We’re going to start things off with another question. We got a couple of questions embedded in here and it’s just sort of to get an idea. For example, this one, “Do you currently have a good understanding of when and how to utilize vapor chambers, as opposed to heat pipes?” So, we just get a feeling of feedback from the audience as to what their expertise level is, so if you can go ahead and vote on that and we’ll take a look at the results later on today.
In today’s webinar, we’re going to basically have six areas that we’re going to talk about. The first one is, “Do I need to use two-phase?” and we’ll look at the benefits and consequences of using solid materials as opposed to two phases. We’ll look at some basic rules of thumbs, and then we’re going to talk about vapor chambers; and always, the question is, are they just flat heat pipes and we’ll answer that question, of course.
We’ll take a look at some sizing; we’re going to talk about how do I integrate them, we’re going to talk about the heat exchange or in other words, the getting rid of the heat part, and how you design those and from a construction standpoint, what they look like. And then finally, what about thermal modeling for these devices? We’re going to talk about that, so we’re going to cover all that stuff today.
Moving on, when to use a two-phase device. And the short answer is always when your design is conduction-limited, because obviously you’re not going to move away from a piece of metal to something else, unless there’s a reason to, and in this case it’s always the conduction limit that’s reached. There’s other reasons you may want to use it, non-thermal goals as we say, such as weight or size. Sometimes the vapor chamber or heat pipe devices can be smaller or lighter than its equivalent performance in a solid material. And we have three examples here.
We took a look at, in the same application, and the same format heat sink, same airflow, same power, everything is held the same; an aluminum heat sink, a copper heat sink, and then a heat sink with a vapor chamber embedded into the base. And just to give you a quick idea on the comparison of these. If we look at the aluminum, the conduction loss and the delta T due to conduction in the base is about 22 degrees. If you go to copper and you keep the base the same thickness, that’s just going to cut it in half to 11 degrees. Often, what happens is they will reduce the thickness of the base to get a little more fin area. And so, in this case it was 17-degree delta T. But if you keep everything the same, the copper has twice the conductivity. So somewhere between 11 and 17 degrees for the copper version. But if you were to put a two-phase device in it, the conduction loss is only four degrees, so you can see why you would go to using a two-phase device.
So, let’s look at some rules of thumb. Everybody knows that the two-phase devices are incredible heat conductors. People talk about huge multiples of better conductivity than aluminum or copper, five to 50 times. I think for us, for this audience, the W/mK numbers are more important. If I’m doing a sort of a local spreading, you might be down in the 1000 W/mK range. And if I’m moving heat across the room or some distance away, you can get those higher numbers, the 50,000 numbers. We say that heat needs to be moved more than, I can surely say, an inch or an inch and a half to two inches, but 30 to 50 millimeters before you look at using to phase over a piece of metal.
There’s a couple of other rules of thumbs; if you’re interested in spreading heat to reduce a hotspot or attach to a local heat exchanger, the ratio of heat spreader to heat source should be on the order of 20 to one, in other words the area of the heat sink versus the area of the device being cooled. Anything less than that, generally metal is good enough. And when you’re designing with a heat pipe or a vapor chamber, we always leave 25% thermal headroom. In other words, for 100-watt heat pipe, I will use that as a 75-watt in application, so I always leave that margin.
So, I think everybody has seen these slides, probably everybody understands how these things work. The key takeaway that you want to keep in mind here, and the question I generally get is these devices are vacuum devices. So, they do work below the working fluid boiling point, and that the limit to the performance of these things is usually the capillary limit, and that is getting the fluid to be returned back to where the heat is taking place. Let’s talk about the wick structure, since we’re talking about the capillary limit. There’s three common wick structures that are used, and if I had to say percentage-wise in the market, probably high-90% of the heat pipes in the marketplace use the sintered powder.
