Answers to those Doggone Thermal Design Questions
By Tony Kordyban
Copyright by Tony Kordyban 2003
Dear Everything, etc.,
Last year, when our company still had a thermal engineer, she designed a heat sink for the microprocessor on my board. Without the heat sink, it was about 25 degrees C over its operating limit. The heat sink brought down the temperature about 20 degrees. That wasn’t quite good enough, so she added a special thermal grease between the microprocessor and the heat sink, and the temperature went down another 10 degrees. It was a great success.
I am now doing a cost reduction on that product. One way of saving money is to use a processor with less horsepower — and less heat, too. Without a heat sink it is only about 8 degrees C over its operating limit. Can I save even more money by eliminating the heat sink and just use the thermal grease to reduce the temperature by 10 degrees?
Penny-pincher fromPalmyra
Dear Penny,
Following is the answer that I am sorely tempted to give, but EVERYBODY ELSE PLEASE UNDERSTAND THAT THE FOLLOWING ANSWER IS WRONG AND INTENDED TO PUNISH A COMPANY THAT LAID OFF ITS THERMAL ENGINEER AND SHOULD HAVE TO PAY THE PRICE FOR GETTING RID OF NECESSARY TECHNICAL EXPERTISE:
Yes, you can reduce the temperature of your microprocessor by just adding a layer of thermal grease to the top surface. In fact, the thicker the grease, the more the temperature will go down. It operates on the same principle as suntan oil, which keeps your skin cooler in the sun and prevents sunburn. The only danger in adding too much special thermal grease is that the temperature might go below the dew point and water would condense on the microprocessor. Grease is like oil, and everybody knows that oil and water don’t mix, so that could be disastrous. In fact, that is why thermal grease comes in very small tubes. If you had a large amount of it all in one place, the large drop in local temperature would make the grease freeze up solid and you could never get it out of the tube. So don’t store the tubes together in the same cabinet, unless you have an ice pick handy.
Because of ethics and lawsuits and such, and probably because it isn’t your fault that the thermal engineer was laid off, I’ll give you the right answer, too. Perhaps from the tone of my wrong answer you have begun to see the flaw in your reasoning already.
There is no magic substance that can reduce the temperature of a heat producing object just by smearing it on, at least not in the steady state. You could get some temporary cooling by rubbing the surface of your microprocessor with alcohol. As it evaporates, the temperature would drop, the extra power absorbed as the latent heat of vaporization of the alcohol. But once it is gone, the component temperature would come right back to what it was before.
[Black paint is another exception. If you increase the thermal emissivity of the surface with high emissivity paint, you can increase the thermal radiation lost from the component, assuming there is a colder surface to radiate to. But thermal grease does not work that way.]
You are the victim of a very human habit called magical thinking. We think certain substances have special “powers.” Here are some common examples you might recognize:
- a Philosopher’s Stone can change lead into gold
- a copper bracelet can absorb arthritis pain
- aluminum absorbs heat like a sponge from hot components
An aluminum heat sink, unlike the first two examples, actually works, but not because of some magical power of aluminum. The heat sink has two features: it is a good spreader of heat and it has lots of surface area. The heat from your component has to end up in the air. The more surface area in contact with the air, the easier the heat will flow into the air, and the lower the temperature of the component. The ability to spread heat well allows the heat to get from the component to where all the surface area is.
How does the thermal grease help? When you clamp a heat sink to the top of a component, you introduce a thermal interface, whether you intended to or not. Not matter how flat and smooth you think the component and the heat sink are, there are microscopic bumps, valleys and pores in each. Murphy’s Law guarantees that they won’t fit together like a jigsaw puzzle. When you press them together, there is only a small amount of solid-to-solid contact. The rest of the gap is filled with air (or vacuum, if your circuit board is in space), and air is a pretty good thermal insulator. The heat has to pass across this joint, which is partly solid-to-solid (good) and mostly microscopic air gap (bad). You have a thermal insulator between your heat sink and component, which leads to a temperature difference.
Throw in a thin layer of thermal grease and squish the heat sink back down. By itself, grease is not such a great conductor of heat, but it is much much better than air, and it is very good at wetting surfaces. When you squish the surfaces together, the grease flows into the cracks and pores and gaps between the sink and the component. Now you have the same solid-to-solid contacts, and a lot of tiny gaps filled with grease instead of air. So heat flows better across the joint and you have a smaller temperature difference between the component and the heat sink.
What happens if you just have grease and no heat sink? Without a sink, there is no extra surface area, so the temperature of the component can’t go down at all. There is no joint between the sink and the component, so you can’t improve the conductivity of the joint. The thickness of grease is now a layer of thermal insulation between the component and the air, so your component temperature will probably increase a little bit.
