Post by thread:[EE]: Heatsink source

On Mon, 1 Jul 2002, Sean H. Breheny wrote: >Hi all, > >I'm looking for a cheap source of large heatsinks. I need something with a >large flat side (5 inch by 8 inch or so), and I only need one of them. I >have checked Digikey and some surplus places and haven't found anything >that large yet. Any ideas? How do you know that you need such a large one ? Would a fan help ? (I'm sure it would). Heatsinks are rated by Rth/Watt not by inches and the heat transfer capacity depends a lot on the construction. Now for some pratical tips: Would 'computer' CPU heatsinks help ? They are built to stay under 70C for up to 70W dissipation. At 100C you could double the dissipation with the same fan. Would mass purchased aluminum profile (rectangle or U) help ? Again think about using a fan. I once made a large experimental PSU by screwing lots of TO220 transistors onto the aluminium/magnesium chassis scrapped from defunct equipment. I had 30 transistors in parallel on it (and the respective emitter resistors). hope this helps, Peter -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

Hi Peter et al, Thanks for the great suggestions, I'll check them out. Yes, I am using a fan. Yes, I already tried making one myself out of some scrap aluminum here, but the Al was too thin a gauge and the heat wasn't conducted to the fins as much as it should be (I also only made fins by bending the edges up so there weren't enough fins). I need to dissipate about 120W. The device is a microwave power amplifier which is 5inches by 8 inches. The thermal conductivity of the case (appears to be some brass-type alloy, at least it is that color) doesn't seem to be that great because you can definitely feel hot spots on it while it is running. That's why I want the heat sink to cover the whole side of the case, even if that isn't really necessary to dissipate the heat. Sean At 06:26 PM 7/2/2002 +0300, you wrote: {Quote hidden}>On Mon, 1 Jul 2002, Sean H. Breheny wrote: > > >Hi all, > > > >I'm looking for a cheap source of large heatsinks. I need something with a > >large flat side (5 inch by 8 inch or so), and I only need one of them. I > >have checked Digikey and some surplus places and haven't found anything > >that large yet. Any ideas? > >How do you know that you need such a large one ? Would a fan help ? (I'm >sure it would). Heatsinks are rated by Rth/Watt not by inches and the heat >transfer capacity depends a lot on the construction. Now for some pratical >tips: > >Would 'computer' CPU heatsinks help ? They are built to stay under 70C for >up to 70W dissipation. At 100C you could double the dissipation with the >same fan. > >Would mass purchased aluminum profile (rectangle or U) help ? Again think >about using a fan. > >I once made a large experimental PSU by screwing lots of TO220 transistors >onto the aluminium/magnesium chassis scrapped from defunct equipment. I >had 30 transistors in parallel on it (and the respective emitter >resistors). > >hope this helps, > >Peter > >-- >http://www.piclist.com hint: The list server can filter out subtopics >(like ads or off topics) for you. See http://www.piclist.com/#topics ------------------------------------------- Introducing NetZero Long Distance Unlimited Long Distance only $29.95/ month! Sign Up Today! http://www.netzerolongdistance.com -- http://www.piclist.com hint: The list server can filter out subtopics (like ads or off topics) for you. See http://www.piclist.com/#topics

On Tue, 2 Jul 2002, Sean H. Breheny wrote: > Thanks for the great suggestions, I'll check them out. A couple of HS from the old Pentium II's might help, they're probably about 3x6. Also, see if you can find anyplace locally that scraps old computer equipment... chances are you'll find BIG heat sinks if you can find 70's/80's vintage IBM gear. I don't know how many tons of aluminum we recycled from old tape drives and such. 3420 reel motor control boards had about 5# of heat sink alone... Dale Oh - did I mention checking commercial radio repair places, or service departments of 2-way sales joints? Ditto anywhere that sells/services PBX or phone equipment, copiers... you get the idea. You'd be amazed at what you can get for the asking if you talk to a geek and let him know you're a geek, too. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