The other two common ones are screen and grooved wick structures. You can see the power density differences between the three designs, the thermal resistance, and the orientation. So, the sintered wick is good for high-power densities, it certainly has fairly low thermal resistance, and it’s good in working in any orientation. If you’re looking at screen wick, it won’t handle as much power density, in fact, relatively low, less than 30 watts per square centimeter. The nice thing about a screen wick is it’s used when you need a higher Q max; it will move more liquid in a horizontal application. And grooved is just, really, the lowest-cost structure you can make.
And this slide talks you through some of the performance limits. As you can see, the primary limit is always the capillary limit and that goes along with the previous slide when you’re talking about wick structures. This shows a calculation for sintered wick heat pipe. So you can see that the capillary limit in the temperature range of electronics is the limiting factor. It talks about the other limits, but we don’t need to go into those in great detail.
What’s important to see here is when we do look at the performance of these devices, in these two charts what we’re showing is the chart on the left are heat pipe performance limits, and the chart on the right the vapor chamber; and it’s by diameter for the heat pipes and width for the vapor chambers. The blue line on the heat pipe will show you the vapor limit, whereas the red line will show you the wick limit. So you can see that the wick limit is the limiting factor, until you start to flatten the heat pipes.
And once you start to flatten the heat pipes, then the vapor limit generally comes down to be in line with the wick limit. In fact, when you’re making ultra-thin devices, that’s one of the things that you need to optimize and that is the sweet spot between the wick limit and the vapor limit. And if you look at the vapor chamber chart on the right, you can see that the wick limit is above the vapor limit and that’s just because these are flattened devices, and the vapor space itself is smaller when you’re looking at rectangular devices versus around device.
Here’s some information on the wick structure. So, we’re seeing two photographs of sintered wick structure; the one on the right in the tube is what we would call just a standard wick structure. And the porous wick structure, the post one is the more open wick structure. Depending on the application, you can vary the porosity and the particle size of the wick structure to give you a different performance. So, you really can’t just say, “Okay. Take a look at the published data from heat pipe manufacturers,” because the wick structure can be tuned to meet a certain application. It’s just like designing a car for a particular application, if you want to go fast or if you want to go off-road, you do different things. You can change the wick thickness, you can change the width porosity, you can change the working fluid. So, a lot of things you could tweak to match the performance of these devices to the requirement of the application, so that’s the key takeaway.
So, how does orientation affects the performance of these? It’s a very common question. A lot depends on the wick structure and whether it was designed to work against gravity. So, we have a chart here that shows you the performance of these different diameter heat pipes versus the gravity angle; and we go from a positive 90-degree angle where the heat source is at the bottom, and the cooling is at the top, to the minus 90 where that’s turned around, where the heat sources at the top. And you can see that once you get to -30, the performance really drops off. A lot of this is a function of the length of the device. And what we really wanted to show you is if I take my eight-millimeter heat pipe, so we have two lines for eight-millimeter, a blue line and a red line. So, we’ve done some gravity optimization on the red line for the eight-millimeter, and you can see how that works much better against gravity.
Moving on to effective thermal conductivity, we call it effective thermal conductivity because the conductivity of these things will change depending on the application. So, for any particular application, we can calculate what an effective conductivity would be, and what we wanted to show you here was how that number changes. So, we’ve put together a little thermal assembly here with a couple of heat pipes and all we’ve done is change the length of the heat pipes, so the power per heat pipe is the same. So, we go from a heat pipe at 75 millimeters long up to 200 millimeters long. Each of those carrying 25 watts and the conductivity numbers go from 6600 W/mK up to 28,000 W/mK. So, you can see if I extended this thing another 200 millimeters, how you would start to get the rather large numbers, but for most electronics cooling applications you don’t move the heat much more than about 150 to 200 millimeters. And for local spreading uses the 6,000 number is pretty typical.
