Post by thread:[EE] Copper Heat Sink

Once upon a time, I had a nice DOC on using copper clad boards as heat sinks. Alas, I can't locate it anymore. Does anyone have a guideline for using double-sided 1 ounce copper PCB as a heatsink?? I am going to be using an IRF3170L MOSFET, which has a TO-262 package (about the same size as a TO-220, bit smaller tab) each dissipating about 5 watts. The transistor has a thermal resistance junction to case of 0.75 degree C per watt. Instinct says that we should be able to dissipate 5 watts through a board, but I need to have numbers to back that up. --Lawrence Lile http://www.projsolco.com Electronic Solutions _________________________________________________________________ On the road to retirement? Check out MSN Life Events for advice on how to get there! http://lifeevents.msn.com/category.aspx?cid=Retirement

----- Original Message ----- From: "Lawrence Lile" <spam_OUTlawrencelileTakeThisOuThotmail.com> To: <.....piclistKILLspam@spam@mit.edu> Sent: Monday, January 17, 2005 1:26 PM Subject: RE: [EE] Copper Heat Sink {Quote hidden}> Once upon a time, I had a nice DOC on using copper clad boards as heat > sinks. Alas, I can't locate it anymore. > > Does anyone have a guideline for using double-sided 1 ounce copper PCB as > a heatsink?? I am going to be using an IRF3170L MOSFET, which has a > TO-262 package (about the same size as a TO-220, bit smaller tab) each > dissipating about 5 watts. The transistor has a thermal resistance > junction to case of 0.75 degree C per watt. Instinct says that we should > be able to dissipate 5 watts through a board, but I need to have numbers > to back that up. > > --Lawrence Lile > http://www.projsolco.com > Electronic Solutions Not exactly the answer you are looking for but go to National Semi's web site and look at some of their "Simple Switcher" parts. The ones I've used have several tables and charts comparing the Rtheta of copper clad to heat sinks. Rob Young rwyoungKILLspamieee.org

On 17 Jan 2005 at 13:26, Lawrence Lile wrote: > Once upon a time, I had a nice DOC on using copper clad boards as heat > sinks. Alas, I can't locate it anymore. Infineon have a paper called "Thermal Resistance Theory and Practice" that covers it quite well. There is also an online calculator at http://www.frigprim.com/online/therm_terr.html that may help. Steve. ========================================== Steve Baldwin Electronic Product Design TLA Microsystems Ltd Microcontroller Specialists PO Box 15-680, New Lynn http://www.tla.co.nz Auckland, New Zealand ph +64 9 820-2221 email: .....steveKILLspam.....tla.co.nz fax +64 9 820-1929 =========================================

> Does anyone have a guideline for using double-sided 1 ounce copper > PCB as a heatsink?? I am going to be using an IRF3170L MOSFET, > which has a TO-262 package (about the same size as a TO-220, bit > smaller tab) each dissipating about 5 watts. The transistor has a > thermal resistance junction to case of 0.75 degree C per watt. > Instinct says that we should be able to dissipate 5 watts through a > board, but I need to have numbers to back that up. Very relevant - albeit not the full story. Micrel application hint 17 Designing PC board heat sinks http://www.micrel.com/_PDF/App-Hints/ah-17.pdf Table 5 is of relevance http://www.asat.com/products/library/bg1002.pdf This will be of interest http://www.maxim-ic.com/appnotes.cfm/appnote_number/862 RM

On Tue, 18 Jan 2005 11:47:20 +1300, you wrote: >> Does anyone have a guideline for using double-sided 1 ounce copper >> PCB as a heatsink?? I am going to be using an IRF3170L MOSFET, >> which has a TO-262 package (about the same size as a TO-220, bit >> smaller tab) each dissipating about 5 watts. The transistor has a >> thermal resistance junction to case of 0.75 degree C per watt. >> Instinct says that we should be able to dissipate 5 watts through a >> board, but I need to have numbers to back that up. > >Very relevant - albeit not the full story. >Micrel application hint 17 >Designing PC board heat sinks Unless the PCB is pretty big or you have forced air, I think 5 watts is unrealistic. I start worrying at anything over a watt.

