IET logo
 
IET
Decrease font size
Increase font size
Topic Title: Maximum let-through energy of fuse into cable
Topic Summary: Is maximum at current corresponding to a 5 seconds disconnect time
Created On: 10 July 2012 05:57 PM
Status: Post and Reply
Linear : Threading : Single : Branch
<< 1 2 3 Previous Next Last unread
Search Topic Search Topic
Topic Tools Topic Tools
View similar topics View similar topics
View topic in raw text format. Print this topic.
 12 July 2012 01:50 PM
User is offline View Users Profile Print this message



Gilbert1

Posts: 19
Joined: 06 July 2012

aha I c....so i'll just have to be a good boy then and await the seal of approval :-)
 12 July 2012 03:32 PM
User is offline View Users Profile Print this message



AJJewsbury

Posts: 11251
Joined: 13 August 2003

aha I c....so i'll just have to be a good boy then and await the seal of approval :-)

That's about the size of it
- Andy.
 16 October 2012 07:47 AM
User is offline View Users Profile Print this message



Spark6

Posts: 17
Joined: 17 May 2012

I have just seen this older thread and have had similar questions to Gilly.

Two things that complicate this analysis further:
(1) Fuse manufacturers usually only show the pre-arcing curve on time current curves. You actually need the total clearing curve to apply the I^2t equation at the 5 second mark.
(2) Does anyone have an opinion or know of a standard that says you need to do the check at 5 seconds. If the minimum fault current (to determine disconnect time) is found to be say 0.3 seconds, do you need to worry about anything above this? I would argue that a high impedance fault would be more likely (that takes you in that 0.3 to 5 second range) so you probably should check this all the way to 5s, but we don't consider high impedance faults to calculate disconnect time, even though they might be more likely.
 16 October 2012 09:35 AM
User is offline View Users Profile Print this message


Avatar for OMS.
OMS

Posts: 19428
Joined: 23 March 2004

Originally posted by: Spark6

I have just seen this older thread and have had similar questions to Gilly.

Two things that complicate this analysis further:

(1) Fuse manufacturers usually only show the pre-arcing curve on time current curves. You actually need the total clearing curve to apply the I^2t equation at the 5 second mark.

Don't they show the trailing edge total I2t curve ?

(2) Does anyone have an opinion or know of a standard that says you need to do the check at 5 seconds.

For short circuit, have a look at 434.5.2 (Note 2) which leads you to BS 7454 - the adiabatic or otherwise nature of conductor size contrasted with time is implied here. There is no specific requirement for 5 seconds, just that you don't damage the cable - 5 seconds is around the point where theings cease to be adiabatic and values of K may change, to benefit the design

If the minimum fault current (to determine disconnect time) is found to be say 0.3 seconds, do you need to worry about anything above this?

For minimum compliance, usually no

I would argue that a high impedance fault would be more likely (that takes you in that 0.3 to 5 second range) so you probably should check this all the way to 5s, but we don't consider high impedance faults to calculate disconnect time, even though they might be more likely.

It's usually iterative - many engineers will want to know the failure point of a design - personally I tend to go around the loop of increasing time and/or diminishing fault current until I reach a failure point - when I know tthat, I decide how credible that point is, in term of realistic faults
.


regards

OMS

-------------------------
Failure is always an option
 17 October 2012 02:48 AM
User is offline View Users Profile Print this message



Spark6

Posts: 17
Joined: 17 May 2012

Thanks OMS,

Do you have a link to to any fuse data that actually shows an I2t curve? (For fuses that comply to BS88-2 or IEC60269-2). I have had trouble finding this. Usually you only get Time-Current curves for pre-arc, peak-current cut-off curve and a single number for max pre-arc energy and max total energy.

So from your response, are you saying that you basically don't need to check cable thermal stress for time durations above the calculated disconnect time for the minimum fault current? I think needing to check to 5s would produce some nasty surprises. I have seen some 20A breakers (fault current limited) that don't appear to protect a 2.5mm2 PVC cable at the 5s point....it isn't limited to fuses.
 17 October 2012 11:00 AM
User is offline View Users Profile Print this message


Avatar for Parsley.
Parsley

Posts: 960
Joined: 04 November 2004

I think needing to check to 5s would produce some nasty surprises. I have seen some 20A breakers (fault current limited) that don't appear to protect a 2.5mm2 PVC cable at the 5s point....it isn't limited to fuses.


2.5mm2 conductor has a K2S2 of 82,656 or 127,806 depending on which K value you select.
Merlin Gerin 20Amp type B breakers appears to limited to approx 30,000A2s at 5 secs, so fine for 2.5mm2 but right on the limit for a 2.5mm/1.5mm2 T&E (29,756A2S) if i've got my calcs right.

