Concept Heatsinks and airflow

[BaK]

To follow a variant of a @Phryq's idea I came up with, I'm wondering if the following setup will cool an heatsink with fins* efficiently?
(*This will be an heatsink with the apertures on its smaller sides as opposed to a regular top-down heatsink)

As you can see, only the bottom part of the airflow generated by the fan is going to travel through the fins.
Obviously the distance between the fins would need to be large enough to decrease resistivity and let the air travel a long distance through an heatsink.


My first pick is an heatsink with fins as it offers more surface to dissipate the heat, but maybe a common extruded heatsink with an 'inner' fan will work better?

What do you guys think?

 

[Phuncz]

It depends as air handles somewhat like a fluid, following the path of least resistance. So either force it down the heatsink with a shroud or reduce the resistance (fin density). If the fins are already spread out enough, it might already work. But ideally you'd want air going through as much as surface area as possible, meaning forcing air through a dense fin structure. Depending on how much energy you need to dissipate, it will influence how much fin density you need.

 

[BaK]

Thx @Phuncz, your appreciation is quite right I think.

I could force more air to go through the fins using a duct, but it will need more room for it and actually I want the fan to cool other elements too.

As you say, I would like to use the more dense fins structure while still letting air going through it.
Let's say this heatsink is supposed to cool down a CPU of 100W, according to its size (12*15*2.5 cm), is there a way to calculate how many fins are needed?

On another hand, with the inner fan extruded heatsink, I will not need any more room for the fan which is a plus.
Question is, with the same volume, what kind of heatsink is supposed to work better?
I guess I either need samples to check that or get a physics teacher to calculate this!

 

[Phuncz]

There will no doubt be simple calculations possible with simple fan layout and a constant air flow and static pressure. But those will also result in poor cooling if meant for CPU or GPU cooling. There's a lot more to cooling components than heatsink size and fin density and you're touching the subject of thermodynamics. I'm not really of much help in this regard either.

 

[Kmpkt]

Pretty hard to determine optimal fin spacing, fan profile and heatsink material/dimensions through trial and error. Honestly your best bet would probably be to test an array of existing heatsinks and fans to find an optimal spacing and size of fin and then try your best to copy that.

The reality is that in order to properly model and optimize a heatsink, you're going to need some pretty hefty tools. You need to be able to accurately model/account for air turbulence (or lack thereof), rate of airflow, laminar flow and boundary layer adhesion among other things. Furthermore you'll need to be able to accurately model thermal transfer from your IHS to the various areas of your heatsink as it is very much not uniform. On top of this, you also need to be able to accurately assess the airflow profile of your fan of choice and optimize that as well. I am presently doing exactly this and according to the thermal design engineer I am working with the COMSOL license we're using is around 5000USD for a student license and closer to 10-15K for a commercial single user license.

Let's say this heatsink is supposed to cool down a CPU of 100W, according to its size (12*15*2.5 cm), is there a way to calculate how many fins are needed?

Again I don't think it is that simple. Look up boundary layer so you can understand how the surfaces of the cooler will reduce airflow as it passes. Fin spacing basically comes down to having the minimal distance at which adequate air can move between the fins for optimal transfer. Too close and airflow will slow/alter on its way through killing cooling. Too far apart and you're going to get good cooling, but you're wasting surface area.

 

[msystems]

I would suggest taking a look at the fin spacing arrangement being used in a natural convection heatsink of similar size. Without forced airflow I don't think the air is going to make it through the entire length of the fins. If you use spacing based on natural convection, your fan should improve the heat dissipation but won't be essential to it's operation. What kmpkt said is the most important thing: Fins that are too far away, you lose efficiency, but too close and it won't work at all.