I've noticed that 4-layer PCBs which have at least GND as a plane, seem to handle heat well. Otherwise, PCB's make poor heat sinks. --Bob Mike Harrison wrote: {Quote hidden}>On Tue, 18 Jan 2005 11:47:20 +1300, you wrote: > > > >>>Does anyone have a guideline for using double-sided 1 ounce copper >>>PCB as a heatsink?? I am going to be using an IRF3170L MOSFET, >>>which has a TO-262 package (about the same size as a TO-220, bit >>>smaller tab) each dissipating about 5 watts. The transistor has a >>>thermal resistance junction to case of 0.75 degree C per watt. >>>Instinct says that we should be able to dissipate 5 watts through a >>>board, but I need to have numbers to back that up. >>> >>> >>Very relevant - albeit not the full story. >>Micrel application hint 17 >>Designing PC board heat sinks >> >> > >Unless the PCB is pretty big or you have forced air, I think 5 watts is unrealistic. >I start worrying at anything over a watt. > > > > -- Note: To protect our network, attachments must be sent to EraseMEattachspam_OUTTakeThisOuTengineer.cotse.net . 1-866-263-5745 Tollfree USA/Canada

maybe in a future Intel will release Pentium 23 or something like that, with reduced power dissipation which will cool itself on the mobo ground plane. I can't wait to throw my cpu cooler out the window. Seriously, even if you had forced air through the pcb, it will get dirty and lose a lot of thermal conductivity. I wouldn't rely on that. people expect circuits, especially PIC based which are supposed to be reliable (like in a control system), to not to need active cooling. So i'd avoid that if it were possible On Tue, 18 Jan 2005 00:21:22 +0000, Mike Harrison <mikespam_OUTwhitewing.co.uk> wrote: {Quote hidden}> On Tue, 18 Jan 2005 11:47:20 +1300, you wrote: > > >> Does anyone have a guideline for using double-sided 1 ounce copper > >> PCB as a heatsink?? I am going to be using an IRF3170L MOSFET, > >> which has a TO-262 package (about the same size as a TO-220, bit > >> smaller tab) each dissipating about 5 watts. The transistor has a > >> thermal resistance junction to case of 0.75 degree C per watt. > >> Instinct says that we should be able to dissipate 5 watts through a > >> board, but I need to have numbers to back that up. > > > >Very relevant - albeit not the full story. > >Micrel application hint 17 > >Designing PC board heat sinks > > Unless the PCB is pretty big or you have forced air, I think 5 watts is unrealistic. > I start worrying at anything over a watt. > > > -

>> Once upon a time, I had a nice DOC on using copper clad boards as >> heat >> sinks. Alas, I can't locate it anymore. In case you missed it, you really want to read this http://www.frigprim.com/articels/thermterr_1.html Links from and is related to the page that Steve referred to. RM

So much for instinct. The online calculator posted by Steve throws plenty of water on this idea. Heat sinks it is! --Lawrence Lile Electronic Solutions http://www.projsolco.com > >Unless the PCB is pretty big or you have forced air, I think 5 watts is >unrealistic. >I start worrying at anything over a watt. > >

>So much for instinct. The online calculator posted by Steve throws plenty >of water on this idea. Heat sinks it is! > >--Lawrence Lile Hmm, yes, the water should be enough to make the difference. But I would think heatsinks are less complicated than water cooling. 8^) (Sorry Lawrence, too much time on my hands today) 8^) Lyle -- No virus found in this outgoing message. Checked by AVG Anti-Virus. Version: 7.0.300 / Virus Database: 265.6.13 - Release Date: 1/16/05

Thanks Russel, This paper says that 1.2 watts is the practical limit, and that above 0.8 watts PCB cooling area starts going up rapidly. Anything over a watt (as someone else said) makes one get nervous. This explains why my attempt at cooling a 2 watt resistor one time, by adding some PCB copper area under it, wasn't very effective. I tried it just for grins, but no calculations. --Lawrence Lile Electronic Solutions http://www.projsolco.com {Quote hidden}>From: Russell McMahon <@spam@apptechKILLspamparadise.net.nz> >Reply-To: "Microcontroller discussion list - Public." <KILLspampiclistKILLspammit.edu> >To: "Microcontroller discussion list - Public." <RemoveMEpiclistTakeThisOuTmit.edu> >Subject: Re: [EE] Copper Heat Sink >Date: Tue, 18 Jan 2005 15:38:01 +1300 > >>>Once upon a time, I had a nice DOC on using copper clad boards as heat >>>sinks. Alas, I can't locate it anymore. > >In case you missed it, you really want to read this > > http://www.frigprim.com/articels/thermterr_1.html > >Links from and is related to the page that Steve referred to. > > > > RM >