Regards
 17 October 2012 11:26 AM
User is offline View Users Profile Print this message



AJJewsbury

Posts: 11251
Joined: 13 August 2003

2.5mm2 conductor has a K2S2 of 82,656 or 127,806 depending on which K value you select.
Merlin Gerin 20Amp type B breakers appears to limited to approx 30,000A2s at 5 secs, so fine for 2.5mm2 but right on the limit for a 2.5mm/1.5mm2 T&E (29,756A2S) if i've got my calcs right.

Sounds about right to me.

Energy let-though for MCBs normally increases with fault current (i.e. the opposite to most fuses) - so worst case is usually at the start of the circuit (lowest Zs = Zdb).

Performance varies with manufacturer, but for generic BS EN 60898 MCBs, you'd need a minimum 2.5mm² conductors (including c.p.c.) for a 20A B type device if the fault current could exceed 3kA (and not exceeding 6kA). (See table A5 in appendix A of GN 3 or the OSG).
- Andy.
 17 October 2012 11:54 AM
User is offline View Users Profile Print this message


Avatar for Parsley.
Parsley

Posts: 960
Joined: 04 November 2004

As Andy stated Performance varies with manufacturer, but for generic BS EN 60898 MCBs, you'd need a minimum 2.5mm² conductors (including c.p.c.) for a 20A B type device if the fault current could exceed 3kA (and not exceeding 6kA). (See table A5 in appendix A of GN 3 or the OSG).

A 32 Amp type C merlin breaker has a A2S of approx 22,500 at 4KA, so in theory a 1.5mm2 cpc will be able to handle the energy let through. So worth checking manufacturers data as the performance may be better than quoted in OSG.

Regards
 17 October 2012 12:00 PM
User is offline View Users Profile Print this message


Avatar for OMS.
OMS

Posts: 19428
Joined: 23 March 2004

Is it likely that an MCB would be lingering on for 5 seconds. ?

Most MCB's (provided you have reached the minimum tripping current will be open in less than 0.1 seconds and usually close to the minimum definite tripping time of 0.01 seconds.

We are clearly into energy limitation at this point as we are into the first few cycles.

Equally, power factor of the fault is still probably quite low at this stage (if you have high fault current, then you probably have low resistance but high reactance distribution)

Given that we are talking about devices intended for household or similar applications, seeing If above 3kA isn't that credible, and almost certainly won't be exceeding 6kA. (regardless of what the loop tester estimates it as)

Now if designers want to take this simple MCB technology and put it into place in high fault current scenarios, then they should also be aware of the impact on cable sizing. If you really don't want to do the sums, then CPC = Line is a good starting point (up to say 10mm2) and a bit of metallic containment works wonders as well (big chunks of steel in parallel with weedy little cables changes the whole dynamic)

The danger arises when we take reduced CPC type cables into generally non domestic situations with high(er) fault levels - and then decide against all the evidence that disconnection could linger on for several seconds.

At the other end of the scale, the last hospital I was at has 2 x 1 x 4C x 300mm2 XLPE/SWA/LSF(Cu) ring mains supplied at each end by 630A circuit breakers - depending on how the ring is configured, and the positionof the fault, disconnection could take up to 10 seconds - now that's a bloody long time to be pouring fault current into a cable - but they are sized to take it, and upstream distribution from the 630A breakers is also sized to take it - but it costs a lot of money.

It's all about the energy - as long as we don't stress the insulation and terminals, then disconnection time isn't important, and things start to be a bit less adiabatic once you have bigger cables and long(er) disconection times anyway. Alternative values of "K" work wonders for the adiabatic expression -

regards

OMS

-------------------------
Failure is always an option
 17 October 2012 11:30 PM
User is offline View Users Profile Print this message



Spark6

Posts: 17
Joined: 17 May 2012

OMS - Any chance you could let me know where I could get hold of that fuse data? (see my previous post).

Yes a type C Merlin Gerin C60N breaker lets through 40,000 A2S at 5s, so it should be fine. Have any of you noticed that the let-through curves show the magnetic trip point at 7.5xIn, when it could be 10xIn with tolerance? Does this mean the let-through could be higher? I'm not sure. The amount of energy at 10x compared to 7.5x is larger by a factor of 1.78. That gets you alot closer to the limit. Does it work like that in practice - probably not, but the manufacturer would need to answer that.

The MCB I was refering to is a Terasaki DIN-T 20A. It is probably fine too, but the let-through curve doesn't show anything below 1kA. So to check the 5 second portion you need to take the current at 5s on the time-current curve. 200A can flow for 5 seconds (worst case). This equates to I2t of 200,000 I2t, easily exceeding what a 2.5mm2 pvc cable can handle. Now again, it is likely that it would be fine if the let-through curve showed the region below 1kA, but based on the information provided you can't be sure. So it goes back to my question, do we really need to worry about checking at 5s, because if we do, this breaker can't be used without more information.

