IGBT transformer driver help

[PT_Dreamer]

Hi,

I made an IGBT gate drive transformer with a toroidal core with a permeability of about 8000H/m and the primary has 20 turns resulting in about 5mH. The unloaded driver circuit outputs a clean square signal with the proper dead time. However when the transformer is connected the signal shows the inflections shown in the picture.
I have monitored the current and there are no signs of core saturation.
Although the voltages are still ok to drive the IGBT gates I would like to understand what is going on.
Any ideas? Thanks!   

[xavier60]

It would be because of reactive current flowing back into  the driver's output for a while after each transition.
The skew at the dead time is undesirable but should not be a practical problem.
I use MOSFET for drivers rather than BJTs.

[PT_Dreamer]

Duh off course thank you! I was reading the plot as an abnormal deep in absolute voltage but it is just the opposite, it is an "abnormal growth in voltage" caused by the sum of the driver voltage and the inductive generated voltage.
Hadn't notice the skew until you mentioned.
Thanks again!

[Phoenix]

That waveform is expected because the driver diodes will be conducting the magnetizing current while it is opposite polarity to the voltage. So you need to add the voltage drop of the diodes to the voltage.

[PT_Dreamer]

That makes sense, strange none of the reference material I saw showed that effect.
Thank you!

[Arctic]

Hello guys,
I am trying to build a gate driver for IGBTs using ferrite toroid core but my output signal is like this.
You can see also the attached schematic.
I am trying many day using different materials but I dont see any difference.
Can you share the core you used?

Thanks!

[strawberry]

leakage inductance
bifilar windings

[T3sl4co1l]

Multiple things:
- TL494 needs pull-up (collector output) or pull-down (emitter output)
- 1X4340, what's that?  Oh, IX4340?  Just the one section of it?
- What NPNs?
- Bottom NPN blows up, gate driver delivers 5A into base to emitter to GND. If wired as labeled, the same happens but B-C (inverted mode operation).  This must be a PNP, emitter up, collector down.  Then you get a complementary emitter follower, which loses about a volt from the driver but can increase drive strength.  But you will need extremely beefy transistors (>>5A!) to improve on the 4340's drive strength.  And the voltage loss seems to only worsen operation into the transformer, I don't know why you'd want it.
- Transformer is DC coupled to ground, through some (220 ohm I guess?) resistance.  So it just kinda farts and does nothing, which is why the waveform is asymmetrical.  In a half-bridge configuration like this, you MUST have a coupling capacitor in series.  Some 10uF+ should do, I guess.  Depends on load current, inductance and required frequency range.
- Will this not be part of a half bridge, so that the TL494's complementary outputs shall be driven to each end of the transformer, and two inverse secondaries drive the finals?
- It's not obvious that the transformer has anything wrong with it but you will want to use twisted pair (or multiple pairs in parallel) to address leakage, as strawberry mentioned.  Note that twisted pair lengths going to the driver and gate, add in series with the twisted pair on the core, increasing leakage further.
- What type is the final MOSFET?  Does it need a gate resistor?  Will it be driven strong and fast enough as shown?  Will it see overshoot?  (These are all easily modeled as the RLC series resonant circuit, between driver, leakage + stray inductance, and gate capacitance.)

Tim