Talk:Heat sink

From Wikipedia, the free encyclopedia
Jump to: navigation, search

Scope of this article[edit]

I just wanted to comment that heat spreading/dissipating devices are by no means limited in application to microelectronics (or, to put it bluntly, PC microprocessors). I would try to improve this article's scope myself, but my focus lies elsewhere right now. This is merely a suggestion for anybody who might decide to improve the article. -- uberpenguin 20:06, 11 March 2006 (UTC)

Yeah, you're absolutely right. In thermodynamics, the term is used to mean somewhere that heat can be transferred to. In engineering, things like engines, power plants, these all have heat sinks, for example a power plant might use a nearby lake as a 'heat sink'. I think the PC microprocessor usage probably came after all this, and is a fairly specific example. Brendanfox 08:02, 13 October 2006 (UTC)
Not only that, but according to a recent paper in Science, an account of which is here, toucans use their bills as heat sinks. Heat sinks in nature seem to be definitely worth going into the article. Lexo (talk) 10:06, 24 July 2009 (UTC)
Very true, sounds like a PC enthusiast has written the whole article, and so many pictures of various CPUs are absolutely redundant. How about pictures of the heatsink on some Audio Power amplifier or a SMPS switching driver transistor?Sub40Hz (talk) 19:55, 20 October 2009 (UTC)

I'll say it again - hopefully someone who knows the topic can do something. I came to this article looking for info on the theories of how heat sinks (colloquial or not) function in general. Heat transfer. This is "how to design a CPU heat exchanger" by and large. Something more about how this concept functions in general would be much appreciated. Jjdon (talk) 20:29, 2 August 2010 (UTC)

Removed pending citation[edit]

It is claimed that some brands of thermal grease that are advertised as containing silver or silver oxide actually contain neither, most notably that of CompUSA.{{citeneeded}}

- 16:19, 15 March 2006 (UTC)

The most common filler in the thermal greases is aluminum oxide. 09:04, 31 March 2006 (UTC)

I nowikied the template so that this talk page does not show up in the articles lacking sources category. -- Kjkolb 18:42, 6 April 2006 (UTC)


Heatsink or heat sink? -- Frap 11:24, 21 May 2006 (UTC)

It's "heat sink" in my dictionary. --Charles Gaudette 00:37, 8 June 2006 (UTC)

Yes it is 2010 now and I ask again; is it one word "Heatsink" or two "Heat Sink"? I have seen both with manufacturers or in documentation... Are you sure it is just one word? Seems both are used -- Surf 10:20, 03 Sep 2010 (MT) —Preceding unsigned comment added by (talk)

Removed pending citation #2[edit]

"Aluminum has the significant advantage that it can be easily formed by extrusion, thus making complex cross-sections possible."

This doesn't ring true with me — copper versus aluminum — they are both very malleable metals. I further asked someone with aerospace materials management experience, and they called it bogus too. If it can be cited to a reliable source, then by all means, put it back in the article. --Charles Gaudette 01:45, 8 June 2006 (UTC)

Not sure whether this was the original authors source, but a source confirming that claim is Groover's Fundamentals of Modern Manufacturing, Third Edition, p419, which says "Aluminum is probably the most ideal metal for extrusion (hot and cold), and many commercial aluminum products are made by this process...", a few pages later on p424 as an example of extrusions ability to form complex cross-sections is a picture of an aluminum heat sink. Brendanfox 23:46, 18 November 2006 (UTC)
I still have a problem with "significant advantage"; for instance extruded copper wire, and the tools and art of the Bronze Age. Does anyone have the ISBN for that book? --Charles Gaudette 01:24, 19 November 2006 (UTC)
The ISBN is 0-471-74485-9 [1]. The book also mentions "copper, magnesium, zinc, tin and their alloys" as typical metals used in extrusion, but it does seem to draw attention to aluminum as being "the most ideal". Unfortunately it doesn't give any particular reason for this. We could change the wording, tone it down a bit, but the basis of the claim (i.e. that aluminum extrudes better than copper) appears to be valid. Brendanfox 02:00, 19 November 2006 (UTC)
(First off, understand I am not fighting against you, just trying to work our way to a better "Heat sink" article.) Even if aluminum is easier to extrude, the underlying thrust implies that aluminum heat sinks are a cheaper manufacturing process (I'll give aluminum a raw materials cost lower than copper). Thanks for the ISBN. I'll have to do some research to catch-up to you, ... American Metals Association, hmm... --Charles Gaudette 19:04, 20 November 2006 (UTC)
No probs. You might also be interested in the graph from slide 47 of this: [2] - Brendanfox 04:01, 21 November 2006 (UTC)

