Can be viewed as a webinar at: http://www.mentor.com/products/mechanical/multimedia/thermal-bottlenecks-shortcut-webinar
Electronics thermal management involves the design of an electronics system to facilitate the effective removal of heat from the active surface of an integrated circuit (the heat source) out to a colder ambient surrounding.This presentation will introduce the concepts involved in rerouting thermal bottlenecks through new heat flow paths.
‘Perfect’ in so much as there’s nothing really that can be done to reduce the temperature rise apart from using a totally different higher conductivity material or making the sphere much smaller. The point is that there is no structural change to get the heat out more easily.
Conduction cooling by a sphere doesn’t occur in reality. Real systems are highly complex, lots of heat sources, intertwined heat removal paths in 3D. Resistor network diagrams are good to highlight a concept but in reality only a full 3D CFD type simulation will provide the resolution required to properly capture all paths the heat takes as it travels from source to ambient.
Bottlenecks could either be a little heat that is highly resisted, a lot of heat that is resisted only a little but definitely a lot of heat that is highly resisted.
Shortcut opportunities are a little more difficult to conceptualise. A shortcut opportunity exists when there is a lot of heat passing adjacent to cooler areas (for what ever reason). Taking the fluid flow bottleneck analogy, a shortcut opportunity would be to indicate that punching a hole in the bottle to vent the air would be a good way of getting more air through the system. From a thermal perspective this is simply adding a new heat flow path to let the heat gush out to cooler areas and thus head to the ambient more easily.
Wall Unit is an application example installed with FloTHERM. With V9.1 it comes with Bn and Sc fields stored.
The Bn/Sc simulation methodology is to look at the variation in one of the numbers in a basecase to identify the first design change to make. In this case we start with Bn and note that it is highest at the interface between the cans and the motherboard where the can connectors abut the surface of the motherboard. The remedial action is to push the connectors through the motherboard (How would Command Center have told you that???).
Note also that the Bn and Sc numbers could be telling you where to ensure your modelling assumptions are correct. If Bn is high in a bit of geometry that has been oversimplified then you’ll need to refine the representation for accuracy. Either way, Bn and Sc indicate sensitive areas of the model to temperature rises.
Can temperature rises go down, comp1 goes up a little (graph shows DECREASE IN TEMPERATRUE RISES)
On a resolve note that the Bn distribution is more ‘uniform’. A perfect design would be where the Bn number is uniform everywhere, where heat is finding it equally difficult to leave everywhere, will never happen of course. The next candidate bottleneck is between comp1 and the thermal pad on the underside of the motherboard. Adding thermal vias is the obvious design remedy.
Comp1 temperature rise drops a wopping 24%
Further Bn modifications could be made, hunting down the next maximum. Now though we swap to Sc prompted design modifications. Back to the can+connectors. Something could be done in that area. It doesn’t tell you exactly what to do, only where to do it. For this application lets push the connectors all the way down onto the heatsink itself.
Can temperature rises drop a wopping 46%, comp1 gets a little warmer
Again, note the more uniform distribution of Sc. For the next modification we use the fact that Sc in fluid near a wall is highly correlated to local Nusselt number, e.g. a measure of the effectiveness of heat transfer, for the heatsink it indicates which areas of the heatsink are working effectively and which aren’t. Heatsink fins are foreshortened and amazingly the temperature rises decrease even more! This is due to a reduction in friction losses through the heatsink due to reduced fin area, thus an increase in air flow rate thus an increase in the effectiveness of the heatsink!!!
We believe that this new approach to using simulation for design may well revolutionize the way in which thermal simulation is performed. This approach will only be available in FloTHERM, not in any competitor products.