Reading MOSFET SOA graph?

[John]

I need a reality check here.

Am I reading the FBSOA graph (Fig. 11) correctly for this IXYS MOSFET?
http://ixdev.ixys.com/DataSheet/99022.pdf

It looks if I wanted to use it as a 100% duty-cycle switch (on for
seconds to minutes at a time), I'm limited to a Vds of about 25V?
Even if my current level is low, say 1A? I've never seen the Vds
rating drop with increasing pulse width before but I've always used
Intl. Rect. FETs and, so far, each one's Vds max has been the same for
various pulse widths in their SOA graphs.

When running a MOSFET as a basic power switch (essentially 100%
duty-cycle), I've always run it without regard to the SOA as long as
my Vds was comfortably below the rated max, I keep the junction temp
low enough, etc. I'm wondering now if that's not a real good idea.

How can I extrapolate a DC curve for something like an IRF1405?
http://www.irf.com/product-info/datasheets/data/irf1405.pdf

Or do I not have to worry worry about 100% duty cycle operation for
the IRF1405 because, unlike the IXYS FET, the IR FET's Vds rating
doesn't change with pulse width?

Thanks!
John
-- remove SPAMMENOT for e-mail responses --

 

[Phil Allison]

"John"

I need a reality check here.

** Only the wise ever imagine this.

Am I reading the FBSOA graph (Fig. 11) correctly for this IXYS MOSFET?
http://ixdev.ixys.com/DataSheet/99022.pdf

It looks if I wanted to use it as a 100% duty-cycle switch (on for
seconds to minutes at a time), I'm limited to a Vds of about 25V?
Even if my current level is low, say 1A?

** Not true at all.

The graph shows that at 24 volts, a current level up to 30 amp is
llowable - since that equates to the 720 watts. Such a condition implies
the device is NOT switched hard on - ie the gate drive voltage is circa
4 volts instead of the needed 10 volts.

All the " DC " curve on fig 11 shows is the combined effects of "on"
resistance and the max allowable dissipation of the device.

Note the heading says "Forward Bias "

- not " Saturated Switching".




......... Phil

 

[John]

Am I reading the FBSOA graph (Fig. 11) correctly for this IXYS MOSFET?

Aha! So the DC curve is really just a visualization of the Pd rating
(730W for that FET) for when the FET is operated linearly, plotting
the Id and Vds that result in the 730W Pd rating.

If I'm hard-switching the gate (e.g., Vgs = 10V), then those curves
aren't applicable and I can use the other graphs to make sure I can
pass the current I need to with the available Vds, how much power I
have to dissipate (using Id and Rds(on)), etc.

That explains a lot, thanks Phil.

John
-- remove SPAMMENOT for e-mail responses --

 

[John]

The graph shows that at 24 volts, a current level up to 30 amp is

Is it just me or does the DC curve not really offer very much? With
the Pd rating, and knowing my Vds and Id, I can easily calculate
whether I might exceed the device's rating. I guess it can save me a
few seconds time if the graph is handy.

But without something like submerged ice-water cooling of the FET and
its heat sink, I'm never going to get even remotely close to the
device's 730W rating when operated linearly at 100% duty cycle!

Might explain why IR doesn't publish DC curves?

John
-- remove SPAMMENOT for e-mail responses --