MOSFET short circuit withstand time

[Techman_7]

How to determine short circuit withstand time of SiC MOSFETs?

 

[KlausST]

Hi,

look at the datasheet. Usually it has some informations on that. Like an SOA diagram...

For more details I recommend to give link to a datasheet.

Klaus

 

[FvM]

Depending on your short circuit condition, supply and gate voltage, the withstand time can be rather short, not necessarily covered by SOA data.

 

[Techman_7]

Or here do we have to limit the junction temeprature by limiting the total MOSFET losses.
This will tell us the max or short circuit current through MOSFET.

Whats your thought?

--- Updated Aug 26, 2022 ---

Hi,

look at the datasheet. Usually it has some informations on that. Like an SOA diagram...

For more details I recommend to give link to a datasheet.

Klaus

Datasheet link -

https://www.onsemi.com/pdf/datasheet/nvh4l020n120sc1-d.pdf

 

[KlausST]

Hi,

what´s your relaistice short circuit condition?
* max voltage
* turn off timing
* loop resistance
* do you have limited current?

Klaus

 

[cupoftea]

Summary is that FETs and Sic FETs are generally no good for short circuit current....they generally cant turn off a short cct current...if you have shorts, then its IGBTs you need.
But if the s/c currnt is within the rating of your fet, then it would be ok.
Are you doing a HV BLDC drive?

 

[Warpspeed]

Or here do we have to limit the junction temeprature by limiting the total MOSFET losses.
This will tell us the max or short circuit current through MOSFET.

Junction temperature rise is the key of course.
Unfortunately power is proportional to current squared, and it all becomes rather explosive very darned quickly.
Unless there is a totally unrealistic amount of series inductance, nothing is going to prevent the inevitable disaster.

 

[Techman_7]

Hi,

what´s your relaistice short circuit condition?
* max voltage
* turn off timing
* loop resistance
* do you have limited current?

Klaus

Realistic short ckt condition = 46A rms; 71A peak;
Max voltage = 800V
Turn off time = ~ 80 ns
Loop resistance = ~ 90mohm
Limited current = 37A rms; 57A peak

--- Updated Aug 28, 2022 ---

Summary is that FETs and Sic FETs are generally no good for short circuit current....they generally cant turn off a short cct current...if you have shorts, then its IGBTs you need.
But if the s/c currnt is within the rating of your fet, then it would be ok.
Are you doing a HV BLDC drive?

Ok. Not using BLDC drive.
Its DCDC converter of 10kW rating.

 

[KlausST]

Hi,

interesting definition of short circuit current in RMS.
Talking in RMS and peak currents ... this sounds as if your voltage (power) source is AC.
AC in the meaning of that the short circuit condition lasts at least one half wave of time.
What is your AC frequency?

What I expect: If the voltage is 800V and the loop resistance is 90 mOhms, then the current goes up to about 8900A.

Also the 80ns of turn off time .. is rather short for my taste. But surely possible and good if true.

May we see the schematic you use to detect the short circiuit and the signal path to turn off the transistor.


Maybe it´s a good way to get ... or to verify your values with a simulation tool .. that just simulates the short circuit condition.

****
I have another question. Do you expect the short circuit to happen any time? Do you have a scenario how the short circuit happens and how it builds up?
What I mean: when you have a device where all the connections are fixed and not accessble, then a short circuit is expected to build up much slower than 80 ns. Thus you have much more time to detect and react.

A rather fast short circuit scenario is when one throws with high speed a metal piece between the 800V.
If you put the metal piece with slow speed you get a scenario like with electric welding where metal gets liquid and even vapourized by the immedaitely existing heat. Then it´s likely to get an arc than a solid (90mOhms) short circuit condition.


Klaus

 

[Easy peasy]

Some SiC mosfets have a quoted 3/5uS short ckt withstand time

But no Si mosfets I have ever seen have any sort of similar rating - especially more modern die shrunk cellular devices with very low Ron, this is because - even if you could sense over current - say 100-200A and start to turn off the fet - the resulting over-volts will kill it - unless you have serious zeners or similar to protect it.

The mechanics of modulating the gate drive to prevent this and give a softer turn off with less overshoot are well into the RF design arena - skills which very few power electronics engineers possess - and then you need the current sense element very close to the fet and it's gate drive circuitry - or as is more commonly attempted - use the Von to estimate overcurrent.

There has never been a successful desat protection for mosfets in the 600V - 1200V ~ 50A bracket that I have ever seen - although plenty of people have tried.

If you have a large inductor limiting the the di/dt in the device - more normal circuits can be used to turn the device off for overcurrents - in the event of a control failure ....