Question on Gate Drivers Peaks Currents

Thread Starter

Xavier Pacheco Paulino

Joined Oct 21, 2015
728
Hi,

I have a gate driver which specifies a minimun peak current of 6A. I'm switching a Power MOSFET at a PWM frequency of 10 kHz, Vgs = 12V. The mosfet total gate charge is 160nC.

For the 12V, I'm using a DC/DC converter @1W/84mA max.
The math throws an average gate current of 2mA. So, is this the current that my converter has to supply, right? Can a 1uF capacitor handle the 6A peak current? As my converter is 84 mA max.
 

crutschow

Joined Mar 14, 2008
34,285
The average gate current is Q*f = 160nC * 10kHz = 1.6mA, rounded to 2mA as you stated, so yes, that's the average current the converter has to supply for that function.

The momentary voltage drop for a 1μF filter capacitor delivering a 16onC charge is V = Q/C = 160nC / 1μF = 160mV, so that should be adequate.
You do want to also put a 100nF ceramic capacitor directly between the driver's power pins and ground for good high frequency decoupling of the transient current.
 

ebp

Joined Feb 8, 2018
2,332
About 2 mA average is correct.

For a 6 A driver you might want more than 1 µF, but 1 µF charged to 12 V is 75 times as much charge as the FET gate, so it should be good. I strongly recommend an X7R or X5R ceramic. The slightly cheaper and smaller types with high dielectric constant (Zxx or Yxx) generally have really horrible negative voltage coefficient of capacitance. If you operate them anywhere near their rated voltage they can be down below 20% of their nominal capacitance. Even the XnR types suffer from this phenomenon, but to a much lesser extent. Using a 25 V cap instead of a 16 V helps.

Sometimes a capacitor that has very good high frequency characteristics can make quite a sharp resonance with inductance in the connections to the power supply and a fast-slewing load current can excite the resonance. If your power and ground to the driver are solid planes, this is unlikely. If they are traces that are side-by-side there will be moderate inductance. Adding a bulk capacitor that isn't particularly good at high frequency near the ceramic cap can dampen the resonance. Since you are using a DC-DC converter, if you place it very close to the driver the inductance should be very small and unlikely to be a problem. Still, if in doubt, I'd suggest provision for something like a 10 µF ordinary solid tantalum. If you need to make the DC-DC converter more remote in order to meet your layout objectives for safety, then I'd definitely recommend the tantalum near the driver. Some of the small DC-DC modules don't like too much capacitance on their output, so you can't get too carried away with adding bulk capacitance.
 

ebp

Joined Feb 8, 2018
2,332
I was assuming from the outset the 1 µF would be a good ceramic type. With the one you have chosen, I see no merit in using a 0.1 µF in parallel. That is/was done where the higher value cap also has "high" inductance due to size-related issues and you can improve the impedance at high frequency with the smaller cap in parallel. You are already planning to use a physically very compact cap where inductance will be very low.

I suspect crutschow may have been thinking the opposite way 'round - the 1 µF would be a tantalum or the like, and in that case a small ceramic in parallel would definitely be a good thing.
 
Top