Okay. What is the difference between the most common versions of these devices, heat pipes, vapor chambers? Heat Pipes are heat pipes, they’re round heat pipes, they’re generally available in 3-to-10-millimeter diameters. Although, for power electronics applications and other applications, we’ve made them up to meters in diameter. Power electronics three quarters or 20-to-25-millimeter diameters are not unusual. Then there’s two other types of vapor chambers and what we call a hybrid one-piece vapor chamber that is really made out of a large diameter tube and a traditional two-piece vapor chamber. Where the two pieces are stamped, the wick structure’s put in both pieces, and then it’s sealed, gently welded the whole way around the periphery. This just gives you some typical dimensions, how to mount them, the relative cost of the devices.
So, we talk a lot about moving and/or spreading heat and what is the difference between those two and what effect doesn’t have on your design? Generally, moving is when you have a heat sink that’s located remote from your hot device. We call that moving heat, so you’re moving it from point A to point B. Spreading heat is where your heat sink is located at the area where you have hot device is, but the heat sink is much bigger than your device so we call that spreading. With the heat pipes what you do is you get a linear heat flow moving heat from point A to point B and A. In a spreading application, you’re spreading heat in all directions.
Occasionally, you want to do both. For example, say you need five heat pipes for a particular application, and those five heat pipes won’t fit underneath your heat source, then you’re going to have some conduction issues and getting the heat spread to those heat pipes, you got attachment issues between the spreader and the heat pipes. In those cases, a vapor chamber can be used to do both spreading and moving. So, it’s not exclusive, heat pipes are used only for moving and vapor chambers are used only for spreading, you can mix and match those things depending on the application.
And then we have question number two for the day, “Does your company currently use heat pipes or vapor chambers?” Obviously, we’re looking for answers as, “Yes. We use them, and, “We’re thinking about using them,” and “No. We have no plans on using them.” And we will move on to the next one. 17. This is a new platform, so bear with me as we learn how to use it.
When moving heat remote heat sink, this is probably 99% of all applications that are out there because from a volume standpoint, most of those applications are in computing applications. Complex shapes are often required, as you can see in the photograph. Heat Pipes can be bent relatively just as easily as a piece of copper tubing. They’re available in volume, the manufacturing capacity in the world is many, many millions per month. The work against gravity is not too bad. For example, for a laptop application, that it will work against gravity up to a 45-degree angle, satisfactory for most of those applications. And in this particular application, it’s a notebook computer, you can see where those two flattened heat pipes used to call three heat sources. And what they’re doing is they’re just daisy-chaining the heat pipes to move the heat from multiple heat sources. That’s fairly common, is to use multiple heat sources on a particular device.
So when we’re talking about spreading heat, and in this particular case, we’re spreading heat in what we call a local heat sink; in other words, it’s not a separate heat thing that’s located somewhere else. You can use the pipes, heat are actually a good choice if you’ve got reasonable airflow, you’ve got plenty of room for fans, you’ve got nominal power densities. In other words, if it’s not a very challenging application, heat pipes make very good spreaders. And here’s an example of one for a telecom application where four heat pipes were put into the base of the heat sink just to move the heat off to the side where the heat sink has been extended.
Now, we will occasionally use vapor chambers for spreading on a local heat sink, also, just like heat pipes. Vapor chambers really are the best choice if you really limited in your space, generally that’s Z, your height is limited, or your power densities are high, or your airflow is low, or it’s just a challenging application where you need every degree you can get. In this example, what we’ve done is we’ve taken a heat pipe design from a previous generation of graphics chip coolers. Obviously, the power densities and the powers have gone up for the next generation, and they wanted to use the same basic shape and form factor. So, we replaced the two eight-millimeter heat pipes with one vapor chamber. And what that did is it gave us a six-degree better cooling versus the original heat pipe design. And the reason for that was just better spreading, and in this case more fin area because the bent heat pipes took up some space for the fins.
