Boost converter output voltage overhoot during igbt switching on

[panoskak]

Hi guys,

I face a problem with a boost converter.

When my igbt is switched on, then I observe an overshoot at the output voltage. I put a snubber
capacitor, but this is not about it (this is actually useful at the switch off state). I googled it, and
found some specs from commercial products characteristic curves with the same "problem". This
overshoot phenomenon duration is of about 10us, which is comparable with my switching frequency.

Describing better my situation:
From the moment that the pulse is on, an overshoot is observed at the output voltage like oscillation,
when this finishes, the igbt seems to switch on and the inductor current increases. After that everything
is ok. When the output voltage is 30V, the peak value of the overshoot is of about 33-35V. At higher
voltages this is more intense.

My diode is hyperfast (some 20-30ns) and my igbt has been tested under other kind of circuits and
it seems to be everything ok, until all these have to operate together.
(test case: L = 1.2mH - 2mH, Cout = 470uF, Rload = 13.6Ohm, Iin_average = 2-3 A, Vin_average = 20V)

A larger capacitor may solve my problem? Has anyone else faced the same thing? You know the answer
because you solved also my previous query with my igbts...

Thank you in advance

 

[FvM]

You can easily calculate (or check in a simulation), that the basic boost converter topology with 1.2 mH inductor and 470 uF capacitor is unable to produce the observed oscillation. This clarifies, that you are talking about parasitic effects of your circuit layout and the involved components, particularly capacitor ESR and ESL, inductor parallel capacitance and ground inductance. Also measurement artefacts e.g. caused by unsuitable probing can play a role.

In other words, it's all about power electronics "craftmanship". You can hardly discuss the problem without looking at the circuit details, including the geometry. Schematics and photos are needed.

 

[panoskak]

You are absolute right. I changed everything and this continues to exist. Inductance, capacitor, load, power supply, but not the switch, temporarily. The igbt turn on and off times, combined with the pulse tr and tf may play a role (who knows...). After that, I will try to change the geometry and I will post again my findings.

Thanks for your immediate reply...

 

[panoskak]

Nothing changed from previous post.

This is a photo from the boost converter, from left to right . Capacitors and inductors are in series. I tried different combinations. The switch is a leg, and I use only the one igbt. I have not connected the point of connection of the two capacitors to earth.

 

[goldsmith]

Dear panoskak
Hi
The value of your snubber isn't proper . what are the values ? just a simple capacitor ? it is not enough . you should use a RC network (series RC , in parallel with your IGBT ) . do you know how is the calculations of an snubber network ? however it is not necessary to improve those overshoots for this purpose .
Best Wishes
Goldsmith

 

[panoskak]

The value of the snubber capacitor is about 0.1 - 10uF, but I do not remember exactly, this value is proposed by the manufacturer without the presence of a resistance. As far as I know, the importance of the snubber capacitor is for the switching off, however the overshoot appears on the switching on. I do not know how to configure a RC network, but due to the fact that the manufacturer proposes just this type of capacitor, I believe that this is not necessary. To be honest, the manufacturer proposes a snubber capacitor to the dc link, for the use of the entire leg. I put the capacitor, and added a wire to connect it to the middle, for the one igbt. In the attached image you can NOT see the snubber, because I added this later. Imagine a capacitor (with low stray inductance) screwed on the two upper holes and a wire between the hole with the black wire and the right upper hole.

When you say "however it is not necessary to improve those overshoots for this purpose" what do you mean. The overshoot duration is about some micro-secs, close enough to my control strategy.

I have to add here that when I reduced the size of the inductor (from 4 to 2 inductors), then the overshoot and the ringing effect was more intense. Additionally, when I added a higher value of inductance then the overshoot appeared again but everything was close to the case of the 4 inductors.

 

[FvM]

I reviewed the photo. Generally speaking, it's not surprizing to get overshoots with this design due to the huge wiring inductances. When probing the output voltage directly at the output capacitor with short ground leads, there shouldn't be much overshoot.

 

[panoskak]

I hope you to be right. I will try to rearrange the design, and if it is possible to use pcb, with all the connections with small cables, or forming welding lines in the dotted board. Thank you again...

 

[panoskak]

After rearranging everything in a more compact design, the same behavior from my igbt. During switching on an overshoot appears. In order to solve this, an other igbt used, from other manufacturer, and the result is an overshoot during switching off, which is expected. Then using the proper values of a RC snubber network and everything seems to work now as supposed. However I did not find the reason for the behavior of the first igbt. The only strange now, and before, is the quite abrupt change in the output capacitor voltage from on to off, and viceversa, however without overshoot.