Hi everybody
I have a full-bridge inverter with IGBT and at 5KW.
One of the gate-emitter of IGBTs is short-circuited when I turn driver off. But when I turn driver on, it's not short-circuited but takes a lot of current so that the amplitude of driver signal decreased.
The driver current decrease with signal frequency decrease (about from 20 KHz to 12 KHz).
My IGBT is SKM100GB12T4.
Why this happen?
i think you need to put resistance about 27kohm between gate and emitter
I think A potential divider of two resistances one between gate switching signal and Gate terminal, while other between gate terminal and emitter terminal should be placed. This can compensate the turn-off DELAY time of IGBT. and also driver current will decrease with decrease in signal frequency
Thanks to all three Mohammad (@Muhammad Waseem, @Mohammed Bashar Sedeeq, @Mohamed Azab )
This is the used schematic:
https://www.loneoceans.com/labs/sales/brickdriver/SchematicBrickDriver11.png
But I remove R2 (27R) because it tooks current from optocoupler and decrease negative voltage.
I would try to exchange the module that shows
this strange behavior with the other identical module
of the other branch and I would verify how the new component
works under the same conditions. If you have the same behavior
you need to change the electronic circuit for turning the IGBT
on and off, if it works well you need to replace the module.
Fabrizio Castaldi thank you for responding.
The another IGBT doesn't show same behavior. I change it but my main problem is why this happen??
Because Gate-Emitter are damaged if the Voltage between them increases by more than 20 V. But I used TVS in the driver circuit.
Actually the behavior of G-E is combination of inductive and capacitive. At dc voltage is shorted (such as an inductor) but at ac voltage has a impedance but it increases with frequency decreasing (such as a capacitor).
Hi Alireza,
Can you explain a little more about how you figured that the device was damaged? Was there any fault? If so, when and how did that happened?
Also, have you checked to see if the device source-emitter still has the voltage blocking capability (i.e. when the gate is off, is source-emitter open? )
Hi
I think you can not separate neutral between upper switch and lower switch.
You can use different different neutral wire for all upper switch and one common neutral wire for lower switch.
But you can carefull about the separation of neutral of all upper switch (used IR2110 it gives the separate neutral for upper switch)
thank you
Also, you can connect an external antiparallel diode in between the emitter and collector. Also connect a 30k ohm resistor in between the gate and emitter.
thank you
I think that the IGBT is already defected before it is stressed. It can happen that some transistors are not well tested and approved before they are dispatched from the factory. When you turn the transistor on it withdraws an excessive current pointing out that such junction is defected and with stressing it will get completely defected such that in the off state it remains short.
Best wishes
Because this state is occurring in reduced the frequency that's having the relationship between voltage and frequency therefore the current reduced.
Finally, I found the cause of it but haven't found why shows this behaviour yet!!!
I used the HCPL316J for driving. This component has desaturation protection and ceases the signal whenever such a situation occurs. I used it for a high Q resonance inverter and have a lot of Voltage!!!
On the other hand, the gate resistor was too small and it runs a lot of currents and switching was very fast. Whenever I had a desaturation situation the OptoCopupler lagging behind and a lot of current feed-in IGBT at the linear area and that was what should not have been!!!!
I increased the gate resistance then there was no IGBT failure and Opto protection works well.
I chose the gate resistances near the values which are calculated from Qg/(tr+ton) and Qg/(tf+toff).
I am glad for you that I have found a way out of the situation. Of course, the use of an optocoupler greatly simplifies the task, thanks to the isolation and flexibility of IGBT control. I think one of the reliable options was to eliminate through currents in IGBTs. And for this, naturally, it was necessary to provide delays between the moments of opening and closing of IGBT transistors located in one arm.
Best regards, Islam.
İslam Asad İsgandarov Thank you for responding.
I've used the dead time for signal separation and making safe and low losses switching.
But if you haven't enough dead time, in the worst-case collector-emitter will short circuit, not gate-emitter.
Yes, I agree with you!
Given my experience and generalizing the opinion of many scientists and specialists involved in this field, I want to note that it is not so easy to engage in practical power electronics.
The practical implementation of such devices requires more experimentation and precaution.
I wish you success in such hard work!
Dear Alireza Tanehkar ,
Thanks to your Tech User Instruction regarding for gate emitter a lot of current consumption .I request you to aware us contineously more electrical / electronic
instrumentation type of instrument regarding user specifications .
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping
- Maps