Thanks OMS, so reading that it looks then as if the worst case is going to be at the PSSC = 11*In end of the curve, and only 10, 16, 20 and 25A devices have to be fast , for the rest T diss could be up to 8 seconds. Now in terms of let-through energy and potential to do cable damage, 8 seconds and 5 seconds of heating time really are  significantly different. 

I wonder how many installations have D type breakers, and are thought OK  because they meet volt drop considerations at normal running current, but would not actually operate the magnetic part  into a dead short - these may need to be re-checked assuming  8 seconds for the  adiabatic heating time . Which is awkward, as the adiabatic calc loses validity for long disconnection times as there is time for the heat to spread into the local volume.

Maybe we should assume that the breaker itself is operating as an adiabatic  thermal device, that the curve is such that at 10* In it is 8 seconds, and a thermal fuse like law (i2t is constant ) applies thereafter, so at

11In Tdiss = 8 seconds* (10/11)^2 = 6.6 secs

12In Tdiss = 8 seconds* (10/12)^2 = 5.5 secs

13In Tdiss = 8 seconds* (10/13)^2 = 4.7 secs

14In Tdiss = 8 seconds* (10/14)^2 = 4.1 secs

15In Tdiss = 8 seconds* (10/15)^2 = 3.6 secs


If the breaker is not adiabatic, but has significant cooling over the 8 seconds, then it speeds up faster with increasing fault current  than this assumption suggests, so these maximum breaking times will tend to err on the safe side.


If so the Zs in the table for 5 seconds would need reducing to something like 10/13 of the current value.


Another one to add to the growing list of potential gotchas.

Thanks OMS, so reading that it looks then as if the worst case is going to be at the PSSC = 11*In end of the curve, and only 10, 16, 20 and 25A devices have to be fast , for the rest T diss could be up to 8 seconds. Now in terms of let-through energy and potential to do cable damage, 8 seconds and 5 seconds of heating time really are  significantly different. 

I wonder how many installations have D type breakers, and are thought OK  because they meet volt drop considerations at normal running current, but would not actually operate the magnetic part  into a dead short - these may need to be re-checked assuming  8 seconds for the  adiabatic heating time . Which is awkward, as the adiabatic calc loses validity for long disconnection times as there is time for the heat to spread into the local volume.

Maybe we should assume that the breaker itself is operating as an adiabatic  thermal device, that the curve is such that at 10* In it is 8 seconds, and a thermal fuse like law (i2t is constant ) applies thereafter, so at

11In Tdiss = 8 seconds* (10/11)^2 = 6.6 secs

12In Tdiss = 8 seconds* (10/12)^2 = 5.5 secs

13In Tdiss = 8 seconds* (10/13)^2 = 4.7 secs

14In Tdiss = 8 seconds* (10/14)^2 = 4.1 secs

15In Tdiss = 8 seconds* (10/15)^2 = 3.6 secs


If the breaker is not adiabatic, but has significant cooling over the 8 seconds, then it speeds up faster with increasing fault current  than this assumption suggests, so these maximum breaking times will tend to err on the safe side.


If so the Zs in the table for 5 seconds would need reducing to something like 10/13 of the current value.


Another one to add to the growing list of potential gotchas.