One might think so, but the reality seems to be rather different from that.
As you say, (even though I²t is, contrary to common parlance, not "let through energy') I²t is related to the energy dissipated in any particular fuse - and, as you say, one might expect that the amount of energy required to melt/vaporise a fuse would be roughly constant.
However, if you look, for example, at the curve for the fuse we have been considering (30A BS3036), you will see that, say, to get a 5 secs disconnection time requires roughly 85A, whereas to get a 0.2 sec disconnection time one needs roughly 300A - which correspond to I²t figures of roughly 36,125 and 18,000 A².secs respectively. If one assumes that the resistance (or average resistance as it heats up) is roughly the same in both those situations , that would mean that it takes around twice as much energy to blow the fuse with 85A than it does with 300A. 5 seconds it just about short enough for the process to be essentially adiabatic, so I would expect the temperature rise for a given energy dissipation to be similar in the two cases.
I don't pretend to understand how the above comes about, but am merely observing it. I one looks at lower currents (for disconnection times greater than 5 secs), the I²t for disconnection appears to gets dramatically higher - about 98,000 A².secs for ~70A/20secs and 360,000 A².secs or ~60A/100 secs. Some of that is undoubtedly due to heat loss when one moves away from the adiabatic situation with these longer times, but I find it hard to believe that it can fully explain such dramatic numbers.
Sure - that's precisely what I been talking about/asking for but, as I have implied, I have found no such information relating to BS3036 fuses. I actually wonder whether it exists as such. Cartridge fuses are manufactured, tested and sold as 'fuses', so one can expect full data to be available. However, with BS3036s, what is sold is just the fuse holder, so performance is to some extent dependent upon user actions (in wiring the fuse). Nevertheless, I would expect at least 'typical' figures to be published - after all, we are seeing t/I curves.
One would certainly expect there to be a finite lower bound to disconnection time for any device which involves 'moving parts'. Some of the parts have to have significant size (hence mass) in order to be able to break the current concerned, and hence there will be a finite amount of inertia to be overcome when the device operates. On the other hand, with a fuse, what is 'happening' is essentially at a molecular level, so I don't see why it theoretically can't approach being 'infinitely fast' (at least, in terms of vaporising - arcing complicates things).
Anyway, returning to the actual question (and particularly given what I've just said about the theoretically possible speed of action of fuses), I don't see how (without data) we can know to what extent a 30A BS3036 can act faster than 0.1 secs (with currents greater than 430/450A) - or what the It would be for PEFCs greater than that figure ('the bottom of the BS7671 curve'). The fact that all the t/I curves for fuses in BS7671 stop at 0.1 secs is obviously arbitrary, and certainly cannot be taken to imply that none of them are capable of operating in less than 0.1 secs.
In terms of how this all started, I suppose the situation we have is, in some senses, 'conservative'. If the CSA of a CPC is adequate for a current corresponding to 0.1 seconds disconnection, then it is unlikely to be inadequate at higher PEFCs than that current - although, at least in theory, I don't think that we can necessarily be certain of that without seeing the figures.
I'm struggling to understand some of the data I'm looking at - for example the below from some fairly old Wylex documentation - which, by implication, appears to apply to all types (B, C and D) of their MCBs. The I²t figures seem lower than I would have expected, implying disconnection times roughly in the range 1.5 - 3.0 msec for 32A MCBs over the full range of currents shown (from about 450A upwards) - is that credible?
Kind Regards, John
Last edited:
), equally true of the t/I curves (which are widely available) - so not, in itself, a rational explanation of why I²t data is so hard to find.
