Open the (mosfet) gate!

More gate voltage makes for lower Rds-on. It's a linear device, not a gate or a switch, despite the word "logic."

Don't trust IR specs anyhow. They are notorious for fudging. This one might be safe at 10 amps and Vg=10 and a heat sink.

So the moral of the story is, if you want the gate opened *all the way* with minimum Rds,on, give the gate Vgs,max then?

Michael

PC motherboard MOSFETs usually switch pretty convincingly with just under 2V on the gate.

The drain current can be anything up to about 95A, but you're unlikely to find any rated higher than 30V and a few are only 20V.

Oh dear, I think we have some nomenclature confusion. (Or maybe I'm confused.) When I (we) say the gate is open, that means something like an open circuit. Rds is infinite (or very large) When the gate is closed, it's like a closed switch with minimum Rds.

Your Fet has only 0.01 ohms more resistance at 5V vs 10V. That's not much at all, and the fet is still basically closed at 5V. If you are really pushing the limits in terms of heat dissipation then, yeah a little more gate drive might help.

Well, "opened" means not conducting to me.

But more gate voltage makes for lower Rds-on, as shown in the graphs on the data sheet. That part is almost fully enhanced at 10 volts on the gate.

Ignoring the poor use of nomenclature, yes, (considering the gate to be a garden gate) higher Vgs opens the gate wider. There will be a point of maximum effect or at least diminishing returns, but logic FETs will turn on at logic level voltages, but turn on more fully at higher voltages. Just don't get carried away and apply 20 volts to a gate that is only rated for 10.

Ah ok, open means open-circuit.

I'm just imagining the King ordering the guards, "Open the gate!" and they open it to 90% only, and the King bumps his head on the partially-open gate as he rides in on his horse.

Michael

King: "I said to open the gate!" Guards: "We did! The specs said it's open!" King: "OFF WITH HIS HEAD!"

lol

That's not atypical of FETs advertised as logic level. The most you can expect is that if it's called logic level then the data sheet will specify operation at 5V, and it won't be too stupid (and, these days, maybe even at 3.3V).

I had asked about logic level FETs in s.e.d not too long ago and I was pointed to some with threshold voltages in the 1 volt ballpark. The IRLML6344TRPbF is rated with a typical threshold below 1 volt, passing 1 amp with Vgs less than 1.5 volts and 10 amps with Vgs less than 2 volts. That's a far cry from needing 5 volts.

My point is that "logic level" is a nice-sounding, mostly meaningless name. If you want to trust what the part is going to do, read the data sheet -- carefully.

I didn't mean that there weren't parts that might work at much lower voltages -- just that you couldn't go shopping for parts based only on them being "logic level".

Yes, yes, good point.

I'm wondering if a voltage amplifier could be used to shove 15V or so to the gate... and I could use a cheaper IRF530 / standard MOSFET.

Michael

If you're just switching the thing on and off, and if you have 12 or 15V available, you want a gate driver. There's a plethora of them out there. And lots and lots, too.

A gate driver would work, but how about this?

Michael

That works, sort of... beware the slow risetime of the gate voltage,

How slow is slow? Operation would be restricted to below 30kHz, 1kHz, ???

Would increasing the resistor to 47k improve anything?

Michael

Frequency isn't the main event, it's risetime. You'd have to simulate with a known MOSFET AND defined load. ...Jim Thompson

Since the NPN gate driver inverts the input signal, is there a PNP solution that won't invert it?

Thanks,

Michael