At 04:04 29-9-2006, you wrote:

The thing to remember is that the gate on a mosfet (or any fet, I
suppose) is a CAPACITOR.  You have to charge it to a sufficient
voltage to bias the mosfet; the charge time depends on the resistance
or current limits of the driving circuit.  You also have to bleed off
the charge to shut it off again.

In fact it is a little more complicated than this - the gate-to-source capacitor is varactor, in that the capacitance changes with the voltage across the junction.

Most power MOSFET data-sheets include a graph of gate-charge (in pico- or nano-coulombs) as a function of gate-source voltage. You have to inject quite a bit of charge to get the voltage through the region where the MOSFET turns on, and drag it out again to get the MOSFET to tun off.

Also, mosfets as switches are either on or off, both of which generate
minimal heat (low current or low resistance).  However, in the
in-between state, the mosfet will generate a lot of heat.  You want to
minimize this by minimizing the time spent in that state.

Power dissipation in the MOSFET is the product of the current through the MOSFET and the voltage drop across it. When the MOSFET is "on" the current is high, but the voltage is (or ought to be) low, and the power dissipated is low. When the device is "off" the voltage drop can be high but the current - and thus the power dissipation - should be zero. During switching, the current is finite - if changing - and so is the voltage drop, so you do get a brief pulse of much higher power dissipation.

Power mosfets (and hexfets) use many individual fets in a parallel
array to handle the high current.  However, this also increases the
gate capacitance, which lowers the effective frequency response for a
given drive circuit.

It doesn't matter all that much how the silicon is divided up - if you want to switch a large current fast you will need to drive relatively large currents into and out of the gate.

For an amp, you probably want to look at the Vgs versus Rds curve,
pick the spot you're going to run the amp at, and see how much actual
current (and thus heat) you'll be seeing.

For a class-D amp you need to work out how many joules each switching event will dump in your switch, and choose your switching frequency accordingly.

--
Bill Sloman, Nijmegen




_______________________________________________
geda-user mailing list
[email protected]
http://www.seul.org/cgi-bin/mailman/listinfo/geda-user

Reply via email to