Heat does not transfer to the heatsink in electronic cooling

[Anonymous]

I am currently working on the thermal simulation of several Mosfets with a heatsink. I placed a small cube in each Mosfet as the heat source. I initially used the thermal study and everything works fine (see in the first picture). Then, I want to have forced air blowing by a fan in my simulation. Since I did not find any option in the thermal study, I chose to run it in electronic cooling for the same model. I added a fan model and made it my forced flow(the configuration is attached). However, after I ran the simulation, the heat from the inside boxes seems not to transfer to my MOSFET packages and the heatsink effectively(see in the third picture). The heat is trapped inside the Mosfets and leads to an extremely high temperature. I wonder if it is because, in the electronic cooling study, I cannot config the convection and radiation that are in the thermal study. Is there any way to solve this issue? Or is it possible to create a forced flow environment in the thermal study? Thanks!

[jorge_garcia]

Hi @Anonymous ,

I hope you're doing well. I'm going to tag the e-cooling team on this but from my experience there is no reason to create the blocks. The components themselves can have be assigned the heatload. Make sure you also assign the materials of the components correctly, that will play a role in the accuracy of the simulation.

Best Regards,

[heath.houghton]

It appears that the simulation is stopping very early on for your model.  The solver is instructed to stop if any components approach an unrealistic number for anything in a normal electronics application (1000 degrees C).  I would suspect that either the heat load is not the same between the two, or the materials assignment is not complete for the electronics cooling simulation.  If you wanted to send the f3d file that includes the electronics cooling, I will gladly take a look.

Regards,

[Anonymous]

Thanks for your help, Heath. My f3d file of electronics cooling is attached. All the material assignments should be the same as what I did in thermal simulation before. The reason that I place the small block inside the Mosfet is that I do not want to make the whole Mosfet as the heat load itself. I am a beginner at this software and I really appreciate your help.

[Anonymous]

Thanks for your help, Jorge. The reason that I create the blocks inside the Mosfets is that I do not want to make the whole Mosfets themselves as the heat loads. The heat inside the Mosfet can transfer to the package and then transfer to the heatsink and air.

[Anonymous]

Hello Heath, sorry to bother you again. I know you might be busy. Have you checked the model that I attached in the last reply by any chance? Do you have any suggestions on it? Thank you!

[heath.houghton]

I was able to take a look.  The heat transfer (heat sink idealization) designation is not needed for your model.  It is a feature we plan on expanding and it's purpose is to shorten run times when you have a very fine detail heat sink.  Your heat sink is bulky enough that we don't need to idealize it.  The software should have run fine either way, and we will look into why it didn't run with the designation.   When I removed the heat sink idealization, the software calculated a very high temperature (max temperature near 600C), causing flow results to look funky.  If I go to maximum resolution (max accuracy), the temperatures come down slightly.  The issue is you have a lot of heat and not much surface area to dissipate it.  The heat sink temperature is still roughly 100C.  If I turn the T0220 encapsulating epoxy into a more conductive material, then I get a die temperature of roughly 110C.  I still don't like the looks of the flow velocity results so I will tinker with the model a little more today.

[heath.houghton]

I added a box around the test assembly to constrain the flow conditions a little.  This allowed the flow solver to give a much improved result.  The thermal results looked like the heat load was partially being produced in the very low conductivity encapsulating epoxy material.  This caused elevated temperatures there.  I took the liberty of changing the material of the epoxy to something more conductive to test that theory.  The temperatures came down to what you see below.  Our development team is looking into the heat being generated in the non-conductive material.  I also took the liberty of moving the fan to the other side as you can see. 

[Anonymous]

It looks a lot better. Thank you very much for your help! I will also try to keep working on my model based on your suggestion and modification.