OMS asked if it was "likely" that an MCB would be lingering on for 5 seconds. I would say yes. We need to work on the assumption that some unlikely events can occur. We need to design installations for the possibility of bolted 3 phase faults - that is a far more unlikey scenario than a high impedance fault occuring on a MCB.

Talking about whether an MCB is installed in a high fault current situation doesn't matter for this discussion, we are only talking about the 5 second region - these can happen regardless of installation environment.
 18 October 2012 01:10 AM
User is offline View Users Profile Print this message



Spark6

Posts: 17
Joined: 17 May 2012

Hi Guys,

I re-read some of your posts. I'm not from the UK. What does C.P.C stand for? What is the OSG and GN 3?
I've looked into this a bit more. BS EN 60898 devices have a let-through energy class. Looks like most are class 3, which has the highest limitation. Does anyone have access to access to BS EN 60898 that can quantify what the range for class 3 is? I assume this is the point Andy was touching on?

I would have still thought this wouldn't answer our question with relation to energy let-through at the 5 second point.
 18 October 2012 10:37 AM
User is offline View Users Profile Print this message


Avatar for OMS.
OMS

Posts: 19428
Joined: 23 March 2004

Originally posted by: Spark6

Hi Guys,

I re-read some of your posts. I'm not from the UK. What does C.P.C stand for? What is the OSG and GN 3?

CPC - circuit protective conductor - the earthing or grounding wire
OSG - On site Guide, a shortform publication by IET to assist compliance/undertsanding of BS 7671 for use in simple installations
GN3 - Guidance note 3 - one of eight guidance notes published by IET to expand on and clarify BS 7671 - GN 3 being focused on Inspection and Testing


I've looked into this a bit more. BS EN 60898 devices have a let-through energy class. Looks like most are class 3, which has the highest limitation. Does anyone have access to access to BS EN 60898 that can quantify what the range for class 3 is? I assume this is the point Andy was touching on?

Class 3 limits are set in Annex ZA for type B and C breakers in a range of nominal current ratings for stated fault levels.

so your " Terasaki DIN-T 20A" if compliant with BS EN 60898 and being a Class 3 devices will have (for Type B), energy limitation of:

3.0kA - 18000A2S
4.5kA - 32000A2S
6.0kA - 45000A2S
10kA - 90000A2S

This is in a range for MCB's (Type B) of 20A, 25A and 32A



I would have still thought this wouldn't answer our question with relation to energy let-through at the 5 second point.

To clarify your thinking, energy let through is defined in units of Amperes x Amperes x Seconds, ie it is the Joule integral (Energy) - it is a total value - it does not exist at a 5 second point or any other point - it's the total energy let through


regards

OMS

-------------------------
Failure is always an option
 18 October 2012 02:38 PM
User is offline View Users Profile Print this message



Spark6

Posts: 17
Joined: 17 May 2012

Hi OMS, thanks for the reply and info.

When I say I2t at the 5 second point, i really mean the current at which it would take 5 seconds to disconnect. So for a 20A type C MCB, that could be 200A. Yes, I understand the let-through is the integral of current squared over this time interval 0-5s.

So, back to my question - the limits set out in 60898 Annex ZA, are they the maximum let-throughs that could occur for any fault duration up to 5 seconds (adiabatic region)? or are they just for the higher fault current end at the fault currents nominated?

If it is the latter - I am still back to the point that you can cook a 2.5mm2 cable served by a 20A type C breaker if you expose it to a high impedance fault that resulted in 200A fault current. It is unlikely (you need the breaker to be operating in the outer tolerance allowed - and increasing resistance due to heat rise of the conductors might make it a tough window to hit) but still possible.

Worth losing sleep about?
 18 October 2012 02:41 PM
User is offline View Users Profile Print this message



Spark6

Posts: 17
Joined: 17 May 2012

OMS - don't suppose you have those annex ZA figures for type C MCBs?
 18 October 2012 06:13 PM
User is offline View Users Profile Print this message


Avatar for OMS.
OMS

Posts: 19428
Joined: 23 March 2004

Originally posted by: Spark6

Hi OMS, thanks for the reply and info.

When I say I2t at the 5 second point, i really mean the current at which it would take 5 seconds to disconnect.

OK that's a different thing though, MCB's won't take 5 seconds to disconnect unless thay are operating in the thermal region (for overload) - you now have much reduced current so the energy let through will be small. Once you reach the magnetic part of operation things speed up - see below

So for a 20A type C MCB, that could be 200A.

At 200A the breaker will operate at 0.1 seconds or less as you have reached 10 x In

Yes, I understand the let-through is the integral of current squared over this time interval 0-5s.

Well to be exact it's the time interval for disconnection to occur - it could be a lot faster than 5 seconds

So, back to my question - the limits set out in 60898 Annex ZA, are they the maximum let-throughs that could occur for any fault duration up to 5 seconds (adiabatic region)? or are they just for the higher fault current end at the fault currents nominated?