And here’s another example of vapor chambers used to reroute heat sink. Challenging application, laser diodes for a very, very high-end 3D projector. Three laser diodes for the three colors, and a common heat sink for all three, so vapor chambers are used for spreading heat. This is just some odd-ball examples we wanted to throw out there for you. Example one is flattened heat pipes that are flattened and soldered into a base and the heat pipes are machined. Often, this is used to, as best possible, get what they call direct contact. The heat pipes make direct contact with a heat source eliminating the two interface layers. Works pretty well. The machining of those heat pipes, you really need to know what you’re doing to not have an issue with machining of the heat pipes. But it’s rather common that it’s done.
Example number two shows an area where they wanted to upgrade the performance of that particular heat sink. It moved to the next generation processor, the i7 chip. There was a copper base on those four heat pipes, and that was changed now to use a vapor chamber and they picked up five degrees better performance. Most of that had to do with the increased power density with the i7 chip.
So, let’s talk a little bit about bending and shaping. As we mentioned earlier, heat pipes are very flexible, no real difference than bending simple copper tubing. Some rules of thumbs around that are the bend radius is 3X the diameter the heat pipe. You could push that a little bit down to two and a half or two, 3X the diameter is a good place to start. Each 45-degree bend reduces the Qmax by about two and a half percent. So, keep in mind that as you’re doing this bending you are reducing the performance of the heat pipe itself. You’d have flattened the heat pipes to, generally, about one third of the original diameter. So, for a six-millimeter heat pipe, you can gently flatten those two millimeters.
You can machine the wall if you’re careful. For example, the picture to the left shows these six heat pipes in a base that are machined after they’ve been installed and flattened. If the heat pipe wall thickness is thick enough that you can take off maybe a tenth of a millimeter, then you can do your machining. Some of the heat pipes on the market today have point three-millimeter wall and you really don’t want to machine those because you’re just getting the wall too thin. So, the one-piece vapor chamber, because it’s made out of a big tube and flattened, you can bend them; you can bend them along the narrow plane, as you can see in the pictures.
The bend radius is typically 10 millimeters. These particular tubes are flattened to 1/10 or 1/20 of the original diameter, but they’re designed that way, so they’re designed that they could be flattened to that extent. Surface pedestals can be formed into the vapor chamber, either to extend above the vapor chamber to reach into a recess for the heat source such as in a processor that’s got a stiffening ring around it.
And the last one to look at is the two-piece vapor chamber where you’re using two stamped parts that are then put together. Gives you more flexibility on the bend, so the stamp bend can be one extra thickness of the sheet metal. You’ll pay a pressure drop penalty for that, but it’s what we call a step and you can see that in the picture on the left. So, the upper and lower plates are stamped flat, and then their wicks put in. A stamp pedestal’s up to three to five millimeters taller or possible, then these things are welded up and processed just like any heat.
Let’s go on to sizing these devices. Based on what we call a typical wick structure, we’ve taken a look at the common sizes of heat pipes, and giving you a guideline as to the typical Qmaxes, what these things can be flattened to, what’s the resulting width once you’ve flattened them, and what happens to the Qmax when you go through flattening. So, you can see, for example, on the three millimeter your Qmax is typically in the 15-watt range. Three millimeters are commonly flattened down to two millimeters, the resulting width is 3.57 millimeters wide. And if you flatten these, assume that you’re down to about 10 watts in Qmax. So, we’ve given you those numbers for the common sizes, and let me walk through an example here, what do I do with that information?
I’ve got a 71 basic, that’s 20 by 20 millimeters, and we’re going to move the heat with 190-degree bend. So, if I look at three six-millimeter heat pipes, so that gives me 18 millimeters in width for a 20-millimeter heat source, good coverage on the heat source. If I look at the Qmax per heat pipe, we’re at 38 watts, so times three, so I get 114 watts of capability. If I assume a 25% safety margin, that means I can do 85 watts, plus loss due to bending, two 45-degree bends, one 90-degree bend, so my performance with those three 6-millimeter heat pipes is at 81 watts. So, we’re okay, because we have 70-watt application. The other option would be to use two eight-millimeter flat heat pipes and we’d do the same calculation here is what’s the Qmax per heat pipe and then the total. And then if I run a 75% margin. So, I’m down to 74 watts, so either of those two designs would work in this case, thermally work. And you can see you get pretty good coverage on the heat source itself.
