MOSFET heat

[Nepo]

Hi guys,

@ Bigglez , I was using a 4 kHz frequency with this last circuit.


I'm going to try two in parallel when the hardware arrives.
So far this is what I have come up with for the case/sink.

http://www.wiredvoid.com/MiscPics/WVFC_SINK2.JPG
http://www.wiredvoid.com/MiscPics/WVFC_SINK3.JPG

[Bigglez]

@ Bigglez , I was using a 4 kHz frequency with this last circuit.

I'm going to try two in parallel when the hardware arrives.
So far this is what I have come up with for the case/sink.

That seems like a lot of finned metal for your needs!
While you can never have enough heatsinking, you
will reach a point where adding more has little
benefit.

If your MOSFets are running hot (don't forget that a
human can't tolerate much heat by touch compared
with a semiconductor device) something may be wrong
with the FET driver stage.

As a rule of thumb, an adequte heatsink will feel
hot to the touch but 'good enough' if it can be touched
for at least five seconds without pain. Not very scientific
but a good indicator of trouble.

At four kilohertz your PWM period is:
1/(4*10^-3) = 250us

If your PWM has 125 steps your have a min pulse
duration of:
250/125us = 2us.

The FET will require 60% or greater on time per
min period, suggesting that the rise and fall times are:
(1 - 0.6)/2 * 2us = 400ns each.

This may not be realistic, and as I've made
assumptions, please check my math and your 4kHz
starting point.

Unless your FET gate driver can keep up with this rate
you will force the FETs into their linear region and
guess what? They will get very hot!!

Can you lower the PWM rate? For a fan motor load
you could operate at 40ms (25Hz) and get equal
results. Perhaps 25Hz is more audiable that
4kHz? In a vehicle I doubt you will hear either.

The design of a sub-microsecond Power MOSFet
buffer is not trivial.

[dyarker]

From first post:
"I was going to use a shorter heat sink, but the thought of it in a sealed box in the summer heat is starting to make me nervous. Is there any problem with using two MOSFETs in parallel? (with smaller heat sinks)" (bold added)

To rephase my previous post: It does not matter how much heatsink you put in a sealed box, you're just heating the air inside the box. It will just keep getting hotter inside the box. Just how hot depends thermal resistance of box material. I really recommend an insulating kit (non conductive shoulder washer, ring and film; and heat sink grease); you may, or may not, need a heatsink on outside of box. It depends on surface area of box and air flow where you intend to install it.

Cheers,

[Bigglez]

I really recommend an insulating kit (non conductive shoulder washer, ring and film; and heat sink grease)

Or a more user-friendly Sil-pad and non-conductive hardware.