| Electronics > Projects, Designs, and Technical Stuff |
| short circuit protection of IGBT? |
| (1/7) > >> |
| coppercone2:
So I read if a IGBT has a short circuit current and even if it manages to turn off, you can still have failures. Assuming you are running something like 300A igbt's, is there some other component you can put ahead of the H-bridge to break the circuit fast enough without this problem? It would have to work within a microsecond or so of the detection. The idea would not be to turn the h bridge off but to cut off power to the H-bridge. I thought maybe some kind of SCR or whatever other weirdness there is that would not be appropriate for power conversion (it would not have to work frequently). And maybe something on the bus bar connected to ground that would short it out while its in the middle of turning off? |
| max_torque:
google "desaturation detection" |
| coppercone2:
Even the short-circuit current is successfully turned-off, short-circuit failure could still happen after several microseconds, which is called delayed failure in [31]. It is verified by experiments and numerical simulation that large leakage current leads to the thermal runaway [37]. A “critical energy (EC)” is proposed to explain the mechanisms of catastrophic failure and wear out failure under repetitive short- circuit operations [38]. When short-circuit energy is below EC, IGBT may survive for more than 10^4 times repetitive short-circuit operations before failed. However, when short-circuit energy is far beyond EC, IGBT may fail after first short-circuit due to thermal runaway thats with the detector I am pretty sure won't cutting power be better then trying to power it down while its over loading? thats what i think. I get the feeling messing with the gate is a bad idea if its really bogged down? it says something about 'local overheating' too. |
| jbb:
And what kind of switch would you use for removing power from the H Bridge? |
| PartialDischarge:
Forget it, the least expensive way to protect an igbt is to let the igbt break, that is besides designing a system with enough tolerances. If there was a switch fast enough to cut the supply it would cost 10x the igbt and add additional inductance to the whole setup, besides weight and size. So in power switching no one does that. |
| Navigation |
| Message Index |
| Next page |