrong?

If I am wrong anyone know anywhere that explains?

Thanks

 

[timhoward]

@Inteificio you have actually pin pointed (or stumbled onto?!) one of the interesting areas where the regs alone don't completely answer the question.
It reminded me of an interesting thread on another forum from years ago which I've attempted to link to.

IET EngX

tinyurl.com

Plugging 800A fault current into the adiabatic with 0.1s disconnection time (the lowest the regs give us) gives a concerning answer, as you identified I think. But you need to read the bit in the box on the graph that says for larger PFCs use manufacturers data.

As above, we need a bit more info from the manufacturer to be reassured that the device will actually operate quicker than 0.1 s.
Manufactures will list a "Rated short circuit breaking capacity", i.e. it will cope with the fault current. At least 6kA usually.
And the manufacturers graphs usually go a lot lower than 0.1s. e.g. this Hager B16:


Plugging 800A into the adiabatic with 0.01s gives you a much more reassuring answer.
Hope that helps.

 

[Julie.]

Hi,

I've never actually come across this energy limiting class being used before. I only knew it existed as I wondered what the 3 on the front of the mcb was =)

Google was not kind with this. Its been discussed quite rarely and most of the time it's been abusing the guy asking the question, not helping! We've all got to learn sometime!

So Schneider puts class 3, 6k b20 as 45kA

Is it is simple as 4500=1200²xs.

S being the disconnection time?

Or have I guessed that all wrong?

If I am wrong anyone know anywhere that explains?

Thanks

It is probably an area that most forget, but the attachment here should help.

Ignore the "special" side of things, that is marketing, as if the other manufacturers don't do the same - they do!!

If you obtain the characteristics from the mcm manufacturer, it should both show how quick it actually trips, and the i^2t let through, which you can use to check coordination.


EDIT:
the characteristic posted by timhoward is typically of manufacturers' data.

At 40x it is going to be current limiting i.e. operates before the half cycle of 50hz completes - so actually less than 10ms.

Even if it doesn't current limit, it would clear on the first zero crossing, which is 10ms at 50hz & 8.33ms at 60hz

 

[pc1966]

Just as @Julie. says, you really need to get the manufacturer's data for more precise let-through values.

There are generic limits for this energy-limiting class of MCB and you see resulting limit values in, for example, Table B7 of the On-Site Guide. However, they are a touch pessimistic and most folks would just use the "standard circuits" of Table 7.1(i) to see if the combination of fuse or MCB, cable, and length are acceptable.

Here is an example taken from the Hager commercial catalogue technical information section for the B-curve MCB. What you see the the I2t let-through drop slowly as you go to higher thermal overload, then suddenly drop when you hit the "instant" magnetic trip region. Beyond that it increases as the PFC increases since the breaker does not open significantly faster with more overload.

But it does open a little faster, as a true "constant time" opening would be increasing at two decades of I2t for one decade of PFC (due to the I-squared part, with t=constant).

Behold:

 

[pc1966]

You can compute the equivalent "virtual opening time" from the I2t and I values. For example, taking the 6A MCB curve as easier to follow:

• At 30A PFC (Icc) the I2t is about 0.009k hence t = I2t / I^2 = (0.009 * 1000) / (30 * 30) = 0.01 = 10ms
• At 700A the I2t is about 1k hence t = (1 * 1000) / (700 * 700) = 0.002 = 2ms

In reality the opening is not instant after physical movement as it arcs, and there are other aspects to the MCB that limit peak current, but it serves to illustrate that an energy-limiting breaker opens in well under a cycle of the AC supply.

RCD are not energy-limiting in any real sense as they typically take a cycle or two of the mains to open (specification is below 40ms I think) and that is not significantly faster once you get beyond 5*In test current. Great for shock or smouldering fire prevention, but not helping much for a high current fault. In fact many domestic RCD are only rated to open 1.5kA which is below the sort of PFC you can see close to the DB. In these cases the manufacturer usually stipulates it should have a MCB that opens below 1.5kA (which is faster) so the RCD is not the one breaking a high fault current.

Even a 50A D-curve MCB meets this at 20*50A = 1kA but you will hardly see that is use domestically as it is hard to meet 5s disconnection Zs for a sub-main (compared to switched-fuse which is often better at selectivity with downstream MCBs), and even harder to meet the 0.4s for a final circuit.