Gate resistor bypass diode on KaizerSSTC

[Zipdox]

So I was browsing loneoceans.com and noticed that some schematics for SSTC bridges had diodes across the gate resistor to make the gate discharge faster.


I've built a variant of the KaizerSSTC3 and I was wondering if I could add these diodes, (1N5819) to decrease switching times. I'm worried that while this would decrease switching-off time it will increase switching-on time because of the load it puts on the gate driver. This would in turn cause a decrease in efficiency in the MOSFETs and increase heat production. Is this correct?

Can someone shed some light on my theory and perhaps give me a recommendation?

[davekni]

When using a GDT to drive symmetric gate waveforms, I use diodes for IGBTs only.  The diode speeds up turn-off, compensating for the IGBT's typically slower turn-off time compared to turn-on time.  FETs are generally fast enough to avoid any cross-conduction without speeding up gate turn-off.

When driving MOSFETs directly with 0-12V signals (rather than +-12V from a GDT), then diodes are usually necessary.

The one case where I can picture diodes being useful with GDT-driven MOSFETs is to further increase dead-time between one FET turning off and the other turning on.  This could help if the natural dead-time due to Vgs threshold isn't enough for the load current to induce the Vds transition.  A bit more power is wasted if the opposite FET turns on before its Vds is zero.

[Zipdox]

So you're saying it won't have much of an effect?

[davekni]

My last post is a wordy way of saying that circuit details will determine optimum gate drive damping resistors and diodes.  Can't say anything definitive without scope traces.  If most of the bridge output voltage step (Vds step) is complete before the opposite FET's Vgs gets above threshold voltage, then diodes won't reduce losses.  (If there's excessive Vgs overshoot, enough to stress the gates, then more series gate resistance is needed for damping.  Need for diodes should be reevaluated after damping is increased.)

[Mads Barnkob]

Dave got on to all the points.

I seem to recall that the main reason is to avoid shoot-through and its properly there for historic reasons or just a extra precaution just-in-case. From the historic view the development of SSTCs, SISGs, BRISG, OLTC and other topologies before SSTC/DRSSTC was refined into what they are today, silicon was even slower and more expensive.