When does TO220F FET blow up?

[cupoftea]

Hi Supposing you have a TO220F FET , which you wire to a gate drive and current source so it dissipates 3W
The TO220F is stood up in free air, 23degC. No heatsink, No copper pour round its leg solder points.
How long would you expect it to work before it went bang?


FET is STF3LN80K5
STF3LN80K5

 

[danadakk]

Tjmax is 150C
Rthermal is 62.5 C / W

So Tjunc = 23 + 62.5 x 3 = 210.5 C

So guessing wont take long.....depends on how much heat is being
removed by PCB layout thru leads, in your PCB case not much. Also
this assumes stagnant air velocity.


Regards, Dana.

 

[cupoftea]

Thanks , Pretty much no heat is being removed from the skeletal PCB pads, or the wires.

 

[danadakk]

Some experiences :

How Mosfets Burn Out with Heat

In my experience, overheating Mosfets fail with all the pins shorting out. I guess the mosfet turns into a conductive blob. Then the currents can 'take out' other electronics too. Poof...then poof...and poof..until something breaks the loop. Is this a typical fail mode? D from BC British...

maker.pro

Section 8 shows some failure modes :

https://www.nxp.com/docs/en/user-guide/MOSFET-Application-Handbook.pdf

Regards, Dana.

--- Updated Aug 5, 2023 ---

Of course there is always considerations, in time, eg. thermal mass, that give rise to
thermal time constant :

https://www.mikrocontroller.net/attachment/600970/APPCHP7.pdf

Regards, Dana.

 

[KlausST]

Hi,

never done this before, still I try to find a physical/mathematical solution.

you have 3W of heat.
* One part of it is disspated as heat into the environment
* the other part is used to rise the temperature of the TO220.

For ease of math, let´s assume the dissipated heat into the environment is linear with temperature difference to ambient temperature.
And let´s just use the heat capacity of a (random) TO-220

Then the critical temperature of 150°C is after about 60 seconds.

Attached a ZIP file containing the EXCEL97 file. Should open with older and newer Excel versions.

Klaus

 

[danadakk]

Of course Tjmax exceeded on a much faster timeframe.

Page 15 and 19 show Tj happens on a much faster timeframe in the TO220 package :

https://www.infineon.com/dgdl/Thermal+Modeling.pdf?fileId=db3a30431441fb5d011472fd33c70aa3

So question is what drives the "bang" the OP is asking for, whole package mechanical expansion
issues of different materials, and or gasses released asTj melts / creates hotspots.....in much faster
timeframe ?


Regards, Dana.

 

[cupoftea]

Thanks, anyway...the STF3LN80K5 blew up on me.....after an 8 hour day of running the 20W flyback in overload and short ccting it repeatedly, the STF3LN80K5 eventually blew up after being on with the flyback having 40W output (overload)...and had been on for 10mins that time. It was running on 90VAC and was in pretty deep continuous mode at this overload level. The gate series resistor was 180R.....it was the only value i had at the time....but Ciss is only 100pF for this FET anyway. The RMS FET current would have been around 700mA RMS...but the switching losses would have been appreciable since it was in deep CCM.

I had also been running it with 59W output from 265VAC.....for kind of 10 mins at a time, repeatedly, throughout the day...and it didnt blow up then...but it was still in DCM in that case.
(overload shutdown was ~60W at 265VAC and ~40W at 90VAC).

 

[Easy peasy]

TO-220 in free air = 1W max, 20 deg ambient, a bit less for the F version