Overcurrent rating selection

[Arbit1]

Hi,
I was designing a dc-dc converter and was selecting a fuse for overcurrent protection on primary side, saw various parameters for its selection, but had one basic doubt, how to determine the current limit at which the fuse should trip
The continuous primary current rating is 4.9A

 

[KlausST]

Hi,

DCDC converters are semiconductor devices. Semiconductors may react rather sensitive on overcurrent. Every device datasheet should tell the SOA usually with current and timing.
In many cases a fuse is too slow to protect semiconductor devices, because they need very high current for fast blow.

And DCDC converters usually work with inductances. Inductors may saturate very hard, means the curren rise gets very fast. Maybe too fast for a fuse to blow. And it works with high switching frequency. Means the current in the semiconductor get's switched of repeatedly. If a "single overcurrent" event is too short to blow the fuse, the current will get switched off, the fuse has time to relax (cool down)...so the I^2t integral for the fuse to blow will take several pulses...and a lot of time..

The other thing is: Often DCDC converters have a built in cycle by cycle current limit. Maybe limiting the primary current peaks to twice the expectable peak. In an secondary_short_circuit condition this may cause the semiconductors and inductors to be stressed, but the primary fuse will see low RMS current. The duty cycle becomes very low, thus the fuse will never blow.

So in my eyes a good protection is:
* peak by peak current limit (this limits the power dissipation, the device may overheat, but it will take several 10s at least)
* aditional overtemperature shut down

If properly designed it will reliably protect the device
* from high load current (with normal output voltage) to short circuit condition (with low output voltage)
* for short time and long time and
* for low to high ambient temperature.

Klaus

 

[Arbit1]

Hi,

DCDC converters are semiconductor devices. Semiconductors may react rather sensitive on overcurrent. Every device datasheet should tell the SOA usually with current and timing.
In many cases a fuse is too slow to protect semiconductor devices, because they need very high current for fast blow.

And DCDC converters usually work with inductances. Inductors may saturate very hard, means the curren rise gets very fast. Maybe too fast for a fuse to blow. And it works with high switching frequency. Means the current in the semiconductor get's switched of repeatedly. If a "single overcurrent" event is too short to blow the fuse, the current will get switched off, the fuse has time to relax (cool down)...so the I^2t integral for the fuse to blow will take several pulses...and a lot of time..

The other thing is: Often DCDC converters have a built in cycle by cycle current limit. Maybe limiting the primary current peaks to twice the expectable peak. In an secondary_short_circuit condition this may cause the semiconductors and inductors to be stressed, but the primary fuse will see low RMS current. The duty cycle becomes very low, thus the fuse will never blow.

So in my eyes a good protection is:
* peak by peak current limit (this limits the power dissipation, the device may overheat, but it will take several 10s at least)
* aditional overtemperature shut down

If properly designed it will reliably protect the device
* from high load current (with normal output voltage) to short circuit condition (with low output voltage)
* for short time and long time and
* for low to high ambient temperature.

Klaus

Thank you so much Klaus this really helped