>I need to dissipate about 120W. The device is a microwave >power amplifier which is 5inches by 8 inches. The thermal >conductivity of the case (appears to be some brass-type >alloy, at least it is that color) doesn't seem to be that >great because you can definitely feel hot spots on it while >it is running. That's why I want the heat sink to cover the >whole side of the case, even if that isn't really necessary >to dissipate the heat. I would almost guarantee that the case is AL with a chemical finish known as alodising. It comes out a pale brass colour. What you need to do is get a catalogue from a dealer that has lots of profiles in it. I almost guarantee that they will have some profiles designed for heatsinks. Get some lengths of this made in an alloy suitable for anodising (the standard alloy they supply may not be suitable) and have it cut to length and anodised black. Once you have a length like this you can bolt a transistor to it along with a pair of thermocouples. Set the transistor on a power supply so it is acting as a load, and drawing a known current. Let the temperature of the transistor and heatsink stabilise, and using the temperature difference between the transistor and heatsink, along with the power dissipation you can work out what your thermal resistance is, and so what temperature your microwave unit will run at. I went through this loop some years back when the company I worked for had a heatsink profile done, and we ended up with something like 0.6 C/Watt :) P.S. I would look seriously at having the active devices mounted straight on the heatsink rather than bolting the side panel to the heatsink. This will remove a potentially high thermal resistance from the path - how flat/well mated to the heat sink do you think the side panel will be - there will always be cavities which will represent extra thermal resistance no matter how much thermal paste you put on there. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Should the black anodizing be sanded off where the semiconductor makes contact? -----Original Message----- From: Alan B. Pearce [spam_OUTA.B.PearceTakeThisOuTRL.AC.UK] Sent: Wednesday, July 03, 2002 4:14 AM To: .....PICLISTKILLspam@spam@MITVMA.MIT.EDU Subject: Re: [EE]: Heatsink source >I need to dissipate about 120W. The device is a microwave >power amplifier which is 5inches by 8 inches. The thermal >conductivity of the case (appears to be some brass-type >alloy, at least it is that color) doesn't seem to be that >great because you can definitely feel hot spots on it while >it is running. That's why I want the heat sink to cover the >whole side of the case, even if that isn't really necessary >to dissipate the heat. I would almost guarantee that the case is AL with a chemical finish known as alodising. It comes out a pale brass colour. What you need to do is get a catalogue from a dealer that has lots of profiles in it. I almost guarantee that they will have some profiles designed for heatsinks. Get some lengths of this made in an alloy suitable for anodising (the standard alloy they supply may not be suitable) and have it cut to length and anodised black. Once you have a length like this you can bolt a transistor to it along with a pair of thermocouples. Set the transistor on a power supply so it is acting as a load, and drawing a known current. Let the temperature of the transistor and heatsink stabilise, and using the temperature difference between the transistor and heatsink, along with the power dissipation you can work out what your thermal resistance is, and so what temperature your microwave unit will run at. I went through this loop some years back when the company I worked for had a heatsink profile done, and we ended up with something like 0.6 C/Watt :) P.S. I would look seriously at having the active devices mounted straight on the heatsink rather than bolting the side panel to the heatsink. This will remove a potentially high thermal resistance from the path - how flat/well mated to the heat sink do you think the side panel will be - there will always be cavities which will represent extra thermal resistance no matter how much thermal paste you put on there. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

"Scott M. Thomas" <scottKILLspamDBT.NET> wrote: > Alan B. Pearce <.....A.B.PearceKILLspam.....RL.AC.UK> wrote: > > What you need to do is get a catalogue from a dealer that has lots of > > profiles in it. I almost guarantee that they will have some profiles > > designed for heatsinks. Get some lengths of this made in an alloy suitable > > for anodising (the standard alloy they supply may not be suitable) and have > > it cut to length and anodised black. > Should the black anodizing be sanded off where the semiconductor makes > contact? The black anodizing is purely cosmetic and serves no useful function. You could skip it altogether, but if you get it, the answer is no, it does not need to be removed. It is very thin and hard. -- Dave Tweed -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

