In uk.d-i-y, ben wrote:

That's interesting to know. The electromechanical bit -- this is
presumably just a relay coil wound and arranged so as to pull the switch
open at the appropriate current? If so, isn't it a worry that a massive
overcurrent will melt the wiring closed-circuit, so that the next time
the MCB is required it doesn't work?

I suppose the manufacturers have thought of that, however .

They, and those nice real-engineer types at the IEE. The mains supply
itself has a finite impedance, limiting the "prospective fault current"
(I think I remember the jargon aright). The combination of the quick action
and the "you won't get more than, say, 6kA out of a normal domestic
main, even with a dead short" allows the mfr to size the coil and so
on of the MCB so that the total energy dissipated won't heat up the
device beyond use. In practice, the current passing in a "dead" short is
further limited by the resistance of the cables between the MCB and
the point where the short is - shorts to neutral or earth usually happen
somewhere Out There in the house, not a few inches from the MCB with a
massive lump of copper between the MCB output terminal and the earth
or neutral busbar.

"Unless", as we say in homage to She Who Allegedly Named RCDs, "you
know better"...