How to eliminate hard switching effects?

[seyyah]

hard switching snubber -resonant -soft

How can we eliminate dips or peaks which occur at rising or falling edges of the switching pulses?

 

[vdaniel]

The question is too common. There are a lots and lots of different kinds of switching pulses.

Varuzhan

 

[IanP]

Without going into details, you can try to use a snubber circuit.
Have a look in here:
**broken link removed**
Regards,
IanP

 

[seyyah]

Switchings are pwm switchings up to 20KHz. Actually i use a snubber. May be it needs tuning. I need some info to calculate parameters, can you offer something? Also someone proposed me to connect an rc filter to the gate , but i couldn't understand the use of this. Does this method make any sense to you?

 

[VVV]

Connecting a filter at the gate will slow down the MOSFETs. With slower rise/fall times the induced voltages (in parasitic inductances) will be lower.

The down side is that you will increase switching losses in the transistors. Generally, only a resistor in series with the MOSFET gate is sufficient. I would not add the capacitor.

 

[seyyah]

A zener diode may suffice may be, what do you think?

 

[power-twq]

you can use RC snubber circuit to have a try.

How can we eliminate dips or peaks which occur at rising or falling edges of the switching pulses?

 

[Ice-Tea]

Maybe this is a stupid suggestion, depending on your experience level, but have you checked your measuring method? Per example, measuring with a scope with long ground leads will also result in perceived ringing on edges?

 

[yongchai]

If you mean the peak/dip at rising/falling edge of 'switching signal' (PWM signal to gate), then you should add a resistor in series to the gate of the mosfet. If I'm not wrong, larger resistance value will increase the rising time, and therefore reduce the peak magnitude. However, switching loss will increase. So, need to set a balance.

 

[seyyah]

R should eliminate but didn't help so much. I can't increase the resistor very much. Snubber may help, thanks

 

[sivakumar_tumma]

use snubbers...

 

[vaibhav.k]

Best snubber circuit is RCD snubber with polyster type cap which reduces almost 60 to 80% spikes

 

[Kurniawan]

first, you must to ensure that your pulse gate of switcher can truly on and off. The second, Your PS must be in continuous conduction mode. Circuit of snubber maybe complex and difficult, if not work in big power maybe u can select device over your design ( current and voltage capability ).

 

[drabos]

I have the same problem, but my device is in integrated circuit. and the inductance is the bonding wire. The breakdown voltage is very little. I can't decrease the rise time. And I can't connect anything with parallel to the bonding wire.
Any idea?