DO not remove the black hard anodized layer. It is an insulator so that the Aluminum is not electrically conductive. If you cut though the black anodized layer the heatsink will be electrically conductive. Also black hard anodized layer does little to effect thermal performance of the heatsink so there is no thermal purpose in removing it. The only reason to remove the black anodized surface is if you WANT to electrically connect the heatsink. {Quote hidden}> Should the black anodizing be sanded off where the semiconductor makes > contact? > > The black anodizing is purely cosmetic and serves no useful function. > You could skip it altogether, but if you get it, the answer is no, it > does not need to be removed. It is very thin and hard. > > -- Dave Tweed > > -- > http://www.piclist.com hint: The PICList is archived three different > ways. See http://www.piclist.com/#archives for details. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> The black anodizing is purely cosmetic and serves no useful function. Black anodizing can have a useful purpose in some cases. It increases the radiated heat, although does nothing for convected heat. This can be a benefit in some "still air" applications. Black anodizing is rather pointless in actively forced convection (fan on heatsink) cases. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> The black anodizing is purely cosmetic and serves no useful function. > You could skip it altogether, but if you get it, the answer is no, it > does not need to be removed. It is very thin and hard. Anodizing is NOT purely cosmetic. IIRC, anodizing helps to decrease the thermal resistance of the sink to the surrounding environment. I'm not sure that black is your optimum colour, though. But as Olin has mentioned, anodizing is relatively pointless in forced air environments. As proof of this, look at the heatsinks used by the newer PCs. They are all unfinished aluminum. I think the point really is that if the thermal resistance of the anodization can be made less than the thermal resistance of aluminum oxide (which quickly coats any bare aluminum) then the anodization makes for a better heatsink. Check out www.onsemi.com/pub/Collateral/AN1040-D.PDF for notes on mounting power semiconductors. The section under "Surface Treatment" has notes on anodization. Regards, Brendan -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

>Should the black anodizing be sanded off where >the semiconductor makes contact? Not normally, but you may wish to mill, not sand, the area where the device mounts. This is to ensure the surface of the heatsink is absolutely flat, and not got any bow or extrusion ridges on it. Try an unmounted chip on the heatsink to see if it will sit flat. In view of the amount of dissipation you are dealing with, you may wish to mill it anyway, as the anodising is another slight thermal resistance. Either way you will need a smear of thermal compound. What we did on the unit I mentioned was to have a 3 inch dia mill run across the extrusion to create a flat area about a 1/16 inch below the extruded surface. By mounting the PCB flat on the extruded surface the top surface of the PCB was then at about the correct height for the stripline transistors to solder directly onto it. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

>The black anodizing is purely cosmetic >and serves no useful function. Wrong. A black anodised heatsink will have better thermal radiation than a natural finish untreated one. You will find all tables giving figures for thermal resistance of heatsinks specify black anodised for this reason. >You could skip it altogether, but if you get it, >the answer is no, it does not need to be removed. >It is very thin and hard. Agreed, until dealing with very high dissipation. Then it is worth machining the surface after anodising, just where the devices are going to be mounted to get the best thermal contact. The anodising will create a slightly rougher surface, which will create a thermal resistance between the package and heatsink, which is not normally a problem. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Is heatsink flatness really that critical? I thought the heatsink compound is denser than aluminium anyway so provided there is heaps of goop on the part it won't matter if the 'sink is flat or not? I've used some real rough bits of alloy for 'sinks in the past, with lots of goop. :o) -Roman Alan B. Pearce wrote: {Quote hidden}> > >Should the black anodizing be sanded off where > >the semiconductor makes contact? > > Not normally, but you may wish to mill, not sand, the area where the device > mounts. This is to ensure the surface of the heatsink is absolutely flat, > and not got any bow or extrusion ridges on it. Try an unmounted chip on the > heatsink to see if it will sit flat. In view of the amount of dissipation > you are dealing with, you may wish to mill it anyway, as the anodising is > another slight thermal resistance. Either way you will need a smear of > thermal compound. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

