Category Archives: Thermal Design

Please, please don’t use pre-heated air in your thermal designs and PLEASE simulate the design using CFD or at least have a nice, calm chat with your mechanical/thermal team prior to design freeze on your electrical layout.   Once a design gets locked down the thermal design options start to get limited or expensive or ugly.  Or all three!

Here’s what we’d like our readers to do instead:

• Do have a nice chat with the EE and ME team together.  Maybe bring some bagels and make it peaceful.
• Do some quick look over of any designs for obvious issues, like, walls of memory in front of hot cpus.   Yes, as ME’s and Thermal Engineers we understand that there are ever sticter rules in layout and routing of high speed memory buses.  BUT, let’s compromise together so we can keep them happy and cool 🙂
• Do some first order modeling, on paper or a spreadsheet.  How do things look relative to your thermal budget? (please click to our nice white paper on how to do first order thermal modeling if you need a reminder)
• Then, as the EE design gets placed out, maybe do some CFD to see how everything looks thermally.

While this seems like it’ll add a week to your design, it won’t.  Maybe two or three days and the end result is a less expensive design that is:

• More reliable with higher MTBF
• Operates at the maximum frequencies
• Is safer

Robin Bornoff, a Product Manager at Flomerics UK (now Mentor Graphics), has some great thoughts on this topic.  We’d recommend having a read of her article, “Get On Board: Solving Thermal Problems at Board Level“.

Tony Koryban is one of the most knowledgeable and most humorous thermal engineers in our industry.  Happily, many of his unique answers are available for us all to learn from.   From time to time, ATS is delighted to bring you some of his wit and excellent answers to various heat transfer questions.

“Dear Everything-You-Know Guy,

What the heck is wrong with my heat sink? Did I get a bad batch of aluminum or something? A microprocessor on my circuit board is running a little too hot, about 60 degrees C above the inlet air temperature to the box. So I added a heat sink that had about three times the surface area of the microprocessor package. You’re always harping on the idea that heat sinks work by increasing surface area. I expected the temperature rise to go down by 40 degrees, you know, in inverse proportion to the increase in surface area. That would give me a case temperature 20 degrees above the inlet air. OK, maybe not quite that low, being properly pessimistic. But the darn thing only went down 5 degrees! What am I doing wrong? Maybe I need a pin fin heat sink?

Heatsunk in Harrisbur”

So what’s the solution?

Dear Sunk,

Keep working on that pessimism. Only by imagining everything that could possibly go wrong will you ever achieve a true understanding of electronics cooling.

There are lots of reasons why a heat sink doesn’t give you the temperature reduction you think you deserve…..[read more by clicking to Tony’s archives here]