how to limit mosfet current spikes

[ahmed osama]

dear all

how can i limit the mosfet current spikes & does it important to takecare of the currents spikes or its avg ??

 

[dick_freebird]

Some of this spiking may be due to using unrealistic simulated,
or overly stiff real gate drive. You have a lot of Qgg that has
to go somewhere. Either out the source, or out the load,
depending on the transition.

You should check datasheets for any pulse or transient
current limits in addition to DC ratings. RMS currents ought
to be no more than the DC, and my rule of thumb is peak
no more than 10X DC (unless you have reliability data).

Look at what happens with (say) a 10-ohm series gate
resistor added. Look at what a drain snubber or simply
representing true parasitics (R,L,C) in the load and the
power device do.

 

[FvM]

Unfortunately, you don't even tell which quantity is displayed in the waveform. How should we comment it?

A high current peak on turning on the FET (Id or Is) can be however expected, because BY249 is a slow rectifier with large reverse recovery time. It can never work at 300 kHz in a real circuit. As mentioned, you also shouldn't drive the MOSFET gate with an ideal voltage source. (Fortunately, no real gate driver is an ideal source...)

 

[sam185]

connect anti-parallel diodes configuration with mosfet.
regards,
Sam

 

[ahmed osama]

Look at what happens with (say) a 10-ohm series gate
resistor added. Look at what a drain snubber or simply
representing true parasitics (R,L,C) in the load and the
power device do.

indeed i tried Rgate from 100 ohm begin the spikes decreased alot it may reach to 40 A but the avg curr jumped to more than 10 A and avg Watt dissp by the mosfet also inc. alot
indeed datasheet says that the mosfet can bear til 140A in pulse

---------- Post added at 19:23 ---------- Previous post was at 18:25 ----------

because BY249 is a slow rectifier with large reverse recovery time. It can never work at 300 kHz in a real circuit.

i do not have alot of diode to choose (plus do not have alot of spice model alsoo) kindly check th following


---------- Post added at 20:58 ---------- Previous post was at 19:23 ----------

Unfortunately, you don't even tell which quantity is displayed in the waveform. How should we comment it?

A high current peak on turning on the FET (Id or Is) can be however expected, because BY249 is a slow rectifier with large reverse recovery time. It can never work at 300 kHz in a real circuit. As mentioned, you also shouldn't drive the MOSFET gate with an ideal voltage source. (Fortunately, no real gate driver is an ideal source...)

i have used 1N5817 instead of BY249 and things became much much more better thanks all

 

[erikl]

i have used 1N5817 instead of BY249 and things became much much more better thanks all

I'd think the 1N5817 is a bit sparse for this application, both w.r.t. voltage (Vr(max)=20V) and to current (Imax=1A).
Better use your BY329X, BY399, or SR360 diode.

 

[ahmed osama]

I'd think the 1N5817 is a bit sparse for this application, both w.r.t. voltage (Vr(max)=20V) and to current (Imax=1A).
Better use your BY329X, BY399, or SR360 diode.

i used more than 4 or 5 1N5817 in parallel to pass more than 1 A and simu. result is not bad

 

[erikl]

Don't put diodes in parallel!

i used more than 4 or 5 1N5817 in parallel to pass more than 1 A and simu. result is not bad

Not exactly the best solution, s. **broken link removed** Depends on how long it should work.

It works well in simulation, as there all diodes are equal. But never so in real world.