In your particular example, you can still achieve a small cost reduction. Keep the heat sink, and eliminate the thermal grease. With the lower power of your microprocessor, you can live with the temperature rise associated with the “dry” joint. And you can save additional money by not having to keep the tubes of thermal grease in those specially heated storage lockers.
—————————————————————————————————————
To the Thermal Wizard of Odd,
A friend of a friend told me that there was a man that got burned on, uh, shall we say, his lap, by his laptop computer. He supposedly was sitting, fully clothed, with the laptop computer on his lap, engrossed in writing a report for an hour. He didn’t notice the burn to his lap regions until the next day. It was like a bad sunburn, enough that he had to seek medical attention.
This story sounds pretty fishy to me, but it should be right up your alley. Did this really happen? Should thermal engineers take this seriously?
Surfing Standing Up for the Time Being
Dear Surf,
This is a true story. I found it published in The Lancet, a respectable medical journal that has been around long enough that its editors still consider anesthesia to be innovative.
You don’t have to take my word for it. The on-line version of the November 23, 2002 issue gives the name of the reporting doctor and his medical institute inEurope. It doesn’t name the embarrassed victim, for obvious reasons. If he were in the US, we’d all know his name and face, and perhaps the appearance of his burned personal area, because he would be on every TV talk show. He would be suing the pants off the laptop manufacturer. Perhaps he wouldn’t be using the phrase “pants off,” though.
The heat dissipation of laptops is getting worse over time, just as it is for most kinds of electronics. Given a shrinking package size, and limited room and battery power for fans, the the surfaces have to get hotter.
How hot do they get these days? I’ll bet that nobody really knows. Not that laptop engineers haven’t measured, but mostly because the temperature of a box cooled by natural convection is very dependent on the surrounding conditions. And it is very difficult to foresee all the possible conditions people might use a laptop in.
This is a tricky problem, because there isn’t much difference between a merely uncomfortably warm surface and one that can damage your skin, if you leave it in contact long enough. In one way, a skin-sizzlingly high temperature would be safer, because at least you’d feel pain and get away from it. Unfortunately, it seems that this particular laptop, under just the right (or wrong) operating condition, got hot enough to burn skin without being hot enough to cause immediate pain. I have done this to myself, when soothing a backache with an electric heating pad. I, too, didn’t feel the burn until the next day, and it wasn’t until I applied the heating pad again that I realized why my back felt sunburned.
What lesson can we learn from this fascinating case? One comes to mind from my old high school coach: Always wear your cup! Who knew it would be needed outside of sports?
Another is that we probably need some better definitions of Thermal Worst Case. Perhaps for laptop computers, the test environments should be defined by the advertising department instead of the safety department. After all, it is the ads that show people using laptops in bed, at the beach, or while curled up on a leather couch in front of a fireplace. I bet none of those environments were tested. Can you imagine how hot a black plastic laptop case gets with the summer sun beating down on it at the beach?
I hope we get this skin burning problem figured out before they start selling those “wearable” computers. I already don’t like wearing clothes emblazened with company logos. I sure don’t want to have a computer logo branded into my skin.
—————————————————————————————————————
Isn’t Everything He Knows Wrong, Too?
The straight dope on Tony Kordyban
Tony Kordyban has been an engineer in the field of electronics cooling for different telecom and power supply companies (who can keep track when they change names so frequently?) for the last twenty years. Maybe that doesn’t make him an expert in heat transfer theory, but it has certainly gained him a lot of experience in the ways NOT to cool electronics. He does have some book-learnin’, with a BS in Mechanical Engineering from the University of Detroit (motto:Detroit— no place for wimps) and a Masters in Mechanical Engineering from Stanford (motto: shouldn’t Nobels count more than Rose Bowls?)
In those twenty years Tony has come to the conclusion that a lot of the common practices of electronics cooling are full of baloney. He has run into so much nonsense in the field that he has found it easier to just assume “everything you know is wrong” (from the comedy album by Firesign Theatre), and to question everything against the basic principles of heat transfer theory.
Tony has been collecting case studies of the wrong way to cool electronics, using them to educate the cooling masses, applying humor as the sugar to help the medicine go down. These have been published recently by the ASME Press in a book called, “Hot Air Rises and Heat Sinks: Everything You Know About Cooling Electronics Is Wrong.” It is available direct from ASME Press at 1-800-843-2763 or at their web site at http://www.asme.org/pubs/asmepress, Order Number 800741.