The Flatness issue is really the removal of air from between the two mating surfaces. Air has a low Cp value, does not move heat energy well. In a perfect world the surfaces would be flat down to an atomic level on both parts. So when the part are brought together there is no air (insulator) between the parts. But we all know that things are never perfect, except in Classes on Thermal to make it easier for us ME's. So the grease is used to displace the air and improve thermal conductivity, not to fill in for large flatness issues. The grease layer needs to be as thin as possible, to reduce thermal resistance. A great thermal interface, if you can stand silicon is Gelvet by Honeywell. Here is the link, it is starting to be used on P4 CPUs to heatsink as an interface. http://www.electronicmaterials.com/products_services/packaging/gelvet_cp_comp_pad.pdf Greg Roman Black wrote: {Quote hidden}> Is heatsink flatness really that critical? > I thought the heatsink compound is denser than > aluminium anyway so provided there is heaps of > goop on the part it won't matter if the 'sink > is flat or not? > > I've used some real rough bits of alloy for 'sinks > in the past, with lots of goop. :o) > -Roman > > Alan B. Pearce wrote: > > > > >Should the black anodizing be sanded off where > > >the semiconductor makes contact? > > > > Not normally, but you may wish to mill, not sand, the area where the device > > mounts. This is to ensure the surface of the heatsink is absolutely flat, > > and not got any bow or extrusion ridges on it. Try an unmounted chip on the > > heatsink to see if it will sit flat. In view of the amount of dissipation > > you are dealing with, you may wish to mill it anyway, as the anodising is > > another slight thermal resistance. Either way you will need a smear of > > thermal compound. > > -- > http://www.piclist.com hint: The PICList is archived three different > ways. See http://www.piclist.com/#archives for details. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

