By Tony Kordyban
Now Available. Order Your Copy Today.
More Hot Air is the long-awaited sequel to the author's previous ASME Press book, Hot Air Rises and Heat Sinks: Everything You Know About Cooling Electronics Is Wrong. This new book continues in the same humorous and easy-to-read style of the earlier book, with all-new, original case studies in the field of electronics cooling. Each case study, told as an anecdote, is designed to teach a basic concept of heat transfer, as applied to keeping electronics from overheating.
Because of the constantly shrinking size of electronics, the job of cooling electronics continues to get tougher. Many people not trained in the basics of heat transfer have been roped into doing this job out of necessity. For those who lack any formal training in heat transfer, the case studies explode many of the myths about cooling electronics and replace these flawed practices with sound engineering, based on actual heat transfer theory.
The case studies and humor in this book are also entertaining to those well versed in electronics cooling. A must-read book for all engineers and their managers concerned with electronics packaging.
Table of Contents
Measurement and Test: Getting the Wrong Answer Direct from the Lab
Chapter 1.1 The Best Worst Case
The requirements say to measure the product temperature under the "thermal worst case environment." But the Reliability Department, the Safety Compliance Department, the Thermal Engineer and the Customer all have different ideas of what the thermal worst case should be.
Chapter 1.2 Blowing the Rel Test
The blowers in the Reliability Test Chamber makes the air flow backward through your chassis. Does that seem like a fair test? Or does it actually tell you something useful about your product design?
Chapter 1.3 The Five-Finger Thermometer
Why your hand is not a good thermal sensor. It not only has calibration problems, but you might literally get burned.
Chapter 1.4 T-types Fried My Brain
There is a reason why different types of thermocouple wire have unique color codes. You can't always tell the difference between the types by just using common sense. It takes brains.
Chapter 1.5 Permutations and Combinations Add Up to Job Security
Maybe it makes sense to stack shelves up in a rack and use one big fan box to cool all of them. But there are lots of thermal reasons not to like that design. Endless combinations of hardware can keep you doing thermal testing for years.
Chapter 1.6 Power Confuses, and Variable Power Confuses Absolutely
For some types of components, power dissipation depends on the component temperature. Sometimes it goes up, sometimes down as the temperature increases. In that case a room temperature test can give you results exactly opposite to what you'd get at an elevated ambient.
Chapter 1.7 How to Get Percent Error 100% Wrong
A story about metrics, and how you can use percent error to get whatever result you want. The important metric for thermal design is the one that measures how well the process is working--usually a temperature difference--not the absolute temperature.
Fans: Increasing the Air Flow and the Trickiness of Your Cooling System
Chapter 2.1 Elbow Room
The boss finally gets me some help to do thermal analysis. But because we have to share a tiny office, we get in each other's way. The same thing can happen when fans are mounted too close together.
Chapter 2.2 Breathing Room
The Marketing Guy questions why the fans need such a large inlet vent. It is demonstrated to him with duct tape and a drinking straw.
Chapter 2.3 The Path of Least Resistance
When air flow is given multiple paths to choose from, it doesn't always follow the path of least resistance. It seems to follow Murphy's Law instead.
Chapter 2.4 Incomprehensible Flow
A frequently asked question is "What's the difference between LFM (linear feet per minute) and CFM (cubic feet per minute)". Incompressible flow is illustrated using melted American cheese.
Chapter 2.5 Fault-Tolerant Cooling
Herbie wants to use a fan/heat sink combination like the one in his personal computer. Does adding a fan/sink increase or decrease the reliability of his circuit board?
Chapter 2.6 Putting the Right Spin on Fan Cooling
Does component temperature depend on whether the cooling fans turns clockwise or counterclockwise? The closer the fan is to the component, the more it matters.
Chapter 2.7 Degrees C and dBs
An important limit to forced air cooling is the audible noise of the fan. As fan RPM goes up, so does the flow, but the noice goes up even faster, according to the Fan Laws.
Chapter 2.8 WKUL-AM
A talk-radio-show host discovers that while paying attention to component temperature, one can forget that the cooling fan itself is a component, too, and has its own operating temperature limit.
Components and Materials: the Sum of the Parts is Sometimes Just a Big Hole
Chapter 3.1 Not Working Within the Limits
What does it mean for an electronic component to have an operating temperature limit? Does it blow up 1 degree over the limit? Does it slowly degrade in function, or does it use up some of its allotted life? Wouldn't it be nice if the component manufacturer would tell us?
Chapter 3.2 Don't Blow It When Sizing a Fuse
Fuses are easy to ignore, but some very common types need to be derated for temperature. Just because they don't dissipate heat doesn't mean they don't get hot.
Chapter 3.3 When It's Hot, They All Go in the Pool
Can the printed circuit board act as a heat sink for a component? Perhaps, but a story about a naughty boy in a swimming pool explains the practical limits of this idea.
Chapter 3.4 Bypass Capacitors?
In thermal analysis of a circuit board, you often ignore all the capacitors, because they aren't supposed to add heat, and there are so darned many of them. But capacitors can generate heat, and their properties can shift with temperature.
Chapter 3.5 A Baffling Temperature Rise
A baffle is often used to deflect hot air from the exhaust vent of one chassis to prevent it from getting sucked into the inlet vent of another. But a baffle is not a perfect way to isolate neighboring chassis, because heat can conduct through the baffle plate. Maybe changing the plate from metal to plastic will help. Don't count on it.