We’re going to jump into mounting options and we’re going to go through this quickly because the mounting options for heat pipes and vapor chambers are the same as any other heat sink. We just showed you some examples here, the upper right-hand corner is a notebook application where they’re simply soldering the two heat pipes to a stamped-out spring clip that probably uses pushpins or screws to hold it down onto the board and onto the device. Once you get up into higher-end applications, higher power servers and things like that, you move up to the typical screws and springs.
The heat exchanger design used for these devices is no different than what’s used in just typical heat sinks. You can use extruded heat sinks, die cast heat sinks, you can use bonded fins, skived fin, fin packs, forged fins, machine heat sinks. I think the most common that are used in the marketplace is probably fin packs because they’re used in notebooks and machine heat sinks that are used in really high-end applications. So, the most common ones are the fin packs and the machined. Although, you can use any of these with the pipes and vapor chambers.
Assembly attachments, I think that it’s safe to say that over 90% of the applications use solder. It does a couple of things; one, it gives you a good mechanical joint, as well as a good thermal joint, and the key there is the thermal joint. The aluminum parts are nickel-plated so that they’re directly soldered between the copper and the aluminum parts using nickel. Sometimes, we will use a thermal epoxy. Often, that’s used when the power densities are relatively low and the parts are relatively large, so either of those are acceptable. The soldering is going to give you better performance.
Let’s walk through the design process for these things. So what we normally do is we’ll estimate the size and number of two-phase devices to manage the heat spread or move the heat. We’ll generally oversize those by about 25% take up for any variation in operation, orientation, or power or any of that kind of stuff. Or estimate the size of the heat sinks that’s required to dissipate the heat. Then what we do is we move that into an Excel model so that we can do some what-ifs, “If I make the fins longer or shorter,” that sort of thing. Once we think we’re getting close, we generally move that design, then, into a CFD model, and then obviously build something and get some test data.
Step one is always to do what we call a volumetric calculation. So, you’re sitting in a meeting where they’re talking about the next product that you’re going to work on, how much power that you’re going to have to take care of and how little space they’re going to give you to do that. The volumetric calculation allows you to get pretty close to a size that’s required to do your cooling based, simply, on a couple of numbers that you’ll have early on in the design phase. Generally, you’ll know the power that you need to get rid of, you’ll know the delta T or the ambient that it’s going to be in, so the available delta T. So for example, if I’ve got 800 watts that’s got an available delta T of 40 degrees C rise above ambient, you often know how much airflow you’re going to have. Some moderate airflow, no air where flow; in this case, we’re using moderate airflow, two and a half meters per second.
So, you can do a quick estimate as to how much volume is going to be required for that solution. So, if you do the quick calculation it says, “Hey, I need 2,300 centimeters square.” Convert that inches cubed, I could need 140 cubic inches. So, you can get an answer in a couple of minutes as to how much space that you think you’re going to need. How close is this to reality? Well, the final design for this particular application was 120 cubic inches. So, the good thing is it overpredicted the space you needed a little, but it’s within 15% of what the real solution turned out to be. That’s close enough for that phase of the design.
The next thing we’ll often do is do the same thing in an Excel spreadsheet. In an Excel spreadsheet, what we do is we just take a little more detailed look at the individual pieces of the design. I see I’m just a couple of minutes, so we’ll move along here. So, we take a look at the heat input, we take a look at the TIM. The interface material; what’s its resistance? What’s its delta T? We do some conduction estimates. We do some estimates on the delta T on the heat pipe or vapor chamber and the key thing is to look at the heat exchanger design, the fin area in the air flow.