>Is heatsink flatness really that critical? >I thought the heatsink compound is denser than >aluminium anyway so provided there is heaps of >goop on the part it won't matter if the 'sink >is flat or not? Well it really depends on how much dissipation you are trying to get rid of. If the contact area is large for the amount of dissipation, then it is not a problem. Think in terms of thermal resistance, and how many thermal resistors you are putting in parallel. The one application I described where we did do it was for VHF land mobile base station equipment which needed to be sited on hilltops and the like, not necessarily easy to get at, so using fans and other such devices requiring maintenance was in the no-go area of development. Hence we were looking at every avenue to get the chip temperature down, and this was part of that exercise. The other factor was that the extrusions do come with a non-flat surface and skimming it with a large diameter milling tool brought a number of other benefits for us. This is not normally noticed on most heatsinks you buy because there is not generally a large flat surface. But get one of those with all the fins on one side, and put a straight edge across it, and do not be surprised if it is not dead flat. If it is, I suspect it may have been milled before anodising - check for the extrusion mould marks down it. >I've used some real rough bits of alloy for 'sinks >in the past, with lots of goop. :o) This is often quite enough for the problem, but again it depends on what the dissipation level is, the original poster is talking of 120W, and I would not want a rough bit of metal and lots of thermal compound for that. From his description it sounds like things have got to the point where there are thermal problems, as in real problems, not sorting things out before a possible problem. It really is a case of horses for courses, and rough bit of metal on a 7805 would probably hold it OK for single digit watt dissipation in an Arizona desert summer, and still keep the chip inside out of thermal shutdown. I'm definitely not advocating every heatsink needs milled flat surfaces :) -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Olin Lathrop <EraseMEolin_piclistspam_OUTTakeThisOuTEMBEDINC.COM> wrote: > > The black anodizing is purely cosmetic and serves no useful function. > > Black anodizing can have a useful purpose in some cases. It increases the > radiated heat, although does nothing for convected heat. This can be a > benefit in some "still air" applications. Black anodizing is rather > pointless in actively forced convection (fan on heatsink) cases. Do you know how hot the heatsink would have to be in order for its surface color to make a real difference? Besides, on finned heatsinks, most of the surfaces face each other, so there is almost no heat transfer by radiation. Also, Greg Hayward <ghaywardspam_OUTCASPIANNETWORKS.COM> wrote: > DO not remove the black hard anodized layer. > It is an insulator so that the Aluminum is not electrically conductive. > If you cut though the black anodized layer the heatsink will be electrically > conductive. > Also black hard anodized layer does little to effect thermal performance > of the heatsink so there is no thermal purpose in removing it. > The only reason to remove the black anodized surface is if you WANT to > electrically connect the heatsink. It wouldn't be a good idea to rely on anodizing alone to provide insulation between a device and a heatsink. Although aluminum oxide is very hard, it is very thin and is still easily pierced by surface irregularities on steel mounting hardware. Always use additional insulators. If you *want* an electrical connection to the heatsink, you must use some sort of toothed lockwasher to bite into the metal and provide a gastight connection. Aluminum oxidizes very readily, and all exposed surfaces are quickly covered with a thin insulating layer. -- Dave Tweed -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Ooh, just had a thought about this. If you really want to keep it cool, and power's not an issue, you could use a pair of peltier devices. IIRC from when I was looking into those, the nicer ones can dissipate up to 75W, but are optimal at 30W, so you'd need about 3-4 of them. That's a lot of current, I know, but it might well be worth it to you. You should theoretically be able to get the case sub-zero without much difficulty. The other advantage is that the other side of the peltier devices would likely get hot enough to make dissipating the heat a bit easier (higher thermal potential, with same thermal resistance means higher thermal flow). Just a thought, though the high current might finish it for you. BGmicro is selling them in series pairs right now, or at least, that's what the catalog they shipped to me said. My $.02 --Brendan -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> IIRC, anodizing helps to decrease the thermal resistance of the sink to the > surrounding environment. I'm not sure that black is your optimum colour, > though. It is. You want the surface to be able to maximally radiate at all wavelengths, therefore black. Ideally the surface becomes a "black body" radiator. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Alan B. Pearce wrote: > > >Is heatsink flatness really that critical? > >I thought the heatsink compound is denser than > >aluminium anyway so provided there is heaps of > >goop on the part it won't matter if the 'sink > >is flat or not? > > Well it really depends on how much dissipation you are trying to get rid of. > If the contact area is large for the amount of dissipation, then it is not a > problem. Think in terms of thermal resistance, and how many thermal > resistors you are putting in parallel. Thanks for the response, but you didn't cover the basic issue. :o) We are using zinc oxide, which I have always believed is DENSER than aluminium, ie will sink better than machined alloy no matter how flat. So if there is a gap between the device and the uneven heatsink, the material that completely fills the void is better than the alloy anyway. So in terms of your "thermal resistor" analogy, isn't it more like the alloy is the high resistance and the goop is like a short circuit? I've used many chunks of alloy as heatsinks over the years, even cast alloy casing of machinery etc. I use lots of goop, even as a "bead" on the sides of the casings which improves thermal transfer over what is possible with just the mating surface. Remember I like large linear PSUs etc. :o) -Roman -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> Is heatsink flatness really that critical? > I thought the heatsink compound is denser than > aluminium anyway so provided there is heaps of > goop on the part it won't matter if the 'sink > is flat or not? > > I've used some real rough bits of alloy for 'sinks > in the past, with lots of goop. :o) The "goop" is actually a pretty poor conductor of heat, but still better than air. You want to use as thin a coating of goop as possible. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> It is. You want the surface to be able to maximally radiate at all > wavelengths, therefore black. Ideally the surface becomes a "black body" > radiator. I am probably remembering something wrong, but my thought would be that if you don't want the sink to pickup stray radiation from the surrounding environment and have to radiate it too, then black is the wrong colour to use. I thought that blackbody radiation was simply for a situation where reflected energy was ignored. With a sink, don't you want it to reflect energy and thereby have less to deal with? I'm probably off on a tangent from the actual course of things. --Brendan -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> > Black anodizing can have a useful purpose in some cases. It increases the > > radiated heat, although does nothing for convected heat. This can be a > > benefit in some "still air" applications. Black anodizing is rather > > pointless in actively forced convection (fan on heatsink) cases. > > Do you know how hot the heatsink would have to be in order for its surface > color to make a real difference? Besides, on finned heatsinks, most of the > surfaces face each other, so there is almost no heat transfer by radiation. I remember seeing some tables where black anodized versus bare aluminum made a noticeable difference. Of course for radiation to be have any cooling effect, the heat sink has to be "looking" at something that isn't also hot. You're right, fins don't do anything useful to increase radiation. They are for convection only. You can prove to yourself that radiation can cause cooling in a practical way. Take an old soup bowl, paint the inside black, and fill with water to about 5-10mm of the top. Insulate the outside of the bowl well and cover the top with a layer of plastic wrap to make a layer of still air above the water. Set this outside overnight with a clear view of the sky. The water can freeze under the right conditions without the air temperature ever dropping to the freezing point. This works only when the sky is clear because the water is radiating its heat to the sky but very little is coming back. Try it yourself. You will have to wait a while for the right weather, especially considering the heat wave we've been having around here the last few days. Some of the folks in the southern hemisphere or at high altitudes could do it this time of year. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