Copper vs Aluminum for heat sinks: it is correct that warm copper can be as readily extruded as aluminum, but copper remains malleable (soft) after such extrusion. This is a positive advantage for some applications, but not for the shapes typically used for air-cooled heat-sinks. Al is also appreciably less dense than Cu, and this eases issues with vibration. Cu heat sinks are (typically) formed from sheet-like material, which allows less malleable product to be created.

The heat sinks cited are widely used with CPUs. But heat sinks in the form of flat plates are used with printed circuit boards. These are also called "cold plates" and "thermal planes". Heat spreaders are also closely related to heat sinks.

I suggest that the title be changed to "Heat Sinks and Heat Spreaders".

In addition to to copper and aluminum, other traditional heat sink materials include copper/tungsten, copper molybdenum and copper-Invar-copper.

In recent years, there have been important new heat sink materials, including aluminum particle-reinforced aluminum (Al/SiC), carbon fiber-reinforced aluminum, highly-oriented pyrolytic graphite, natural graphite, and diamond particle-reinforced silicon carbide, among others. The latter has been used in IBM servers. CarlZweben 22:18, 5 March 2007 (UTC) CarlZweben

Two considerations for improving article[edit]

First, consider noting that the term "heat sink" is a misnomer and that this expression is not found in heat transfer textbooks. A heat sink is a collection of extended fin surfaces (Hot Air Rises and Heat Sinks, Kordyban). Second, consider distinguishing between bonded fin heat sinks and extruded fins, augmented bonded fin heat sinks, etc.

P.K., August 6, 2007

I'm an old-timer, a retired lifelong electronic tech with a couple of years as an associate editor at Electronic Design magazine, and the term "heat sink" has continually been a very mild irritant; it's borderline slang (embedded update follows), although by very wide usage, I'd say that nearly everybody concerned knows what it means. I was mildly sorry to see the term as an article title; it seems somewhat too close to slang.

Update, 20110907 (Sept.): I read a following comment that rebuts effectively the idea that "heat sink" is a slang term. This would not be an isolated instance, it seems to me, of what might have originally been slang becoming an accepted and well-defined term, although at the moment, I can't think of a good example. Live and learn! Regards, Nikevich (talk) 10:19, 7 September 2011 (UTC) (Now using DHCP, so my numeric IP is subject to change.) (End update)

I never studied thermodynamics, but it seems to me that in formally-defined study of heat transfer, there are heat sources and heat sinks. When concentrating on the transfer of heat, it only distracts to give any details of the nature of the sink, it seems to me. Therefore, in such academic discussions, an heat sink is best considered simply as an abstract entity. (The ultimate heat sink is outer space, I'd say, with the atmosphere and maybe bodies of water as intermediaries.)

My technical life spanned octal-based tubes to VLSICs, and when power transistors came into use, calculations of thermal resistance from junction to case and from case to ambient (typically via an heat sink) were (and, I'm sure still are) commonplace.

The term is catchy, so to speak, which has made it popular, like "baud" or "Watts RMS", a term that is technically incorrect*, despite apparently widespread belief otherwise.

  • In dimensional analysis of "Watts RMS", one finds contradictions, iirc. There are definite problems when one attempts rigorous analysis. Quite a few years ago, there was a discussion in the Journal of the Audio Engineering Society about "Watts RMS". It asked, (paraphrased) "If there were such a thing as 'Watts RMS', how would it be defined?". What was interesting is than no two numerical definitions agreed! [Updated : Nikevich (talk) 10:19, 7 September 2011 (UTC)]

Regards, Nikevich (talk) 02:25, 13 October 2009 (UTC)

Heat Sink Color[edit]

I have heard that Black Heat Sinks work better that non anodized or painted. Yet, I can't find conclusive evidence either way.

I suspect that there has been a misunderstanding of black bodies being ideal sources of thermal radiation. I think that in reality, the ideal source property describes the lack of environmental interference when making measurements.

