For a traditional  fuse  the I²t changes slowly  with fault current and it is safe to assume for large faults that  it is more or less constant and so is the energy let through and the downstream damage done while it blows.

However  an MCB with moving parts becomes almost a constant time device at the really fast end of the curve, as the speed limits of what can be done with mechanical contact separation are reached- so the damage, the let through energy and the I²t all rise as the square of the current.  A modest increase in PSSC can have a devastating consequence for the downstream wiring, and how chunky it needs to be to not have its insulation cooked, and as Lyle notes, the reverse is also true, a few tens of milliohms down the wire, and no amount of vampire hunters hammering silver stakes into the wiring is going to kill it in any more serious way than firing the trip.

However to keep the burden of testing on makers of MCBs and so on to a reasonable level, this is generally only tested at the worst case - the full 6000 A or whatever, and a few other points, so again you are in a rather grey area  and guessing what the performance will be in between.

It is maybe helpful  to work out how fast the breaking is at 6000 A to get say 100,000 J/ohm  of the 32A C type breaker

I² = 36million A², so if I² t is 0.1 million we are looking at  1/360 of a second, or about 3 milliseconds . Now that is a sort of rectangular equivalent time, as if the full fault current flows and then stops dead. It is important to realize that it does not.

Really as an arc forms and stretches, there  is a low and variable resistance in the path that rises and then becomes infinite, so the current waveform has something of a tadpole shape, by which I mean a large head and a thin tail, probably only dying out totally at the next zero crossing of the AC cycle. (and we sort of know it relies on this this as we can set fire to an ordinary AC only MCB if we ask it to interrupt a DC of the same voltage.)

For a traditional  fuse  the I²t changes slowly  with fault current and it is safe to assume for large faults that  it is more or less constant and so is the energy let through and the downstream damage done while it blows.

However  an MCB with moving parts becomes almost a constant time device at the really fast end of the curve, as the speed limits of what can be done with mechanical contact separation are reached- so the damage, the let through energy and the I²t all rise as the square of the current.  A modest increase in PSSC can have a devastating consequence for the downstream wiring, and how chunky it needs to be to not have its insulation cooked, and as Lyle notes, the reverse is also true, a few tens of milliohms down the wire, and no amount of vampire hunters hammering silver stakes into the wiring is going to kill it in any more serious way than firing the trip.

However to keep the burden of testing on makers of MCBs and so on to a reasonable level, this is generally only tested at the worst case - the full 6000 A or whatever, and a few other points, so again you are in a rather grey area  and guessing what the performance will be in between.

It is maybe helpful  to work out how fast the breaking is at 6000 A to get say 100,000 J/ohm  of the 32A C type breaker

I² = 36million A², so if I² t is 0.1 million we are looking at  1/360 of a second, or about 3 milliseconds . Now that is a sort of rectangular equivalent time, as if the full fault current flows and then stops dead. It is important to realize that it does not.

Really as an arc forms and stretches, there  is a low and variable resistance in the path that rises and then becomes infinite, so the current waveform has something of a tadpole shape, by which I mean a large head and a thin tail, probably only dying out totally at the next zero crossing of the AC cycle. (and we sort of know it relies on this this as we can set fire to an ordinary AC only MCB if we ask it to interrupt a DC of the same voltage.)