The exchange of thermal radiation between surfaces is a function of 1) Surface emission properties, 2) Surface absorption, reflection, and transmission properties, and 3) Properties of the medium that lies between the thermal radiator and receiver. The radiating surface can only transmit energy through radiation to surfaces that it can *see*. So on a finned heatsink most of the radiated heat is just back on an adjacent fin, not a cooler surface away from the sink. If there is a concern of radiated energy being absorb by the heatsink then a polish aluminum or copper suffice should be used as it will reflection more of , not absorbing, the heat energy. Brendan Moran wrote: {Quote hidden}> > It is. You want the surface to be able to maximally radiate at all > > wavelengths, therefore black. Ideally the surface becomes a "black body" > > radiator. > > I am probably remembering something wrong, but my thought would be that if > you don't want the sink to pickup stray radiation from the surrounding > environment and have to radiate it too, then black is the wrong colour to > use. I thought that blackbody radiation was simply for a situation where > reflected energy was ignored. With a sink, don't you want it to reflect > energy and thereby have less to deal with? > > I'm probably off on a tangent from the actual course of things. > > --Brendan > > -- > http://www.piclist.com hint: The PICList is archived three different > ways. See http://www.piclist.com/#archives for details. -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> > It is. You want the surface to be able to maximally radiate at all > > wavelengths, therefore black. Ideally the surface becomes a "black body" > > radiator. > > I am probably remembering something wrong, but my thought would be that if > you don't want the sink to pickup stray radiation from the surrounding > environment and have to radiate it too, then black is the wrong colour to > use. This is basic physics. You can't have something that receives radiation well but is poor at emitting it or vice versa. If you did, you could easily construct a "free energy" machine. Black objects get hotter in the sun than shiny silvery ones because they absorb more of the incoming radiation. Therefore they also radiate more at the same temperature. A perfect mirrored object would never radiate at any temperature. > I thought that blackbody radiation was simply for a situation where > reflected energy was ignored. A "black body" produces the maximum possible radiation due to its temperature. A perfect black body must therefore also absorb all incoming radiation, which means there is no reflected radiation. Reflection has nothing to do with black body radiation. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Roman Black <@spam@fastvidKILLspamEZY.NET.AU> wrote: > Thanks for the response, but you didn't cover the > basic issue. :o) We are using zinc oxide, which I have > always believed is DENSER than aluminium, ie will sink > better than machined alloy no matter how flat. Actually, density has very little to do with the two parameters that are critical to thermal managment: thermal conductivity and specific heat. Lead is the densest metal, but it falls short of other metals in both areas. Aluminum scores fairly high in both areas, plus it has the advantage of being relatively easy to shape. Zinc oxide powder suspended in silicone grease is not a better conductor than the aluminum, so you really do want a very thin layer. A bead of it on the edges of a device is probably counterproductive (wasteful, and messy besides). -- Dave Tweed -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Olin Lathrop <KILLspamolin_piclistKILLspamEMBEDINC.COM> wrote: > I remember seeing some tables where black anodized versus bare aluminum made > a noticeable difference. Of course for radiation to be have any cooling > effect, the heat sink has to be "looking" at something that isn't also hot. > You're right, fins don't do anything useful to increase radiation. They are > for convection only. > > You can prove to yourself that radiation can cause cooling in a practical > way. ... All very true, but knowing the environments in which *most* heatsinks wind up, I think it's safe to say that radiative cooling is a non-issue. On spacecraft, heatsinks work *only* by radiation. The fins are black and usually angled at least 60 degrees relative to each other. > You will have to wait a while for the right weather, especially > considering the heat wave we've been having around here the last few days. Man, it's brutal out there. We New Englanders aren't built for this kind of heat. -- Dave Tweed -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