The nice thing about this is it breaks it down, it says, “Hey, look, here’s your bits and pieces.” Your TIM is seven degrees, your base conduction is 17, your heat exchanger design is 18-degree delta T. And you can see that when I vary that from just an aluminum heat sink down to a two-phase heat sink, my 17 degrees goes down to four degrees. This generally will get within about 10% of the final solution, but it allows you a whole lot of flexibility to do what ifs just by changing some quick numbers.
Obviously, the next step is to go through a CFD analysis. Don’t think we have to talk about the benefits and drawbacks of that. I think there’s a presentation from Mentor that’s going to go through some of that stuff. So, one of the potential drawbacks of that is you have to have enough validation of your model before you can rely on it. So always validate your model; either you have enough information from the previous design, or before you believe this you want to build and get some real data.
Then I think the last question of the day, “How much thermal design assistance does your company require or expect from its members?” Virtually none, we have our vendors build to print; some feedback from the vendors is nice; or we rely a lot on third-party or vendor thermal engineers for their expertise; or obviously don’t know. And so, that’s the last question of the day. And if we go to questions and answers.
Graham Kilshaw:
Thanks, George. Great presentation, as usual. We do have quite a few questions, and attendees, if there’s questions you’re thinking of, now’s the time to enter them. We’re up to 12 questions so far. Let’s see if we can get through them all. Attendees, if you want to ask a question, just use the Q&A button at the bottom of your screen and send them in right now. All right. George, let’s see what we can do. First question comes from Jose at Intel, “Do you have a catalog of HP and VC and their capacity based on the size, length and flatness?”
George Meyer:
No. Sorry. Because the performance of these things changed as soon as you change any of the attributes of the application. If I change the heat source size, if I change the heat sink size, if I change the airflow; if I change any of those things, the performance of the heat pipe or the vapor chamber’s going to change. What we generally do is say, “We’ll give you some guidelines. When you have an application, call me and I’ll give you the number.”
Graham Kilshaw:
Okay. I think we’ll do some follow-up on these questions outside of the presentation as well, so we can get those contact numbers back to everybody. Next question, George, comes from Russell at Bluecore. Don’t know if you can answer this question or not, “Out of interest, how are the wicks manufactured inside the copper tubes?” Can we answer that one or not?
George Meyer:
Yeah. We’ve got a bunch of little elves that do that. No. I’m kidding. Depending on the wick, in other words, a grooved wick is going to be put in during the manufacturing of the tube. Either they draw the groove into the tube or they will swage the tube over a mandrel and put the grooves in. The screen, if you’re using a screen wick, is just cut, rolled up and slid in, and often, then expanded to touch the wall well. The sintered wick is actually poured in as a powder and then goes through a sintering process, so the powder’s sintered to the wall and sinters to the powder itself.
Graham Kilshaw:
Okay. Hey George, just a heads up, we’re up to 20 questions right now. I don’t know how far we can get, but let’s give it a shot, see how many we can get through. Next question comes from Mark Smith, “What’s a good rule of thumb for derating heat pipe Qmax based on flattening, bending, orientation, et cetera?”
George Meyer:
I think we touched on some of that in some of the earlier slides. Bending was about two and a half percent loss for every 45-degree bend. Flattening affects more the Qmax for the vapor than it does the wick structure. So, unless you flatten beyond the wick capability, flattening generally doesn’t have a big effect on the performance. But I think, like you said, we’ll come back and answer all of these in detail over the next few days, and send them out to everybody. Go ahead. Read the next-
Graham Kilshaw:
If we can touch on as many of them as we can and then we can do detailed answers for everybody, I think, after today’s event. We’re now up to 23 questions. Okay. Two questions here from Hazem. His first question is, “Can we derive an accurate lumped thermal model for heat pipes and vapor chambers?” I’ll let you answer that first, George, then I’ll go to his second one.
George Meyer:
I’m not quite sure what you mean by lump. Probably bulk, I’m assuming. Can we derive a accurate bulk thermal model? I’m not quite sure what you’re asking there, Hazem, but if I think I know-
Graham Kilshaw:
Let me try the second question, that might shed a little bit of light. His second question is, “Is CFD reliable to model vapor chambers of two-phase heat spreaders?”