Lyle, all I can say is: "Aaarrgh" ;-) {Quote hidden}>From: "Lyle Hazelwood" <spamBeGonehaze9844spamBeGonebellsouth.net> >Reply-To: "Microcontroller discussion list - Public." <TakeThisOuTpiclistEraseMEspam_OUTmit.edu> >To: "Microcontroller discussion list - Public." <RemoveMEpiclistTakeThisOuTmit.edu> >Subject: RE: [EE] Copper Heat Sink >Date: Tue, 18 Jan 2005 13:35:55 -0500 > > > >So much for instinct. The online calculator posted by Steve throws >plenty > >of water on this idea. Heat sinks it is! > > > >--Lawrence Lile > >Hmm, yes, the water should be enough to make the difference. >But I would think heatsinks are less complicated than water cooling. > >8^) > >(Sorry Lawrence, too much time on my hands today) 8^) > >Lyle >-- >No virus found in this outgoing message. >Checked by AVG Anti-Virus. >Version: 7.0.300 / Virus Database: 265.6.13 - Release Date: 1/16/05 >

At 11:30 AM 1/18/2005, Lawrence Lile wrote: >So much for instinct. The online calculator posted by Steve throws plenty >of water on this idea. Heat sinks it is! > >--Lawrence Lile >Electronic Solutions >http://www.projsolco.com You mean: "So much for instinct. The online calculator posted by Steve throws plenty of STEAM on this idea. Heat sinks it is!" (same excuse as Lyle <grin> ) dwayne -- Dwayne Reid <dwaynerEraseME.....planet.eon.net> Trinity Electronics Systems Ltd Edmonton, AB, CANADA (780) 489-3199 voice (780) 487-6397 fax Celebrating 20 years of Engineering Innovation (1984 - 2004) .-. .-. .-. .-. .-. .-. .-. .-. .-. .- `-' `-' `-' `-' `-' `-' `-' `-' `-' Do NOT send unsolicited commercial email to this email address. This message neither grants consent to receive unsolicited commercial email nor is intended to solicit commercial email.

D'oh!!! > >You mean: "So much for instinct. The online calculator posted by Steve >throws plenty of STEAM on this idea. Heat sinks it is!" > >(same excuse as Lyle <grin> ) > >dwayne > >-- >Dwayne Reid <EraseMEdwaynerplanet.eon.net> >Trinity Electronics Systems Ltd Edmonton, AB, CANADA >(780) 489-3199 voice (780) 487-6397 fax > >Celebrating 20 years of Engineering Innovation (1984 - 2004) > .-. .-. .-. .-. .-. .-. .-. .-. .-. .- > `-' `-' `-' `-' `-' `-' `-' `-' `-' >Do NOT send unsolicited commercial email to this email address. >This message neither grants consent to receive unsolicited >commercial email nor is intended to solicit commercial email. > >

On Tue, 18 Jan 2005, Lawrence Lile wrote: > Thanks Russel, This paper says that 1.2 watts is the practical limit, and > that above 0.8 watts PCB cooling area starts going up rapidly. Anything over > a watt (as someone else said) makes one get nervous. This explains why my > attempt at cooling a 2 watt resistor one time, by adding some PCB copper area > under it, wasn't very effective. I tried it just for grins, but no > calculations. You can cool it by mounting it on 'standoff' tubes (brass usually). These can be finned at the top. There is also a 'mushroom' style heatsink that can be screwed directly onto a TO220 etc mounting hole. I saw it in some equipment and have been looking for something like this ever since. Another trick is to mount the board with the parts flush with the bottom (metal) plate of the casing, with some thermal compound for contact. Peter

There's a PCB layout technique called "via stitching" that is basically a number of regularly spaced plated-through vias (holes) in the pad's area that connects copper between the two sides (assuming a double-side PCB; it was also work multiple layer boards). During board assembly, the via are filled with paste (you could also hand solder the vias, although I not sure if wave soldering would fill all of the vias). This adds additional metal to the heat sink. I use this method on all of my voltage regulators and power transistors (TO-220, TO-252/D-PAK, TO-263, SOT-23, etc.). If you'd like to see a photo of what this looks like, please let me know. Douglas Wood {Original Message removed}

It's also a pain to solder/desolder small components to/from. Makes for a very good heat sink. Computer motherboards use these liberally around the switching regulators that power the CPU. The negative end of the filtering caps (you know, the caps that fail and have to be replaced?) usually ends up in one of these via stitched areas. But it's not so bad that I'm going to throw away essentially free computers either... -Adam Douglas Wood wrote: {Quote hidden}> There's a PCB layout technique called "via stitching" that is > basically a number of regularly spaced plated-through vias (holes) in > the pad's area that connects copper between the two sides (assuming a > double-side PCB; it was also work multiple layer boards). During board > assembly, the via are filled with paste (you could also hand solder > the vias, although I not sure if wave soldering would fill all of the > vias). This adds additional metal to the heat sink. I use this method > on all of my voltage regulators and power transistors (TO-220, > TO-252/D-PAK, TO-263, SOT-23, etc.). If you'd like to see a photo of > what this looks like, please let me know. > > Douglas Wood > > {Original Message removed}