Forced Air Cooling

[techturkey]

I am working on a design which contains electronics that have to be cooled by air passing through a duct in a heatsink. The air is forced thru the duct by a fan. The duct contains corrugated foil which increases the surface area of the duct, effectively turning it into a heat exchanger.
Does anyone know of the whereabouts of any standard calculations that can be used to relate the fan rating, duct size, inlet and outlet temperatures, and power dissipated by the electronics please?
Thanks,

 

[Bgoldstein]

That's can be a pretty involved problem, depending on how complicated the geometry is. You should be able to do it using some standard equations, but anything complicated will need CFD. The easiest thing to calculate is the outlet air temperature: Air temperature rise = (heat added to air)/ [ (volumetric flow rate) x (air density) x (specific heat of air)

One of the better information sources I've seen for equations dealing with heatsink cooling is on R-Theta's web page (

http://www.r-theta.com,

go to product info, then Extrusion, then download the first part of the catalog.) A lot of equations in there, but if you go through all the technical stuff it will tell you what you need to know.

Some tips: the fan rating's quoted by fan manufacturers is for the fan working in open air, with no restrictions. You are going to have to look at the pressure drop of your system (which depends on the heatsink geometry, and where the air enters and exits the system) to figure out actual fan speed based on fan pressure vs CFM charts.

For a 1st cut, ignore the effects of radiation and natural convection (which are usually negligible if you have a forced flow from a fan in the temperature ranges seen by electronics.) Also, you will need to assume some type of thermal resistance between your electronic components and the heatsink, depending on how you are physically connecting them.

Hope this helps - there are some simpler approximation methods out there then using the full equations, but I've not had much success with those (especially when trying to estimate pressure drops, and thus fan performance). If you have some specific questions, feel free to post them and we can try to help you out.

 

[ko99]

I agree with Bgoldstein. Pressure drop is critical and I don't know of any on-line programs or text-book equations that handle this well. Usually you are required to input a flow or velocity, but you have to guess unless you know the fan operating point.

Remember the fan pressure is due to more than the duct and sink. It includes the airflow obstructions before and after the duct, including inlet and exit vents and any external restrictions. Any other fans sharing these paths can dramatically effect the pressure.

I also agree that a CFD model is the best way to answer the question in lieu of actual parts.

There are alternatives, however.

You could estimate the pressure drop. Depending on your experience level you may get within 20-50%. If you're very conservative and still have good margin, then you don't need more accuracy.

Also, a mockup might work (actual sink, fan(s), and approx vent sections; foam-core for everything else). If you measure the flow you won't need to know the pressure. Use a hot-wire anemometer to find velocity average (ft/min) over the inlet or exhaust area, then multiply by the area (ft2) to get CFM. It's best to construct a hood for this.

Good luck

ko (

http://www.ecooling.biz)

 

[Uge]

Also try to break every section of the electronic board into a 2D heat dissipating model and start from there. You will have to make assumptions as you go down the board wrt to air speed and direction but calculating it this way is a good starting point w/o CFD.
Funny thing, I did my thesis on this exact thing but with a huge 7 foot high unit. In the end I recommende we oversize the blowers, at a minimal cost, that is if noise is not an issue I do recommend a blower if noise is an issue if this is a large unit - ie quiter and longer life.

 

[alpharam]

Cooling process can be computer modelled and presseure loss thus estimated.

 

[Tunalover]

techturkey-
David S. Steinberg's text Cooling Analysis of Electronic Equipment (I think that's the title) is a great, practical source for this kind of problem (it costs about $140 new though). The good thing about the text is that it was written by a practicing engineer who focused on getting practical results (as opposed to most texts written by academics that worship formality, mathematical rigor, and generality).

What the folks above have said is true: the problem you are attempting to solve can be very involved and may require CFD to get a good approximation. Of course it all depends on how close your answer has to be. At least handcalculation per Steinberg's book is a step in the right direction and will certainly help you validate numerical or experimental results.

Tunalover

 

[ko99]

Great practical text. I still use it from time to time. Title is 'Cooling Techniques for Electronic Equipment'

ko (

http://www.ecooling.biz)