Relying on loads not being able to overload

[JohnW2]

Of course the suggested plasma forms within the MCB!

As I implied, that's what I suspected.

In that case, it has nothing to do with what interests us - namely the process going on in a CPC during fault conditions which I still think is (when t<0.01s) going to be just about as close as one can get in practice to a perfectly adiabatic process (very little heat is going to have a chance to move from the conductor is such a short period of time).

As I've said, the only problem with undertaking an adiabatic calculation (for the CPC) with these very small disconnection times is finding the appropriate value of I²t to use for the calculation.

Kind Regards, John

 

[JohnW2]

With disconection times less than 0.01 you need to refer to the manufacturers A2S/I2t, where large fault current exist, these figures are derived through test and measurement, though its still definitely an adiabatic process.

Sure, but in keeping with what you suggest, most of the I²t data I've seen from manufacturers relates to extremely large fault currents - e.g. the example tabulations of manufacturer's I²t data provided in the IET Design Guide covers fault currents from 10kA to 50kA.

For the purpose of this discussion, what we need are I²t figures for normal domestic fault currents - in fact, right down to the current which will operate the magnetic trip of the OPD (what BS7671 calls Ia).

I'm still rather confused. Some people seem to be implying that use of manufacturer-supplied 'energy let through' (I²t) figures (if they are available for the currents which interest us) for calculations is in some way a different process from using figures for I and t from sufficiently detailed and complete t/I curves (if they are available). I must be missing something, because the two processes seem to me to be exactly the same thing. Can you help me understand?

Kind Regards, John

 

[JohnW2]

In an attempt to drag this thread back onto topic ...

The conclusion of the last few pages of discussion is probably that it is reasonable to assume a disconnection time of no more than 0.02 seconds, since it seems that will probably achieved by most/all MCBs with a fault current equal to the "Ia" of the MCB (e.g. 5*In for a Type B). What we have seen so far suggests that, at least with some MCBs, getting down to 0.01 secs may require a current higher than “Ia” (e.g. perhaps 6*In for a Type B, per Hager curves), and therefore would probably not be appropriate for calculations if one wished them to remain correct for all possible levels of fault currents.

On that basis, I have re-tweaked my earlier tabulation to reflect a disconnection time of 0.02 secs (and, to be conservative, I've used the reduced 'maximum Zs' figures which will probably come into force next year, and have also included 'minimum Zs' figures based on 253V, as well as 230V):

Untitled

• JohnW2
• 17 Apr 2014
• 1

As can be seen (and as would be expected), on this basis the range of Zs values for which CPCs are adequately protected against fault current becomes much wider. For example, in terms of one of the early extreme examples we discussed, it appears that the CPC (and live conductors) of 1mm² T+E would probably be adequately protected by a B50 MCB for Zs values in the range 0.31&#937; to 0.70&#937;.

Kind Regards, John

 

[THRIPSTER]

John,

The discussion has always been on track and, due to the discourse that has taken place and feedback you have received, you have revised the figures you believe acceptable for minimum Zs downwards from 0.63 to 0.28 ohms in your latest table (in the case of 1+1 T&E). Therefore, you must believe that the discussion is relevant and on track.

I have raised the issue of 0.1 disconnection time on the IET forum and have been referred to 434.5.2 (which was pretty much where I started from) and the use of manufacturers I²t data for specific MCB's below that time. In the mean time, I have a query with Mark Coates and await to hear whether he is able/willing to respond.

Regards

 

[JohnW2]

John, The discussion has always been on track and, due to the discourse that has taken place and feedback you have received, you have revised the figures you believe acceptable for minimum Zs downwards from 0.63 to 0.28 ohms in your latest table (in the case of 1+1 T&E). Therefore, you must believe that the discussion is relevant and on track.

You sound a little defensive, but I was in no way being critical of you, nor was I underestimating the importance of the part of the discussions you initiated - which, as you say, have led me to feel that it is reasonable to examine CPC protection on the basis of an assumed disconnection time of 0.02s, rather than the original 0.1s. My comment about getting the thread back 'on topic' mean literally what it said - considering what the revised calculations meant in terms of fault protection of the CPC of an 'undersized cable' in which it was being assumed that overload protection was not required.

FWIW, on the evidence that we have so far seen, I don't think it is 'safe' to assume a disconnection time of 0.01s if one wants one's calculations to be generally applicable to all compliant circuits. The t/I curves from at least two of the three (and probably also the third) manufacturers we have seen indicate that a disconnection time of 0.01s is not necessarily achieved at a fault current of just "Ia" (i.e. for a circuit just compliant in terms of ADS), whereas all three would achieve 0.02s at that current.

I have raised the issue of 0.1 disconnection time on the IET forum and have been referred to 434.5.2 (which was pretty much where I started from) and the use of manufacturers I²t data for specific MCB's below that time.

To be fair, it says one should use the I²t data either from the relevant Standard or from the manufacturer.