George Meyer:
Yes and no. In other words, if we go through the calculation and I tell you what its effective thermal conductivity is in a W/mK number, then CFD is a perfect place to do it. If you’re trying to model the liquid flow in the wick structure, the boiling flow at the surface of the wick structure, and the vapor flow in CFD, that’s a bit more challenging.
Graham Kilshaw:
Okay. Let’s keep going. Next question comes from Mark at CI, “Why would I use one-piece versus two-piece?”
George Meyer:
Cost. It’s cheaper. Yeah. That’s the short answer. Yeah. Generally-
Graham Kilshaw:
We like the short answers as well as the long ones.
George Meyer:
Yeah. Yeah. Okay.
Graham Kilshaw:
Here’s another question on cost, comes from Ranjiv. Ranjiv asks, “On the cost reduction point of view, what are the parameters I need to consider to change from vapor chamber to heat pipe solution?”
George Meyer:
Okay. So, you have a vapor chamber solution and you want to change that to a heat pipe solution?
Graham Kilshaw:
Right.
George Meyer:
Their cost for the heat pipes themselves are generally lower in cost than for the vapor chamber. But you have to look at how many heat pipes you’re going to need, and then what’s the cost of installing those? Do I need to machine some fancy grooves into the base or are they just straight grooves? In other words, how much is my machining cost going to be? And then you have to bend the heat pipes and press them in. So, generally speaking, the heat pipe base heat sink is going to cost less than a vapor chamber, but not always. So, you have to look at the cost of integrating the heat bytes into the base, whereas the vapor chamber’s just one piece that’s put on as the base.
Graham Kilshaw:
Got it. Okay. And question from our friends over at Intel, “Do vapor chambers allow bends in the XY plane?”
George Meyer:
We can see from the pictures that we can bend them in the one plane, it gets much more challenging to bend them in the other plane. If I have a relatively long vapor chamber, I can do some bending before I flatten it. So, it is possible, it’s just not very common to bend the vapor chamber in the long beam direction. So, you can bend them in fin direction, but not in the width direction.
Graham Kilshaw:
Okay. Next up, question from Jack at [Eldip 00:42:41], “George what’s the best solder material and process to attach the HP or VC to an aluminum metal sheet?”
George Meyer:
Okay. The first thing you want to do is nickel-plate the aluminum sheet, just because it makes the soldering a whole lot easier. The next thing you have to be careful of when choosing a solder, most of the solders are above the boiling point of the working fluid. Let’s just use water, for example, when I’m going to solder above 100 degrees C, I start to have a positive internal pressure in the heat pipe or vapor chamber. Now, because the heat pipe’s round, it can withstand a positive internal pressure without much distortion. Whereas my vapor chamber’s a flat device; flat vessels are terrible pressure vessels, so as I get to say 105, 110, or 120 degrees C, that internal vapor pressure is going to want to make that nice flat vapor chamber turn back to round. What we do is we designed a fixture that holds that pressure, holds the vapor chamber flat, during the soldering process. And we have a little technical paper on how to do that, so we can also offer that to anybody who wants it.
Graham Kilshaw:
All right. I’m not sure we’re going to get through all the questions, they’re coming in thick and fast. We are up to 33 questions in this presentation. George, you obviously hit a good note with everybody today, so they got lots of interesting questions. If your question doesn’t get answered today, folks, don’t worry. We will get back to you either through the system or through email. We’ll connect with George, we’ll make sure everybody gets an answer one way or another. But for the benefit of the audience, let’s hit on a couple of other interesting ones. This one comes from Tomas at TOMAR electronics, “George, when you’re talking about result accuracy, are you using the centigrade scale, Fahrenheit, K? What is it?”
George Meyer:
I think most of us use degrees C. At least, in the thermal business, degrees C is the most common. I always use degree C.
