Boost converter MOSFETs

Question

I built a boost converter that converts 42 V to 57 V.

While it works well with a FDH055N15A MOSFET at 20 A, I am trying to understand why it won't work properly with a HY3912W MOSFET. I only have two FDH055N15A MOSFETs left and a whole box full of HY3912W's I would like to be able to make use of.

NOTES

Inductor L22 in the schematic and pictures is bridged with coper wire of zero resistance or inductance.

I wound the inductors myself and haven't actually tested the inductance or resistance but thry works fine with thr FDH055N15A:

I am guessing that the on resistance is too high as the base pin is only driven to 12 V by the TC4420 driver that is also powered from the same rail as the fan.

I am struggling to understand from the datasheets what the difference would be if the 12V rail was increased to 18 V or 20V. I am thinking of trying to power the rail at 20 V and swapping the fan out for a 24 V fan hoping it will run at 20 V. If that is a success, re-print the boards splitting the 12 V rail into a 12 V and a 20 V rail. I rather over purchasing £7 each MOSFETS for these boards I already have.

FDH055N15A Datasheet

HY3912W Datasheet

The datasheet of the FDH055N15A shows an on resistance at 10 V or 20 V:

The datasheet for the HY3912W shows an on resistance at 10 V only:

I am not sure either will help me understand why the HY3912W runs at 20% duty cycle producing as much heat in the heatsink as actual output current where the FDH055N15A will run happily at about 12% duty cycle producing 20 A.

Am I barking up the wrong tree and just being thick? To be honest, I don't really know what I am doing, my whole life is generally just trial and error.

Answer 1

I believe you will find that the problem is caused by the stray inductance in your gate drive trace. The two parts have dramatically different reverse transfer capacitances.

This is the capacitance between the drain and the gate. When the drain is high, and you attempt to drive the gate high, the drain voltage decreases with a high dv/dt. The reverse transfer capacitance couples this change in voltage to the gate, which tends to drive it low, turning the FET back off. When this occurs, the drain voltage can go back "high" and the FET will turn back on. So when you scope your circuit, you will see an oscillation on both the gate and the drain. The opposite happens when you turn the FET "off." This oscillation will occur on each edge.

This "back drive" will be markedly less with the 23 pF of reverse transfer capacitance of the FDH055N15A when compared to the 465 pf of the HY3912W (20 times higher). Your only solution to use the HY3912W is to drive this capacitance by supplying high current to the gate. Many times board designers take great care to provide short, wide current paths for drain-source current, but fail to do so for the gate drive, whose instantaneous current is quite high.

Good luck!