On Tue, 2 Jul 2002, Sean H. Breheny wrote: >Hi Peter et al, > >Thanks for the great suggestions, I'll check them out. > >Yes, I am using a fan. Yes, I already tried making one myself out of some >scrap aluminum here, but the Al was too thin a gauge and the heat wasn't >conducted to the fins as much as it should be (I also only made fins by >bending the edges up so there weren't enough fins). > >I need to dissipate about 120W. The device is a microwave power amplifier >which is 5inches by 8 inches. The thermal conductivity of the case (appears >to be some brass-type alloy, at least it is that color) doesn't seem to be >that great because you can definitely feel hot spots on it while it is >running. That's why I want the heat sink to cover the whole side of the >case, even if that isn't really necessary to dissipate the heat. Try to think of immersion cooling. You can seal the whole amplifier in a box (old paint can soldered shut) and partially fill it with oil. If you use a gallon can it will only get moderately hot on your 120W, without a fan. 120W sounds like you want a high performanec CPU cooler to me, and that should be enough. To fix the hot spots obtain thin copper flashing, smear it with thermal compound on both sides and tighten the screws as much as you dare (get the module's data sheet). The flashing serves as ductile washer and will embrace any asperities in the surfaces. The base of the amplifier is annealed tempered brass. I also have one of those modules, for 800MHz, about 10W. It's important that you do not cause the base to bend. If it does, it will crack the chips glued directly onto it (they are inside the epoxy part). Peter -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

> You can prove to yourself that radiation can cause cooling in a practical > way. Take an old soup bowl, paint the inside black, and fill with water to > about 5-10mm of the top. Insulate the outside of the bowl well and cover > the top with a layer of plastic wrap to make a layer of still air above the > water. Set this outside overnight with a clear view of the sky. The water > can freeze under the right conditions without the air temperature ever > dropping to the freezing point. This works only when the sky is clear > because the water is radiating its heat to the sky but very little is coming > back. Try it yourself. You will have to wait a while for the right > weather, especially considering the heat wave we've been having around here > the last few days. Some of the folks in the southern hemisphere or at high > altitudes could do it this time of year. AFAIK the "ancient" Romans & Egyptians used to make ice by this method. RM -- http://www.piclist.com hint: The PICList is archived three different ways. See http://www.piclist.com/#archives for details.

Olin Lathrop wrote: > This is basic physics. You can't have something that > receives radiation > well but is poor at emitting it or vice versa. If you did, > you could easily > construct a "free energy" machine. Too strong statements, as for me. Mother Nature constracted >something that receives radiation well but is poor at emitting it< : trees, grass. Why "free energy" machines - rather "radiation energy" machines. In case of trees and grass - "sun radiation energy" machines. Human made things are ones, based on energy saving films, transparent to main sun radiation, while reflecting infrared. This films are often built in windows. This helps room to be more warm at winter times. Impressive experiment: get two windows, one with the film and other without. Place behind them powerful infrared source: 1kwt reflector lamp at low voltage with a red filter. Standing before ordinary window you feel lamp heat, before window with the film you feel no heat. The film at that is invisible. Mike. -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email RemoveMElistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

>On spacecraft, heatsinks work *only* by radiation. The >fins are black and usually angled at least 60 degrees >relative to each other. Umm 'taint necessarily so :) My previous project was involvement with a power supply for an earth resources satellite. The major heat radiating point is open to space (as in not covered by those gold coloured blankets) to get the radiation for getting rid of the heat, and is covered with a highly reflective silver (actual metal not known) foil to minimise heat absorption from the sun. Even inside the assembly where I had LM117 type regulators mounted the aluminium was alodised, resulting in that gold/brass colour mentioned previously, rather than black anodised. The rest of the outer surface of the assembly was black anodised though. However this is not exposed in a manner allowing it to radiate heat to space. -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email spamBeGonelistservspamBeGonemitvma.mit.edu with SET PICList DIGEST in the body

Dave Tweed wrote: {Quote hidden}> > Roman Black <TakeThisOuTfastvidEraseMEspam_OUTEZY.NET.AU> wrote: > > Thanks for the response, but you didn't cover the > > basic issue. :o) We are using zinc oxide, which I have > > always believed is DENSER than aluminium, ie will sink > > better than machined alloy no matter how flat. > > Actually, density has very little to do with the two parameters that > are critical to thermal managment: thermal conductivity and specific > heat. > > Lead is the densest metal, but it falls short of other metals in both > areas. > > Aluminum scores fairly high in both areas, plus it has the advantage > of being relatively easy to shape. > > Zinc oxide powder suspended in silicone grease is not a better > conductor than the aluminum, so you really do want a very thin > layer. A bead of it on the edges of a device is probably > counterproductive (wasteful, and messy besides). Thanks Dave, yes I stand corrected. I looked up thermal conductivity on google and got a list of metals and insulators values. Heatsink compound was a lot worse than aluminium, quite surprising the amount that it was worse. Also surprising was just how badly water scored on that list too, but I don't think it takes into account any convection. -Roman -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email RemoveMElistservTakeThisOuTmitvma.mit.edu with SET PICList DIGEST in the body