Graham Kilshaw:
Okay. Tyler asks, “What encapsulated gases are used in most heat pipes?”
George Meyer:
Encapsulated gases? Well, hopefully none. All you want in there is the working fluid and the working fluid in its vapor phase. For example, water. There’s couple ways to process these things, I’ll just walk you through one, and it’s just like a refrigeration system; you put the liquid into the device and you draw vacuum on it. What that vacuum does is it draws the air out, it also then draws out a little bit of the water vapor, and then you seal it off. So, what you have is only the working fluid in there under a vacuum.
Graham Kilshaw:
Okay. Question comes from Steven at Rocor, “How much of a role do you think CFD will play in advancing the design of the internal structure of vapor chambers?”
George Meyer:
That’s a really good question. I think as you start to tweak these things to get out the last bit of performance, and as you start to look at different types of nanostructured wick structures and things like that. I think CFD is going to be key early on in the development of unique wick structures. But as far as using these things just to model the heat sink, that’s done every day. But it’s very uncommon to use CFD to model the internals of a heat pipe. So, I’m not quite sure if it answered the question, but I do believe as you develop new unique wick structures, CFD will come into play for that.
Graham Kilshaw:
Okay. Question from Raytheon, “Thank you, George, for the presentation. You mentioned large diameters are possible for heat pipes. On the order of meters, I believe you said. How large have vapor chambers been made?”
George Meyer:
I’ll just walk you through a couple of examples. I think the biggest one I’ve seen is probably the size of today’s TVs in the neighborhood of a 47-inch diagonal or a couple of feet by a couple of feet. That’s not a limit by any means. The longest heat pipes I’ve seen are probably on the order of 10 to 20 meters and probably the longest ones that are being used now are the ones for the Transatlantic pipeline where they’re keeping the ice frozen. So, as far as vapor chambers, how big can they be? I don’t think there’s a real limit, the largest ones I’ve seen have been between five and 10 square feet.
Graham Kilshaw:
Pretty good size. Pretty good size. All right. I think we can handle two more questions because we’re almost up to 40 questions. Don’t think I’ve ever hosted a webinar that had 40 questions, so that’s pretty impressive. So, let’s try the last couple of questions here, George. How do you decide the area and length of the evaporator and condenser sections?
George Meyer:
Well, the area and length of the evaporator is generally fixed by what you’re trying to cool. If I have a one by one inch heat source, well then my evaporators is one inch. The condenser size is generally determined by how much heat sink I need. If I need six inches of heat sink, well then that’s what you have for condenser area. So, the application generally dictates what those are.
Graham Kilshaw:
Okay. And our final question for today because we’re a little bit over time. This is quite an interesting one, comes from Tommy. Question in the application in data centers Tommy says, “This is for a data center application. If I have multiple heat sources on the server, can we route heat pipes from various heat sources to a central condenser?”
George Meyer:
Yes. It’s quite common to do. You just have to make sure that you have enough capillary action to move the liquid past that first or second heat source to the third heat source. So, it’s very common.
Graham Kilshaw:
All right. I’m sorry we couldn’t answer all the questions, but if you asked a question, you will get an answer. Either directly from George or through us in the next day or two. Just want to remind the attendees that the video recording of George’s presentation will be available in the next hour or two, very shortly. You’ll be able to log in through the same link that you watched the live version. And a copy of the slides will be available automatically in 24 to 48 hours from now. Again, you’ll receive an email from us to notify you when those are available. Thanks, everybody from all around the world attending today. We had a quick look at who was here today; Toshiba, Honeywell, Microsoft, Facebook, Qualcomm, and GE, to name just a few. It was a really great presentation, George. Thanks to everybody who attended. George, thanks to you. Great to host you again, and we look forward to seeing you here at Thermal Live another time. We’ll have another presentation on Thermal Live in the next hour. Thanks, everybody.
George Meyer:
Bye-bye
