Hello Samo Smodiš

The SiC diode has a much shorter turn off time. So there is much less energy absorbed in the switching edge. The energy stored in the inductance and the parasitic capacities however doesn't disappear. In stead of getting dissipated in the diode the energy simply swapps between a voltage in parasitic capacities (when the ringing current crosses the 0A) and the current flowing in your wires to the diode (current maximum, voltage over parasitic capacities crosses 0V)

Options you have:

1. Check if the ringing harms your application at all.

2. Minimize parasitc inductances (wires etc.)

3. Minimize the stray capacities of the coil

4. Use a diode with lower junction capacity (this simply shifts up the ringing frequency. The number of periods to decay however remains similar)

5. Use a smaller switching transistor to reduce the energy stored in the Cds. (Leads to higher static losses!)

6. Add an RC snubber (then the snubber resistor dissipates what the silicon diode dissipated before. Regarding efficiency you loose what you gained adding the SiC diode)

7. Consider a change of concept to zero crossing switching. This takes you to a resonant mode design.

8. Are you really sure the ringing didn't exist using the silicon diode? (Could be the turn off behavior was simply modeled a little bit more beautiful or the wire inductance of the diode was not part of the model)

Ciao, Ricardo

Dear Samo Smodiš ,

Most probably, when the diode is in reverse bias then the abrupt changes on the voltage at the diode cause by the voltage steps will ring in the diode and coil because the diode has reverse capacitance and bulk resistance Rs. It is o that the bulk resistance is appreciably larger in the SiC diode than the Si diode. This causes that ringing is sustained for longer time in the SiC diode.

Best wishes

Dear Abdelhalim abdelnaby Zekry

I think the key point is the more abrupt turn off of the diode. After the turn off the diode simply acts as a capacity.

If the bulk resistance would be higher the ringing would even be damped better.

I also noticed a change in the modeling style of most manufacturers. Old silicon components often didn't include the parasitic inductances of the package because they didn't matter much at the speed of silicon power diodes. Modern SiC diode models often include the parasitic inductance because there the switching speeds are higher and the inductance has a higher impact. As a consequence simulation results sometimes are not really comparable due to the change of model style.

Comparing the SPICE netlists of both diodes would help to understand what happens.

But since a model always is a simplification of the truth I recommend measuring both designs. (hoping they both use the same layout and layout parasitics are equal)

With best regards, Ricardo Erckert

Yes I agree completely with you that one needs more thorough and comprehensive investigations including the parasitic associated with the encapsulation of the device.

I also greet you for your very constructive view for the ringing problem here.

I tried really to compare the parameters of the two diodes when reverse biased and I found that the major difference is the higher diode resistance of the SIC diode and the capacitance of the SiC diode is slightly smaller than than that of Si diode. So I assumed that this may be the cause of the damped oscillations observed. As the parallel resistance increases the damping will be less and the oscillations may last longer. This was my thinking to the problem.

I was not also sure that this may not be major cause of extebded ringing, so I said most probable!

Hello Abdelhalim abdelnaby Zekry , hello Samo Smodiš

What was the turn off time of the original Si diode compared to the ringing period? The slower the rectifier the lower the energy content in the high frequency range that can excite ringing.

If the rise and fall times of the current in the diode are about the the same as the ringing period the energy available to excite ringing has already approximately decayed to half compared to an abrupt diode turn off.

On the other hand why should we care about ringing as long as it doesn't trigger anything critical (such as emergency turn off or hot carrier aging in the switches). It is simply energy swapping between the reactances.

Some people may complain about EMC. For those there are two answers:

- we can add a snubber or make the diode slow and kill efficiency (bad idea but cheap)

- we can add EMC filters at the supply and the output (costs money but efficiency stays high)

My usual approach in such cases if I can't improve the layout anymore: Give both options a price tag and let the customer decide. (The most important parameter in a spec for volume production always is rated in Dollar!)

Wish you success, Ricardo

Oscillations occur at the resonant frequency of the circuit formed by the inductance and parasitic capacitance of a closed diode. This phenomenon is discussed in the book "Investigation of switching power supply circuits in SwCAD / LTspice"

https://www.twirpx.com/file/3235308/ (Russian)