> > This is basic physics. You can't have something that > > receives radiation > > well but is poor at emitting it or vice versa. If you did, > > you could easily > > construct a "free energy" machine. > > Too strong statements, as for me. Mother Nature > constracted >something that receives radiation > well but is poor at emitting it< : trees, grass. Actually trees and grass receive and emit equally well. However, since their temperature is significantly lower than that of the sun's, they mostly receive compared to the amount they transmit. None of this violates the basic physics I was talking about. They are also moist with has a whole other set of effects on their temperature, none of which are relevant here. > Why "free energy" machines - rather "radiation energy" > machines. In case of trees and grass - "sun radiation > energy" machines. Because the existance of these machines doesn't disprove the existance of things that are good receivers but poor emitters (or the other way around). They are converting incoming radiation into other forms of energy. I thought it was obvious how you could make a "free energy" machine with objects that differ in their reception/transmission. I further thought everyone would agree that such a free energy machine can't exist, thereby proving that objects with different reception/transmission characteristics can't exist. Apparently I was wrong. Suppose object A radiated better than it received, and object B received better than it radiated. Now suppose I surround these objects with perfect reflectors so that they are both "seeing" each other. Both objects start out at the same temperature. After a while, A will radiate more onto B than B will radiate onto A, thereby lowering the temperature of A and raising the temperature of B. I could now use this temperature difference to extract work, getting energy out of the system without having put any in. Furthermore, this process could go on indefinitely. That means power is produced by the system without anything being used up or any power added. Let's hope at least that you think that is a violation of the basic laws of physics, else I'm wasting my time and invite you to go patent a perpetual motion machine. > Human made things are ones, based on energy saving > films, transparent to main sun radiation, while reflecting > infrared. This films are often built in windows. This helps > room to be more warm at winter times. These things exist, but don't disprove what I was saying because the sun and the room aren't at the same temperature. All the coating does is alter the absoption/radiation effeciency spectrum, it does not allow more absorption than radiation at a specific wavelength. ***************************************************************** Embed Inc, embedded system specialists in Littleton Massachusetts (978) 742-9014, http://www.embedinc.com -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email listservEraseME.....mitvma.mit.edu with SET PICList DIGEST in the body

On Thu, 4 Jul 2002, Roman Black wrote: >Is heatsink flatness really that critical? >I thought the heatsink compound is denser than >aluminium anyway so provided there is heaps of >goop on the part it won't matter if the 'sink >is flat or not? The flatness has to do woth breaking the chips on the carrier being mounted, and not with heat. Heatsink compound is denser than aluminium in the weight sense (maybe) but it is more ductile. So it fills the voids which is what it is supposed to do. >I've used some real rough bits of alloy for 'sinks >in the past, with lots of goop. :o) The goop is not snake oil. If you want real good conductivity you want the sealant to be thin. This means flat surfaces. The copper flashing shim I proposed usually solves the problem of not so perfect surface finish assuming it can be tightened well. Peter -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email EraseMElistservmitvma.mit.edu with SET PICList DIGEST in the body

On Wed, 3 Jul 2002, Brendan Moran wrote: >> It is. You want the surface to be able to maximally radiate at all >> wavelengths, therefore black. Ideally the surface becomes a "black body" >> radiator. > >I am probably remembering something wrong, but my thought would be that if >you don't want the sink to pickup stray radiation from the surrounding >environment and have to radiate it too, then black is the wrong colour to >use. I thought that blackbody radiation was simply for a situation where >reflected energy was ignored. With a sink, don't you want it to reflect >energy and thereby have less to deal with? The sink will pick up stray radiation from the surrounding but it will radiate itself SIGNIFICANTLY more than that because it is hotter. The gain obtained like this far outweighs losses due to captured radiation. The black effectively improves the radiation coupling of the object to the medium, in both directions. Peter -- http://www.piclist.com#nomail Going offline? Don't AutoReply us! email RemoveMElistservEraseMEEraseMEmitvma.mit.edu with SET PICList DIGEST in the body