The Anodize article does state that anodized aluminum has lower thermal conductivity than plain aluminum. Anodized aluminum would be protected from corrosion, which may be more detrimental.

Does anyone have a source with the answer?

Here is what I have found so far: From

  How does the color of a heat sink impact its thermal performance?
  In natural convection a black or dark colored heatsink will perform 3% to 8% better
  than an aluminum heatsink in its natural silverish color. This is due to the fact
  that dark colors radiate heat more efficiently.
  In forced air applications, surface color does not increase a heat sink's performance
  due to the increase in convection. The color would provide cosmetic benefits only.


  Does heatsink color affect heat dissipation?
  Black is the best thermal body in terms of being a absorber or emitter. Let's assume
  a vacuum situation, if the surface "A" of a black heatsink is totally covered at T1
  (temperature 1) by another black body at T2 (temperature 2), the black heatsink would
  get the energy reflected from another black body at reflection energy Ad(T14-T24) that
  we call "Stefan-Boltzmann" law of thermal radiation, here refers to Stefan Boltzmann
  Constant, it is 5.6697 x 10-8 W/m2?K4. Therefore if based on above, T1 is the
  temperature obtained from the black heatsink onto CPU, T2 is the ambient temperature
  around CPU. So if T14-T24 is a positive value, we know black is the best heat
  dissipation transistor since there is no thermal source around CPU inside PC case.


  Do not paint a heat sink. Most heat sinks are anodized aluminum. Painting a heat sink
  (especially if it's a thick coat of paint) is like putting a blanket on the amplifier.
  If you absolutely must paint the heat sink, use the lightest, thinnest coat of paint possible.


Eet 1024 20:59, 13 November 2007 (UTC)

Black does radiate better than other colors. I remember doing experiments on this in my undergrad degree. However color is not the only concern here. Paint usually is a poor heat conductor (especially household paints, this is why there are special heat sink paints), especially compared to metals, and therefore painting a heat sink using regular paints actually hinder heat dissipation. You enhance radiative dissipation (due to blackbody radiation), but hamper convective dissipation (due to air flow). Using very thin layers of paint means you get the full benefits of having a black color, and minimizes that harm you do to convective dissipation. I hope that clarifies. Headbomb {ταλκWP Physics: PotW} 13:37, 10 July 2008 (UTC)

I think the websites are saying other things equal black is better, which of course makes sense since black radiates more (although what matters is being black in the infrared, which is far from the same thing as being visibly black). But if the way you get it to be black is by adding a layer of paint that harms the thermal conductivity, it's not worth it; radiation is usually a very small proportion of heat transfer. The radiationsink website is basically saying that in natural convection heat sinks, radiation is a small (~10%) contribution to heat transfer, whereas in forced-air heat sinks, it's an utterly negligible contribution. --Steve (talk) 16:07, 10 July 2008 (UTC)
I agree with Steve Byrnes. Black in the infrared is good, but not at the cost of an insulating layer of paint. JRSpriggs (talk) 21:43, 14 July 2008 (UTC)
The best thing is to maximize the surface area. If you cannot make it wider, then make it corrugated. JRSpriggs (talk) 18:38, 15 July 2008 (UTC)
Found it: Kirchhoff's law of thermal radiation. It all depends on the application, as stated in the radianheatsinks page above. Will consider adding to the Heat_sink#Performance section. Thanks, Eet 1024 (talk) 05:01, 23 September 2008 (UTC)


Do any heatsink use material other than alluminum and copper? Since Silver, for example, has thermal conductivity greater than copper, some manufacturer might be interest in this material. —Preceding unsigned comment added by (talk) 14:43, 7 August 2008 (UTC) can a heat sink ever be made based on "maxwell's demon"? (talk) 17:51, 7 October 2008 (UTC)

Silver would not be used since it is a precious metal. For the same reason that gold is not used. Also, if you are using air, the major thermal resistance is the convection resistance. It is worth more to optimise the air side of the heat sink than to use a very expensive material. For the same reason that copper is used, but only if really necessary. Dtc5341 (talk) 10:17, 20 January 2010 (UTC)

