Natural Convection Heatsink for 80W power dissipation?

[Magnethicc]

Hi guys,
Does anyone have experience with convection cooled heatsinks for a power dissipation of around 80W?
We are looking for a design of such heatsink and we came up with a vertical fins stretching from the bottom to the top and was wondering if there is a better design.

This is a top view of the heatsink.

Thank you for your replys.

 

[wwfeldman]

it looks like there are 6 dissipating devices on each heat sink
if you are using natural convection, there needs to be space above and below so the air is free to flow naturally.

the manufacturer of the heat sink should specify how much area each dissipating device needs to
get the specified thermal resistance. power dissipated = thermal resistance x temperature difference
between two sides of heat sink - hot side, with mounted dissapaters and cold side, where the air flows.

Understanding Thermal Resistance - learn.sparkfun.com

 

[crutschow]

There needs to be free air flow around the sink and, if it's in a container, that also has to have large openings to allow the heat to escape.
The heat sink size will also depend upon the highest ambient temperature it will see.

 

[andre_luis]

Grossly speaking, you could represent the isotherm shape of the steady air around the heatsink built with narrow blades, as a flattened surface with a quite smaller area compared to the current heatsink geometry; the smaller the gap, the smaller the effective surface. This happens because the air in the deepest regions warms cumulatively without sufficient time to spread out. The following illustrations shows in the red curve a sketch of what is be the effective surface of a heatsink originaly made for forced air, but being used fanless.

So, in short, there are heatsinks suited to either work exclusivelly with or without forced air, and form the aspect ratio of the blades of the one you have, at first sight seems not adequate to work with natural convection.

 

[Easy peasy]

you need to state the total dimensions and the max ambient air temp - also the ability for free air to convect up through and away from the top of the heatsink

adding a fan to a smaller finned heatsink can increase efficacy markedly.

 

[dick_freebird]

I'd put the fins facing outward and transistors in the
middle. Get more / better air that way, not throwing
heat at each other. And potentially not suspended
over an airflow-blocking PCB plane.