Mosfet max drain current rating

[OM222O]

Hi
I was wondering if the mosfet max drain current rating is like a limit or rather something to watch out for. I want to build a charge pump and use totem-pole fets for charging / discharging the capacitors and I'm not sure if that can damage the fets, since the current is only limited by RDSon and ESR of the capacitor which is usually way too small to do any sort of "limiting". I know in cases like capacitve discharge welders, multiple fets are used in parallel but I think in those applications the overall capacity and current are much much larger than a charge pump.

If the max drain current does not limit the current, what's the best way of achieving this? adding resistors in series with the capacitors?

[Zero999]

Presumably you're talking about the surge current, when the power is first turned on? The maximum drain current is irrelevant. What's more important is whether the MOSFETs can withstand the large power surge or not. Take a look at the surge and safe operating area information on the data sheet.

[OM222O]

it's not a surge current, rather a repeated pulse which charges / discharges the capacitors every cycle. I'm not sure which ratings I should consider. in lt spice simulations it seems the mosfet kind of limits the current to 2 or 3 amps peak, but just using a resistor with the same RDSon value yields upwards of 100A so I'm not sure what the limiting factor in the fet is or if the simulation is wrong.

[magic]

Safe operating area, transient thermal impedance.

[Zero999]

it's not a surge current, rather a repeated pulse which charges / discharges the capacitors every cycle.

But those repeated pulses are small, compared to the turn on transient.

I'm not sure which ratings I should consider. in lt spice simulations it seems the mosfet kind of limits the current to 2 or 3 amps peak, but just using a resistor with the same RDSon value yields upwards of 100A so I'm not sure what the limiting factor in the fet is or if the simulation is wrong.

Resistors just waste power.

What's the load current? It will have to be very high to get 100A pulses. Please post some schematics.

[David Hess]

The absolute maximum drain current is specified in the datasheet and is typically twice the maximum continuous drain current.  It is most easily controlled with a resistance in series with the gate.

Note that in a charge pumping application, loss is independent of switching speed.  The MOSFET will dissipate the same amount of power per switching event whether switching is fast or slow.

[OM222O]

I updated ltspice and ran the simulations again, this time I added a third model with a 1K gate resistor. The first interesting thing is that this time around, the current is not limited to just 2 to 3A when using the mosfet  it's actually more than the single 50m\$\Omega\$ resistor (similar value to RDSon of the fet). adding a 1K gate resistor seems to limit the current slightly, but not by much. a 10K one will reduce it to a peak of about 170A but I'm not sure if that's safe for operation. reducing the capacitor size helps but that also reduces the output current capability of the charge pump at the end.



I don't want to build the circuit, only to have it blow a dozen fets because I wasn't careful with the design, but here is the circuit that I'm planning to build (with a -12V LDO post regulator which I didn't include here):


I once tried it with TO-92 BJTs (2N2222 and BC-327 40). a single phase design only had a 0.185W capability (12V at 15mA. I'm sure it was constant power since putting a 500ohm load, dropped the output voltage to about 9.6V). I tried adding two more phases to it, but it just kind of died, I'm not sure why. so I want to use higher current devices such as fets in the next version, but I'm not sure if they will die as well or not.

Edit: the push pull amplifier design is needed since the DC40106B is only capable of driving 1 or 2mA at 12V.