A while back I had a similar problem and it turned out to be ringing due to trace inductance would drive the VS pin so far negative that it would destroy the high side driver chip. Here is an app note that explains it and the fix. See section 2.4, 4.4 and figure 18.

I've read the App Note you give
I saw that in fig.18 there is a schotky diode add between Vs and Gnd, and Rgate between Vs and Vout. Did you use it to prevent negative spike, and is it works?
The explanation in app note is too short for me to understand. I'm a bit confused for the function of those component. Sorry

 

Yes, it worked. I have not had a failure since I added it. Here is a picture of the schematic. The driver chip is an LM5104.



Note: D2 is located right next to U4.

 

The MPU you are using tri-states the I/O ports during reset which leaves the inputs to your gates floating. The design does need pull-downs to address this issue. If the gate inputs float high, the H-bridge will be destroyed.

 

@HW-nut
I am not familiar with the microcontroller the OP is using. Does it have a PWM module? If so, are there dead time provisions?

 

Yes, it worked. I have not had a failure since I added it. Here is a picture of the schematic. The driver chip is an LM5104.

View attachment 107156

Note: D2 is located right next to U4.

I have bought a lot of IR2110, is it OK not to change IR2110 with LM5104? I dont wanna waste a lot of money lol
How can you calculate the value of R between Vs & source and choose the right schotky diode for the circuit?
is there any app note which I can read? because I never use schotky diode before

Thank you

 

The MPU you are using tri-states the I/O ports during reset which leaves the inputs to your gates floating. The design does need pull-downs to address this issue. If the gate inputs float high, the H-bridge will be destroyed.

OK. Thank you for your advice

 

@iyunxpenxhunx
Stop, the LM5104 is not a replacement for the IR2110. The LM5104 is a 100Volt part and the IR2110 is a 500 Volt part.

..............is there any app note which I can read?............

Yes, it is in post #6.

The LM5104 creates the "dead time" necessary to safely drive the Mosfets. The IR2110 does not create "dead time". You need to create dead time in your microcontroller or you need to find a high voltage driver chip that does.

If you do not understand what I have said or the app note in post #6, then you need assistance on your end, to make this work.

 

@iyunxpenxhunx
Stop, the LM5104 is not a replacement for the IR2110. The LM5104 is a 100Volt part and the IR2110 is a 500 Volt part.

But can I apply that circuit to IR2110, especially for R Vs&source and schottky diode?

I'm sorry if my english bad and a little bit misunderstood

 

But can I apply that circuit to IR2110,..............

Yes, absolutely.

 

Yes, absolutely.

one more thing, how to choose the right schotky diode for that circuit? what kind of parameter which I need to see?

 

I'll report if I have any update with my circuit after some revision

Thank you for all your advise

 

one more thing, how to choose the right schotky diode for that circuit? what kind of parameter which I need to see?

It needs to be a fast switching diode that is rated for the voltage. You can use a fast conventional diode to get the voltage up there. Current is very small. The one I use has a 1 amp rating. You could use a diode with a lower current rating.

 

Make sure you add the pull-down resistors to the MPU output so the gates aren’t floating during reset.

 

It needs to be a fast switching diode that is rated for the voltage. You can use a fast conventional diode to get the voltage up there. Current is very small. The one I use has a 1 amp rating. You could use a diode with a lower current rating.

Make sure you add the pull-down resistors to the MPU output so the gates aren’t floating during reset.

OK thank you very much
I'll update the report as soon as possible

 

Thank you everybody, problem has been solved.
Thank you very much