Well, they are the maximum energy let through that could occur at that fault level - they are not referenced to 5 seconds because we are talking about a joule integral - t will be what it is, usually between 0.1 seconds and 0.01 seconds

So from my examples for type B above when you have a 6kA fault level the let through energy would be 45000A2S - for a 2.5mm2 copper conductor with XLPE insulation and a K value of 100, then the cable withstand is 62500A2S. For a 1.5mm2 with K at 115, the cable withstand is 29800A2S. In the first case the cable is protected, in the latter it isn't. checking again at 3KA, the energy let through of the device is 18000A2S - the cable exceeds this in terms of thermal capacity.

So the lesson for today is don't just chuck in MCB's and expect them to protect circuits not designed for the prevailing fault level, even if you use current limiting MCB's.



If it is the latter - I am still back to the point that you can cook a 2.5mm2 cable served by a 20A type C breaker if you expose it to a high impedance fault that resulted in 200A fault current. It is unlikely (you need the breaker to be operating in the outer tolerance allowed - and increasing resistance due to heat rise of the conductors might make it a tough window to hit) but still possible.

Not likley due to the operating time of the MCB. At 200A both the thermal and magnetic protection will be operating , but yes, in theory putting 200A down a 2.5mm2 cable for 5 seconds will cook it

Worth losing sleep about?

Not really - there a probably millions of 20A MCB's on 2.5mm2 circuits in both low and high current fault situations - the country isn't littered with burnt out buildings last time I looked


Regards

OMS

-------------------------
Failure is always an option
 18 October 2012 06:16 PM
User is offline View Users Profile Print this message


Avatar for OMS.
OMS

Posts: 19428
Joined: 23 March 2004

Originally posted by: Spark6

OMS - don't suppose you have those annex ZA figures for type C MCBs?


For 20A, 25A and 32A type C

3.0kA - 20000A2S
4.5kA - 37000A2S
6.0kA - 52000A2S
10kA - 100000A2S

let me know if you need other device ratings (up to 63A)

regards

OMS

-------------------------
Failure is always an option
 18 October 2012 06:56 PM
User is offline View Users Profile Print this message



mawry

Posts: 219
Joined: 26 April 2004

Fair play OMS you should write a book!
 18 October 2012 07:41 PM
User is offline View Users Profile Print this message


Avatar for OMS.
OMS

Posts: 19428
Joined: 23 March 2004

The Taffia Memoirs -

If I wrote a book, you lot would pick bloody great holes in it - now where's the fun in that -

OMS

-------------------------
Failure is always an option
 18 October 2012 08:45 PM
User is offline View Users Profile Print this message



mawry

Posts: 219
Joined: 26 April 2004

Plenty of fun for us!
 19 October 2012 12:11 AM
User is offline View Users Profile Print this message



Spark6

Posts: 17
Joined: 17 May 2012

Thanks OMS. I'd love to get all the content of Annex ZA but that might be pushing the friendship :-)

I'll need to agree to disagree with your statement:

"At 200A the breaker will operate at 0.1 seconds or less as you have reached 10 x In"

As an example, for a Schneider 20A C60 C Curve breaker, you only need to miss 200A by a fraction and you go can from magnetic to thermal - the difference between less than 0.1 seconds and over 3 seconds clearance time. For the same breaker, 170A (8.5x) could take 5 seconds to disconnect. Yes, you have gone into the thermal region, but if we work on the assumption that no heat will be dissipated from the faulted cable for 5 seconds then you still need to consider it. So from 170A to 199.99A, I'd assume you still need to apply the adiabatic equation to this region regardless of if it is thermal or magnetic. If you slip into the thermal zone and it is still possible to clear the fault in under 5 seconds, this is obviously far worse than clearing a 200.01A fault in under 0.1 seconds.

Maybe the confusion is that I am talking about C type breakers and it is more common to use B type breakers in the UK? This isn't an issue with B type breakers - just C and D.

I guess my example is bordering on theoretical, but it is possible. It has just always left me feeling slightly uneasy. I should point out that the let-through characteristic for the same device suggests some sort of current limitation even around the 5 second region - not sure how that works given that it is obviously not interrupting in the first half cycle anymore, but that should further support that this is not really worth worrying about.

I agree there is a lot of 2.5mm2 cable fed from 20A MCBs and it doesn't seem to be a problem - but I wonder if every now and them a cable gets a little bit toastier than it should be and nobody really notices.....
IET » Wiring and the regulations » Maximum let-through energy of fuse into cable

<< 1 2 3 Previous Next Last unread
Topic Tools Topic Tools
Statistics

See Also:



FuseTalk Standard Edition v3.2 - © 1999-2014 FuseTalk Inc. All rights reserved.