Driver for high power IGBT

[Jagajack]

Hi All,

I am using a SEMIKRON IGBT for my 5kW H-bridge inverter.
Which driver is best for 1200V/100A IGBT gate?
1.Pulse transformer
2.Optocoupler

Thanks.

 

[mrinalmani]

Best depends on timing requirements, which you have not mentioned.
Simplest would be a driver IC. Transformer may require en extra active pulse build-up circuit. Between opto-coupler and transformer, opto may be easier to implement practically.

 

[Jagajack]

Thanks for your response.
My inverter operating frequency is 10kHz with 50% duty ratio. But how to drive a hide side switch using optocoupler?
Is totem-pole a better solution. If so how it could be implemented?

 

[mrinalmani]

There are opto-couplers available with totem pole output, you may use them. Otherwise an optocoupler with transistor output can be used drive a totem-pole that will in turn drive the high side MOSFET.

 

[Jagajack]

Could 4 optocouplers with individual isolated supply be used to to drive my H bridge?

 

[FvM]

I presume that the IGBT datasheet suggests gate drive parameters (voltage, gate resistance). What is it? I guess for the a 100 A IGBT, bipolar gate voltage drive, e.g. +/- 15 V or +15/-8V is suggested. To follow the suggestion, you'll use a gate driver circuit that is capable of bipolar drive. Another important point is actual bus voltage in your application.

 

[D.A.(Tony)Stewart]

Driving high power IGBT's is not like driving a Relay. There are different design rules for 3rd and 4th generation IGBT's where Rg affects speed and power dissipation due to high gate load charges during switching. The best driver is a faster low impedance driver during switching time to minimize transient power loss during switching.

So a driver capable of 100A might need a gate pulse current of many Amps which is not viable with your suggestions.

**broken link removed**

 

[Jagajack]

I presume that the IGBT datasheet suggests gate drive parameters (voltage, gate resistance). What is it? I guess for the a 100 A IGBT, bipolar gate voltage drive, e.g. +/- 15 V or +15/-8V is suggested. To follow the suggestion, you'll use a gate driver circuit that is capable of bipolar drive. Another important point is actual bus voltage in your application.

The IGBT I am using is SKM100GB12T4.
The gate resistance Ron=5 ohm and Roff=10 ohm.
The DC bus voltage is 600 V.
Why bipolar gate voltage is recommended for IGBT with gate threshold voltage >5V?
0 to +15 V gate signal could be able to switch the IGBT as desired. Correct me if I am wrong.

- - - Updated - - -

Driving high power IGBT's is not like driving a Relay. There are different design rules for 3rd and 4th generation IGBT's where Rg affects speed and power dissipation due to high gate load charges during switching. The best driver is a faster low impedance driver during switching time to minimize transient power loss during switching.

So a driver capable of 100A might need a gate pulse current of many Amps which is not viable with your suggestions.

**broken link removed**

But these drivers are very costlier.

 

[D.A.(Tony)Stewart]

The SKM100GB12T4 has an internal RgINT=7.5 Ohm, adding external 1 Ohm increases slew rate and internal energy per pulse with rising gate resistance.

Gate capacitance vs Gate Voltage is measured in mJ from 1/2CV^2*t for energy required the reverse charge level.

Since module is rated for +/-20V Vg abs. max. and specified operation is +/-15V there may be some benefit to using an avalanche effect driver starting at -15 vs -8V vs 0V in the energy required to switch the gate faster from the region of -8 to +15V is my guess. Avalanche effect drivers are common and do not follow simple linear RC characteristics as driver impedance is not linear. Think of it more as a switched capacitor discharge driver.

The fastest low (ESR & ESL) impedance driver <=1 Ohm is desired for low dissipation in power module but also increases EMI radiation di/dt. So you can now see how switching 100A on module requires switching many Amps on Gate in short period.