When does a MOSFET need a fan to cool?

[eem2am]

Thanks,

Interestingly, this tiny little DO15 TVS is supposed to be able to dissipate 5W continuously at 75C ambient....

....or am i missing something here?

http://www.datasheetcatalog.org/datasheet/comchip/P6KE7.5C.pdf

 

[AndrewT]

because it can run reliably ....ing hot !
If placed close to PCB materials it will usually char the board within a few dozen hours of operation.

 

[eem2am]

The datasheets for the P6KE TVS says it can do 5W continuous when soldered to 40mm^2 of copper on PCB.

However, we dont have any appreciable copper PCB area, and are wondering what maximum power we can assume continuously with our situation.

These "DO" type diode packages ......i don't see any reason why they should have any greater power dissipation capability than a TO220...and a TO220 without a heatsink has very limited power dissipation capability.

Would you agree with my gut-feeling that the contiuous power dissipation capability is around 300mW max, with no appreciable PCB cooling area.?

 

[marce]

As Andrew as said these devices would scorch the board, they are generally mounted in metal standoffs to raise the body a few mm above the board surface.
The copper stated in the data sheet is for a test board, it has no bearing on real life thermal design of products. The only way to properly engineer a design thermally is with very expensive CFD software and or a thermal camera. If you get involved with MTBF and reliability predictions you instantly de-rate most components by 50% just to get reasonable MTBF.
Also with PCB copper the max area that is usable is 2" x 2" this gives a figure of 16-23 degC/W, thermal management using PCB's mis a whole new world, that involes either simple designs on metal clad boards or multiple planes and numerous thermal vias.

 

[eem2am]

well i understand they run hot, but there is no indication on the datasheet that its ok to run them at scorching temperatures....no indication of the litetime limitation?

 

[marce]

A program called relex is uded to calculate MTBF, this is based on a number of standards,
and this software determines the lifetime of a product based on numerous sources of data you feed in. With any semi-conductor based device lifetime is mainly down to the temperature the junction runs at, so the more you de-rate a semi device and the cooler you run it the longer it will last...

 

[AndrewT]

there is no indication on the datasheet that its ok to run them at scorching temperatures....no indication of the litetime limitation?

there are lots of clues, better than clues really. They tell you the maximum power dissipation. They tell you the maximum ambient temperature. They sometimes show you the power de-rating graph for other duties. Read the datasheet.
They do not generally tell you the surface temperature of a device. Similarly they do not generally tell you the temperature of the lead out wire as it exits the package. If you need that information then build and measure. If sufficient data is available then you can simulate, but never rely on simulation until you can prove that the results are reliable, i.e. compare the model to the built assembly.