I continue to be confused by the way in which people seem to feel that when it is called 'energy let through' "I²t data from the manufacturer" is something different from the I²t one would/could determine from appropriate t/I curves provided by the manufacturer. We have now found such curves from three of the most major manufacturers (Hager, Wylex and MK) which go down to below 0.01s, so I can't see what useful (for our purpose) additional information would exist in "I²t data from the manufacturer" if it could be obtained. Am I missing something?

In the mean time, I have a query with Mark Coates and await to hear whether he is able/willing to respond.

It would obviously be interesting if he did. I would, as above, be particularly interested to hear whether he really feels that it is 'safe' to make an assumption of a 0.01s disconnection time (for 'generally-applicable' calculations) when the evidence we have seen is that the MCBs from at least some manufacturers will not necessarily achieve that (although they will achieve 0.02s) at fault currents which are only just beyond the "Ia" of the device.

Kind Regards, John

 

[christof22]

With disconection times less than 0.01 you need to refer to the manufacturers A2S/I2t, where large fault current exist, these figures are derived through test and measurement, though its still definitely an adiabatic process.

Sure, but in keeping with what you suggest, most of the I²t data I've seen from manufacturers relates to extremely large fault currents - e.g. the example tabulations of manufacturer's I²t data provided in the IET Design Guide covers fault currents from 10kA to 50kA.

For the purpose of this discussion, what we need are I²t figures for normal domestic fault currents - in fact, right down to the current which will operate the magnetic trip of the OPD (what BS7671 calls Ia).

I'm still rather confused. Some people seem to be implying that use of manufacturer-supplied 'energy let through' (I²t) figures (if they are available for the currents which interest us) for calculations is in some way a different process from using figures for I and t from sufficiently detailed and complete t/I curves (if they are available). I must be missing something, because the two processes seem to me to be exactly the same thing. Can you help me understand?

Kind Regards, John

Hi John

Manufacturers info on I2t starts much lower than the 10 kA, if you are looking at low fault current levels then i doubt there will be much of a difference between BS7671 and the manufactures info, as fault currents increase energy limiting kicks and manufactures CPD will differ in how they limit I2t.

Cheers
Chris

 

[JohnW2]

I'm still rather confused. Some people seem to be implying that use of manufacturer-supplied 'energy let through' (I²t) figures (if they are available for the currents which interest us) for calculations is in some way a different process from using figures for I and t from sufficiently detailed and complete t/I curves (if they are available). I must be missing something, because the two processes seem to me to be exactly the same thing.

Hi John ... if you are looking at low fault current levels then i doubt there will be much of a difference between BS7671 and the manufactures info, as fault currents increase energy limiting kicks and manufactures CPD will differ in how they limit I2t.

Yes, that all makes sense, but it's not quite what I was asking about. I was talking about the alleged difference (i.e. a difference which some people seem to imply exists) between (a) a value of I²t ('let through energy') provided by a manufacturer for a certain device and fault current, and (b) the value of I²t which one would calculate (for the same fault current) from the values of t and I obtained from a t/I curve supplied by the same manufacturer for the same device. Unless I'm missing something, they have got to be the same, yet some people seem to talk as if they are 'different approaches'?

Kind Regards, John

 

[EFLImpudence]

Not a dig but it confuses me a lot.
It's quite often printed in documents as well.
So, for the benefit of others who may be confused.

I2t is just poor typing for I²t isn't it? Not something else?

I2t actually means I x 2 x t or does it?

 

[christof22]

I'm still rather confused. Some people seem to be implying that use of manufacturer-supplied 'energy let through' (I²t) figures (if they are available for the currents which interest us) for calculations is in some way a different process from using figures for I and t from sufficiently detailed and complete t/I curves (if they are available). I must be missing something, because the two processes seem to me to be exactly the same thing.

Hi John ... if you are looking at low fault current levels then i doubt there will be much of a difference between BS7671 and the manufactures info, as fault currents increase energy limiting kicks and manufactures CPD will differ in how they limit I2t.

Yes, that all makes sense, but it's not quite what I was asking about. I was talking about the alleged difference (i.e. a difference which some people seem to imply exists) between (a) a value of I²t ('let through energy') provided by a manufacturer for a certain device and fault current, and (b) the value of I²t which one would calculate (for the same fault current) from the values of t and I obtained from a t/I curve supplied by the same manufacturer for the same device. Unless I'm missing something, they have got to be the same, yet some people seem to talk as if they are 'different approaches'?

Kind Regards, John

Hi John, the Time current curves in BS7671 will be the same as the manufacturers I2t down to the 0.1 second disconnection times.

Cheers
Chris

 

[JohnW2]

Hi John, the Time current curves in BS7671 will be the same as the manufacturers I2t down to the 0.1 second disconnection times.

Sure, but I'm not talking about BS7671 (whose t/I curves don't go below 0.1s anyway), I'm talking about t/I curves published by the manufacturer of a particular device.

