mosfet driver question

"Scott"

** That is your problem.

The IRF840 is a standard MOSFET and NOT a "logic level" type.

Means you have to supply the gate with 5 to 6 volts at least for it to switch on hard.

3.3 volts will only drive some examples partially on and others not at all.

.... Phil

There is a gate capacitance connected to the drain. That sends transient voltages from the 130 volts to the logic chip when switching. A 13 volt spike can be sent onto your digital chip pin if the overlap capacitance for gate-drain capacitance is 10% of total gate cap.

The data sheet specifies the gate threshold voltage can vary between

2.0V and 4.0V. Your 3.3V will, as you have discovered, sometimes work, and sometimes not.

The gate threshold voltage is specified at a drain current of 250uA. Since you're running four times that, you'd need more than 4.0V to be sure you get full saturation. At 5.0V, you'd be good to about 2A. The transistor can take up to 20V, so anything between 5V and 20V should be OK.

--
RoRo

"Globemaker"

** FOAD you pile of psycho shit.

You want to use something like the FQT4N20L available from Mouser for $0.55/ea in small quantity.

Ok, sounds good.

At the source? do mean the low side of the load? or are you actually measuring the Source of the FET? you did say it was ground?

I am going to shoot at the hip and assume you mean the low side of the load being the source of voltage, which thus connects to the Drain of the part..

If you look a the PDF from Fairchild, the Vgs(On)/Th (Threshold) ranges from 2.5 - 4.0 volts, the parts you have from China must be cheap rejects that are borderline and the uC is placing them linear state.

You really should be driving them in this window for a switch application. You need more voltage..

With that type of app, you really should be using some sort of isolation. With that, you can drive a basic opto from the uC which in turn can then switch on the gate signal from a higher voltage source.

Or, just employ a gate driver chip that has a charge pump on it.

Jamie

Ok, now this makes sense... I was misreading the data sheet and thought anything above 2.0V would switch it.

Seems I botched the message when I was typing it late last night. Actually I did mean I was measuring at the drain on the mosfet, not the source.

So the consensus is I need more drive voltage, by using some additional drive logic, either an optoisolator or a mosfet driver. The alternative being to choose a mosfet that is logic level switchable.

That does give me some good places to start from, thanks group!

Scott

In my experience 3.3V is never enough for driving power Mosfets. You'll need at least 5V.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

The threshold voltage is most likely what is hitting you but also be aware that if you get a fast edge on the drain, the miller capacitance of the fet will source or sink some current into the pin of the IC through your 120 ohm resistor. Some uCs don't like that and will latch. Mosfet drivers are specifically designed to deal with these miller currents.

I would either put a buffer stage between the micro and the transistor. Or at minimum increase the gate resistance as high as possible to slow down the rising and falling edge on the drain to something I can live with.

It always a good idea to switch fets as slow as you application will allow. Fast edges when not necessary are almost always going to cause noise issues.