The material aspect of sink performance is really glazed over in very little detail about how material selection affects the properties of the sink. Also, the pricing and historical trending information on material cost (aluminum vs copper) seems too detailed for the scope of this article. A simple, "copper tends to costs #-# times as much as aluminum" with a citation, is probably sufficient. (talk) 17:50, 20 September 2010 (UTC)

Everybody seems to be missing the primary purpose of the heat sink, which is to increase the heat mass of the system and thereby slow down temperature rise time. So the ideal material has a high heat capacity first, and high thermal conductivity second. For metals, these properties are related, for other materials, they are independent. BTW, Aluminum passivates, Copper and Silver do not... Over extended service, Aluminum may be superior. —Preceding unsigned comment added by (talk) 17:46, 5 May 2011 (UTC)

Proposed major revision of article[edit]

To whom it may concern,

I am planning a major update of this article (heat sinks) and would like some constructive comments. Bear in mind that everyone has his or her own opinions, etc. Find below the introduction that I am going to use and proposed article outline. The article will be updated in segments.

1 Introduction

The term heat sink is colloquially used to describe two objects: 1. The ambient air, river or sea water, or 2. Finned metal object

This article will focus on the later of the two objects. Examples of heat sinks are the heat exchangers on a refrigerator, air conditioning systems, the radiator (also a heat exchanger) on a car. Heat sinks are also used on, for example, cooling of lasers, electronics and light emitting diodes.

A heat sink does not have a “magical ability to absorb heat like a sponge and send it off to a parallel universe” [1]. To explain how a heat sink works, the heat transfer theory of a heat sink will be discussed first. Discussing the heat transfer theory of a heat sink will also seek to answer the question of: “When is a heat sink not a heat sink?”

The term heat sink is never used in a heat transfer text book [1]. When the text books do mention the objects commonly referred to as heat sinks, they use the term extended surfaces. This also relates the “magical ability” of heat sinks, as mentioned earlier. A detailed theoretical model of a heat sink will be discussed based on theory and heat transfer coefficient correlations from heat transfer text books, and fluid dynamic theory.

2 Basic heat sink heat transfer theory model

3 Methods to determine heat sink thermal performance

3.1 Heat transfer theoretical model

3.2 Experimental data

3.3 Numerical data

4 Design factors which influence the thermal performance of a heat sink

4.1 Material

4.1.1 Fin efficiency

4.1.2 Spreading resistance

4.2 Fin arrangements

4.2.1 Pin fin

4.2.2 Straight fin

4.2.3 Cross cut

4.2.4 Flared

4.3 Fin spacing

4.3.1 Natural convection

4.3.2 Forced convection Ducted Unducted

5 Engineering applications

5.1 Processor/Micro processor cooling

5.1.1 Attachment methods methods

5.1.2 Thermal interface materials

6 References

[1] Kordyban, T., 1998, Hot air rises and heat sinks – Everything you know about cooling electronics is wrong, ASME Press, NY

If there are no comments on my proposed updates, I'm going to update the article tomorrow (26 March 2010 0700 CET) —Preceding unsigned comment added by (talk) 10:12, 25 March 2010 (UTC)
Major revision updated. Comments please. I still want to update the LED section... I know of material, but need to sort out the copyright issues. Dtc5341 (talk) 05:38, 26 March 2010 (UTC)

Support for your view: Copper vs Aluminum: although it is correct that warm copper can be as readily extruded as aluminum, such copper remains malleable (soft) after such extrusion. This is a positive advantage for some applications, but not for the shapes typically used for air-cooled heat-sinks. Al is also appreciably less dense than Cu, and this eases issues with vibration. Cu heat sinks are (typically) formed from sheet-like material, which allows less malleable product to be created. Also, I think the article could be further improved by at least mentioning heat pipes in the main text, rather than consigning them wholly to one of the references. Finally, it could be helpful to provide an empty paragraph entitled "non-electronic applications" or some such - an open invitation to other fields. PhysicistQuery (talk) 13:59, 18 June 2010 (UTC)

Peer review[edit]

Colloquial term?[edit]

I have removed the claim that heat sink is a colloquial term and the claim that it is never used in academic books. This is demonstrably untrue. Google books gets well in excess of 600,000 hits and the first listed is a a book actually called Heat Sink. Also on the first page of results is Horowitz's book which is widely used on Wikipedia as a reliable source by some electronics editors. Thinking the term may possibly be avoided in academic journals I tried Scholar and got in excess of 400,000 hits. These include multiple hits for the Journal of Heat and Mass Transfer (example) who one would think, if anyone, would be avoiding a coloquial term. There are also many hits for the prestigous IEEE Transactions.