If, say (simple numbers for illustration!), a manufacturer said that the I²t (even if they called it 'energy let-through') for a certain device at a fault current of 500A was 2,500 A²sec, would you not expect that the t/I curve published by that manufacture for the same device would indicate that the disconnection time at a fault current of 500A was 0.01s - so that I²t (500² x 0.01) would again be 2,500 A²sec? (In other words, the manufacturer's 'energy let through' figure would be telling you nothing that you could not ascertain from their published t/I curve for the same device)>

And I suppose I should add ... if not, why not?

Kind Regards, John

 

[JohnW2]

I2t is just poor typing for I²t isn't it? Not something else?

I've certainly been assuming that, not the least because the literal mathematical interpretation of "I2t" would make absolutely no sense in terms of what we're discussing.

If you need more clues, few, if any, mathematician (and probably few engineers) would ever write "I2t" to mean that all three were to be multiplied together - they would almost invariably write "2It" or "2tI".

Kind Regards, John

 

[christof22]

Hi John, the Time current curves in BS7671 will be the same as the manufacturers I2t down to the 0.1 second disconnection times.

Sure, but I'm not talking about BS7671 (whose t/I curves don't go below 0.1s anyway), I'm talking about t/I curves published by the manufacturer of a particular device.

If, say (simple numbers for illustration!), a manufacturer said that the I²t (even if they called it 'energy let-through') for a certain device at a fault current of 500A was 2,500 A²sec, would you not expect that the t/I curve published by that manufacture for the same device would indicate that the disconnection time at a fault current of 500A was 0.01s - so that I²t (500² x 0.01) would again be 2,500 A²sec? (In other words, the manufacturer's 'energy let through' figure would be telling you nothing that you could not ascertain from their published t/I curve for the same device)>

And I suppose I should add ... if not, why not?

Kind Regards, John

Hi John

At around 15 kA the fault current will contain a DC component, the fault current is no longer symmetrical, its now asymmetrical, this makes direct calculation difficult and will differ from one device to another.

Cheers
Chris

 

[JohnW2]

Hi John ... At around 15 kA the fault current will contain a DC component, the fault current is no longer symmetrical, its now asymmetrical, this makes direct calculation difficult and will differ from one device to another.

Be that as it may, as I said before, we are talking about 'usual domestic fault currents', not 15 kA or above - and, indeed, the example I asked you about used only 500A.

Kind Regards, John

 

[christof22]

Hi John ... At around 15 kA the fault current will contain a DC component, the fault current is no longer symmetrical, its now asymmetrical, this makes direct calculation difficult and will differ from one device to another.

Be that as it may, as I said before, we are talking about 'usual domestic fault currents', not 15 kA or above - and, indeed, the example I asked you about used only 500A.

Kind Regards, John

Hi John

I believe i have answered your question, down to 0.1 seconds you will derive the same answer, below 0.1 there are other factors, energy limiting, asymmetrical faults.

Example 32 amp type B, 500 amp fault current, I2t = 1.69(kA2S)

Disconnection time = 0.01

Your method would give (500)2 * 0.01 = 2.5(kA2S)

Current limiting lowers the peak kA, as i said this cannot be directly applied to the time current curves, you need to look at the energy let through curves or the tabulated tables.

Cheers
Chris

 

[JohnW2]

I believe i have answered your question, down to 0.1 seconds you will derive the same answer, below 0.1 there are other factors, energy limiting, asymmetrical faults.
Example 32 amp type B, 500 amp fault current, I2t = 1.69(kA2S) Disconnection time = 0.01
Your method would give (500)2 * 0.01 = 2.5(kA2S)

Well, therein lies my problem of understanding.

Are telling me that 1.69 kA²s is the value for I²t given by the manufacturers when I=500 and t=0.01 ? If so, the figure they are giving you is clearly not actually I²t - since, as you illustrate, with those values of I and t, I²t should be 2.5 kA². Are you saying that they report, as "I²t", something which is not actually I²t, or what? Am I misunderstanding what you're saying?

Current limiting lowers the peak kA, as i said this cannot be directly applied to the time current curves, you need to look at the energy let through curves or the tabulated tables.

I must be dim, or going mad, since that makes absolutely no sense to me. If they can determine (empirically or however) the actual I²t for any I, then their published t/I curves surely could (and should) plot the actual t [i.e I²t divided by I²] against I. The curve would then tell the full and correct story, over the entirety of whatever range of values it was plotted.

More specifically, if they can tell me a value for I²t for any value of I (taking into account the various other consideration you mention), they can divide that by I² to get the actual corresponding value of t for any value of I (let's call it T) - so why don't the t/I curves they publish show a plot of T against I? Indeed, if their published curves are not plotting T against I, what on earth are they plotting against I? I do feel as if I need some help before I pull the rest of my hair out

Kind Regards, John