You need to check that Vgs min is lower (with some margin) than Voh min for=
 the=20
driver.

Then you need to calculate the average current requirement for whatever dri=
ves=20
it based on switch frequence and total gate charge.

Then you need to calculate the switching and static power dissipation based=
 on=20
frequency, load current, rise and fall time for the gate driver, duty cycle=
 and=20
Rdson.

Then you can calculate the temperature rise based on thermal resistance to=
=20
ambient. Be careful here though since datasheets can have figures based on =
a=20
minimum cooling copper area on the board (SMD components).

Then you can calculate the highest junction temperature based on max ambien=
t=20
temperature and verify that it is less than max operating junction temperat=
ure=20
for the transistor.

You also have to check that you don't get too high switch ringing voltages=
=20
based on stray capacitance, stray inductance and load inductancs. You may g=
et=20
voltages higer than Vds breakdown over the transistor.

Generally you can say that with lower Vds breakdown and current rating the=
=20
lower the power dissipation will be so don't choose a device larger than it=
=20
needs to be.

You can simulate the circuit with LTSpice (free) and get some indication fo=
r=20
power losses but that requires that you have good models for your component=
s=20
and also add stray capacitances and inductances to the simulated circuit.

At 100W (24V * 4A) you can quickly overheat the transistor with a bad drive=
r=20
circuit.

/Ruben

> Hi all,
>=20
> I'm thinking about switching chips, which would mean going to a 3.3V
> micro. The project is a medium sized LED dimmer (4A capacity at 24V).
> I made a smaller dimmer using a logic level MOSFET that worked well,
> and a larger one that I used a gate driver on.
>=20
> It would be ideal for this project to not use a driver as I've got to
> make a bunch of these, and the costs would add up.
>=20
> But...I am not familiar enough with MOSFET design to know if I can
> successfully drive them directly at 3.3V. Or if by doing that I will
> get excessive internal losses.
>=20
> Any thoughts or parts selection ideas? I used the TrenchFET line on my
> last project and it seemed pretty decent.
>=20
> Josh
> --=20
> A common mistake that people make when trying to design something
> completely foolproof is to underestimate the ingenuity of complete
> fools.
>         -Douglas Adams
> --=20
> http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
> View/change your membership options at
> http://mailman.mit.edu/mailman/listinfo/piclist
>=20
>=20
> -----
> No virus found in this message.
> Checked by AVG - www.avg.com
> Version: 2015.0.6176 / Virus Database: 4460/11038 - Release Date: 11/21/1=
5
>=20



--=20
http://www.piclist.com/techref/piclist PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.