Chapter 3.6 24K Gold Heat Sinks: Worth Their Weight in Aluminum
Computer hobbyists called "overclockers" tout gold-plated heat sinks to reduce microprocessor temperature. The definitions of the three modes of heat transfer--conduction, convection and radiation--explain how gold-plating doesn't help.
Chapter 3.7 Improving the Weakest Player
A salesman promotes his new printed circuit dielectric material with its huge improvement in thermal conductivity. It is 10 times better than the ordinary dielectric material, so why doesn't the temperature of the curcuit board get any better?
Chapter 3.8 Getting Lost in the Cracks
When power gets high, one has to be as strict as Mother Superior from The Flying Nun, that is, very picky about how the heat sink is attached to the component. At high power, the thermal resistance of that joint can make or break the whole thermal design.
Radiation: No, Stefan and Boltzman Were Not a ’70s German Heavy Metal Band!
Chapter 4.1 Seeing (Infra) Red
The basic physics of radiation heat transfer are explained, using Herbie's girlfriend Vernita as the source of radiation. One of the laws of radiation is Murphy's Law, in that radiation is only there when you don't need it, such as when thermal resistance of a component package is measured in the industry standard test.
Chapter 4.2 The Beauty of IR Is Only Skin Deep
Can an IR camera see through clothes? Can it at least see through the metal skin of a chassis? Can it see plumes of hot air? No, but the IR camera is still a useful tool.
Chapter 4.3 "Negative Result--Very Important, Too!"
Why is it nearly impossible to get an infrared picture of a circuit board that looks like the color temperature map from a computational fluid dynamics (CFD) software tool, and what can be learned from the lack of agreement of these important thermal analysis tools.
Chapter 4.4 Selective Surfaces
Selective surfaces can protect outdoor enclosures from solar radiation. But you can't control the selective surface once it leaves your hands.
Tales of the JEDEC Knight
Chapter 5.1 *Circuit Board Not Included
Can drilling holes in a circuit board make components cooler? An Internet chat room discussion suggests it is so. Maybe it's true if the holes are actually vias. The JEDEC definition of 0j-a (thermal resistance between junction and ambient) already includes a board as a heat sink, so counting on your board as an additional heat sink is not likely to help much.
Chapter 5.2 Thermal I/O
A set of Moses-style commandments are given unto component package vendors. Thou shalt design packages with real paths for heat to get out, give users the details of those path(s), and allow users to measure junction temperature directly. Will these commandments be obeyed any better than the previous 10?
Chapter 5.3 JEDEC Standard: Stake in the Ground, or Stick in the Mud?
A thermal engineer uses sophisticated CFD and experimental methods to optimize the design of a new component package. Unfortunately, he optimizes the value of the JEDEC-defined 0j-a, instead of something.
A Collection of Not Even Loosely Related Stories
Chapter 6.1 The Milk-box Problem
How fast will a bottle of milk reach the freezing point in a picnic cooler outside at 20 below zero? This is important, because I might have to get out of bed early to bring in the milk. And it helps me to understand something about the transient temperature response of outdoor electronics to extreme changes in weather.
Chapter 6.2 Specs, Lies and Red Tape
The spec sheet for an electronic chassis says that it needs 50 degrees C inlet air at 100 CFM. Can you trade off temperature for flow rate? How hot can the inlet air be if the flow rate is 300 CFM? Why are spec sheets always wrong?
Chapter 6.3 Thinking Kinks Jinx Sinks
A lsit of reasons why a heat sink hardly ever does what you think it should.
Chapter 6.4 The Magic Pipe
A fairy tale about heat pipes, in which Hodgepodge the Hedgehog helps the Three Bears. His heat pipe, which seems to work like magic but is very real and based on science, equalizes the temperature of Papa and Mama Bear's porridge so they can all live happy ever after.
Chapter 6.5 When 6% Is 44%
A small improvement in the efficiency of a power supply is actually a large change in its heat dissipation. Don't be fooled by a 6% change sounding insignificant.
Chapter 6.6 So Crazy, It Just Might Work
A summary of some innovative thermal engineering ideas from a real technical conference. Did they really get funding to develop a heat sink with a gooey center?
Telecom: A Field With Myths and Mistakes All Its Own
Chapter 7.1 Thinking Inside the Box
Why the air flow in telecom equipment is supposed to go from bottom to top. Some new kids on the block are trying to sneak in equipment with side-to-side flow, just because it allows them to fit more equipment in a rack.
Chapter 7.2 "Just Slap It in an ETSI Cabinet and Voila!"
Converting a telecom rack designed for the U.S. market to meet a European standard is a little more complicated than just learning to read a requirements document written in French. The operating range for temperature is different, and they talk about air pressure when they really mean altitude.
Chapter 7.3 NEBS: the Bible of the Central Office
A summary of the thermal rules in the telecom industry standard document, Telcordia's GR-63-CORE. Violations of these rules are punished in this world, and possibly in the next, too.
Chapter 7.4 The New NEBS: More a Horror Tale Than Another Bible
In 2001 there was a rumor that NEBS would be rewritten. It wasn't, but a new thermal managment standard was issued, and it seems to have been ghost-written by Stephen King.
Chapter 7.5 Normal Room Temperature: the Latest Worst-Case Thermal Condition
Herbie slows fans down to meet the NEBS audible noise limit. But does that make normal room ambient the worst case thermal situation? Not because of the ambient, but because the air flow is the lowest?
Chapter 7.6 The Weakest Link in Air Cooling
It is the 21st century already. We don't have our personal jet packs yet, and we're still struggling to get heat out of a room with air-conditioning. It turns out that air-conditioning technology is fine--it's the puny humans in the equation that are limiting the air cooling. Where are our robot servants to save the day?