|
answers to those doggone thermal design questions
by tony kordyban
mr. cool,
sparky is working on round 2 of his board and wants me to redo the heat sink for the big gate array. the original heat sink turned out to be overkill, because his original power estimate was way too high. by too high, i mean he told me it was 10 watts, and my cfd analysis told me i needed a big, honkin' heat sink to keep the case temperature down to 90c. when we finally tested it, the case was only 60c. he would like to recover the space eaten up by the heat sink to stick in some more components he forgot the last time. i asked sparky for a better power estimate for the gate array, but he said that the board design was too complex to calculate it. he asked if i could get the power dissipation from the thermocouple measurements and the package thermal resistance. what steps would you take to optimize this heat sink?
fred from red river
dear fred,
what a surprise! the circuit designer can't figure out the power dissipation of his own custom designed chip. but sparky isn't on the wrong track with his suggestion.
when you have nothing better, i think it is perfectly reasonable to come up with your own power estimate for a component based on your temperature measurements. i have done it myself when i can't trust the values i have to work with.
but i wouldn't follow sparky's idea to the letter. i think what he wants you to do is use the thermal resistance of the package, such as rj-a (the resistance between junction and ambient) to estimate the power value based on this relationship:
tj = ta + power x rj-a
where tj is the junction temperature and ta is the ambient temperature. there is one major algebraic problem with this method. you can't solve a problem that has one equation and four unknowns.
technically, rj-a isn't unknown. you can probably find it in a data sheet. but its value is bogus. to find out why you'll have to read all my previous columns.
you don't really know tj or ta either, because in your test you probably only measured tcase and the inlet air temperature, which are not even close to the same things.
since you say you analyzed this board with a cfd tool, that is the gizmo i would use to back out a better value of power. you had a case temperature prediction from cfd based on 10 watts, but it was much higher than what you measured. if you think your cfd model is reasonably representative of the real board in most other ways, then just guess a lower power value and re-run the cfd too. if the temperature is still too high, make the guess lower and run it again. it shouldn't take more than 3 guesses to zero in on a pretty good power estimate.
here is a hint to help you guess faster: (tcase - t air) is directly proportional to the power, more or less.
this power value is probably a lot more accurate than any "estimate" that sparky can make by counting up gates. but even so, since it can't be verified by some other method, i wouldn't trust it 100%. it is plenty accurate enough to help you optimize the heat sink. the proof of how good it is will be when you build the new board and test it.
and, of course, it wouldn't hurt to throw in a safety factor to account for part-to-part variation in power. just how big a safety factor depends on the chip technology. sparky, or the application engineer for the chip maker could give you a handle on whether you need 10%, 50%, or even 100% safety factor in your power estimate. don't laugh at 100%. once i had to work with a gallium arsenide chip that varied from 0.8 watts to 2.8 watts just due to manufacturing variation. now there was a heat sink that looked like overkill a lot of the time.
to mr. know-it-all:
why is everyone inventing a new term for every thermal event of one kind or other? for example, i have seen such terms as thermal resistance, thermal inertia, thermal impedance, thermal conductance, thermal inductance, thermal intrusion, etc. where can someone with a double digit iq find simple and correct definitions for these confused, complicated and head-twisting terms? i hope i am not blowing hot air on cool stuff.
thermally confused from thermopolis
dear confused,
i never heard of these terms either, until i got involved in electronics cooling. when it comes to the electronics industry, it's like steve martin's complaint about france: "those people have a different word for everything!"
you have come to the right place, though. i will try to give you a simple definition of the terms you are confused by, although i can't promise to give you a reason why we need them when perfectly good heat transfer terms from the days of isaac newton still work.
|
thermal resistance
|
1. the characteristic of a system that leads to a temperature difference (t) when heat (q) flows through it, according to this equation: q = t/r. (in 1-d conduction q = ka/x t, so r = x/ka. but in convection, r = 1/ha, so there is no single complete definition.)
2. a way of hiding the true nature of a 3-d network of heat transfer paths. example: rj-a lumps together the solid conduction through a component package, with the convection and radiation from the surfaces to the environment, into a single, meaningless value.
|
|
thermal inertia
|
the mythical ability of a component to continue to go up in temperature even after the power has been shut off
|
|
thermal impedance
|
same as thermal resistance, but may have extra confusing terms used in transient analysis; from impede, as in "this term impedes my ability to calculate an accurate junction temperature".
|
|
thermal conductance
|
the reciprocal of thermal resistance; usually, in solid conduction equal to ka/t, but if that's all it was then we wouldn't need a new word. so it, too, usually lumps in convection and radiation terms and treats them as if they are all a single entity.
|
|
thermal inductance
|
never heard that one, but it wouldn't surprise me if somebody was using it someplace.
|
|
thermal intrusion
|
i don't know. you'll have to use your imagination on that one.
|
but, hey, that gives me an idea. how about if we start making up our own thermal terms and use them to confuse the electronic designers and their managers and have a little fun at the same time? here are some that occurred to me in the span of five minutes.
|
thermal incontinence
|
the behavior of the highest power component on the board, in which it dumps its waste heat all over the local environment, causing all the neighboring components to become hot, too.
|
|
thermal impecuniosity
|
the paradoxical complaints from the engineering manager that you increased the material cost of a board by $15 when you added a heat sink, even though it protects a $2,000 custom chip from burning up.
|
|
thermal incandescence
|
the ability of a component to change from an electronic device to an optical one when too much current passes through it and it starts to glow in the visible light spectrum.
|
|
thermal jurisdiction
|
whose job is it to add the heat sink to the bill of materials? it's not an electronic part, and not really a mechanical part. and who has to write the spec for thermal grease?
|
|
thermal prescience
|
that creepy feeling you get when you see a new component for the first time and you just know it's going to give you trouble.
|
|
thermal intransigence
|
sometimes called "thermal persistence"; the unexplained ability of a thermal problem to not respond to anything you do to make it go away. no matter how big a heat sink you apply, no matter how big a fan you point at it, the part temperature won't go below its temperature limit. the only known remedy is to bribe or blackmail the appropriate authority into canceling the project (see thermal impropriety).
|
about 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 20 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 b.s. in mechanical engineering from the university of detroit and a master’s in mechanical engineering from stanford. in those 20 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 at https://www.amazon.com/hot-air-rises-heat-sinks/dp/0791800741. this advice column is an extension of that educational effort.
|
Facebook
Twitter
Google+
Linkedin
Reddit