It might be possible to say that the term heat sink as a component arises from the theoretical thermodynamic use of the term where it is opposed to a heat source, although I have no source to hand giving the etymology. Often, ideal sources and sinks are considered that can supply or absorb any quantity of heat without changing temperature, much as an ideal voltage source can supply or sink any amount of current.

I am not sure exactly what the source, Kordyban, has said on this, but apparently the book is written in a humorous style. Perhaps someone can provide the exact quote, but it should probably not be taken too serioulsy. SpinningSpark 10:08, 20 June 2010 (UTC)

Quick clarification[edit]

"But one of the advantage of copper is that it is less loud than aluminium." I am a lay on the topic; I read this and think of loudness as referring to noise level. Is that how it is meant? If so, why would either material make any noise? -- (talk) 02:38, 2 December 2010 (UTC)

The other direction[edit]

Primus EtaPower billy.jpg

How about heatsinks that are designed to operate in the opposite direction, i.e. transfer heat from the air to a cold object? An example would be this picture which appears in the Billycan article, a cooking pot with a heatsink welded to its base to trap as much heat from the stove as possible. Rwxrwxrwx (talk) 09:52, 26 December 2010 (UTC)

This is not genuinely a heat sink, it is a heat store. The definition of heat sink implies that there is only one direction of heat flow. Heat goes in from the fuel, but the metal is acting as a heat source as the heat goes out to the food. This is similar to the old-fashioned cooking pots with thick copper bottoms. At the very least, this need a reliable source (preferably journal publication) decribing this as a heat sink before inclusion is warranted. SpinningSpark 07:27, 3 March 2011 (UTC)

Linking to heat sink article[edit]

"Can I link to my website here? I have a lot of good heatsink illustrations." — Preceding unsigned comment added by Davidfarzam (talkcontribs) 23:26, 2 March 2011 (UTC)

No, we do not usually approve of links to uncontrolled private sites. You can, however, upload images if you are prepared to release them under a free licence. SpinningSpark 07:02, 3 March 2011 (UTC)
I have also removed the batch of three external links recently added. The first one is a calculator, which is not needed as the article already contains the design formulae. Calculators are often created by commercial web sites merely in order to generate links to their site rather than a genuine attempt to be useful and are particularly dubious when they fail (as here) to explain details of the calculation. The second one timed out on me and appears to be dead, at least temporarily (WP:EL#EL7, WP:EL#EL16). The third one is a power point presentation which would be most useful as an aide-memoire to someone who actually attended the lecture, but not here. The material is already adequately covered here and anything that isn't can easily be added. See WP:EL, especially WP:EL#EL1. SpinningSpark 08:10, 3 March 2011 (UTC)

Equation 2 Defination[edit]

In the section: Basic heat sink heat transfer principle there is an equation,

\dot{Q} = \frac{T_{hs} - T_{air,av}}{R_{hs}} (2)

Where the value of R is not defined as to what it is. There should be some clarification. — Preceding unsigned comment added by Dabozz88 (talkcontribs) 13:33, 3 October 2011 (UTC)

Someone please fix this; its frustrating to come across explanations that you have to decipher yourself. (talk) 01:26, 2 December 2013 (UTC)Anonymous

Undeclared Variable in Section 1[edit]

In the basic heat sink heat transfer principle section, the variable c_{p,in} is not declared, and does not explicitly appear in the associated figure. Equation numbering is also done manually rather than using the LaTeX-based markup for equation numbering, which may lead to future inconsistencies. — Preceding unsigned comment added by Nummify (talkcontribs) 17:06, 17 February 2012 (UTC)

Undeclared Variable in Section 4.1[edit]

Lot's of variables used in equations from 5 to 13 are not declared, so these are not easy to use. — Preceding unsigned comment added by (talk) 06:51, 12 June 2014 (UTC)

m n[edit]

what are m and n called, what are their units??????????? — Preceding unsigned comment added by (talk) 14:27, 10 November 2013 (UTC)