Heat Pipe Design Fundamentals And Product Applications
Billy: Welcome, and thank you for joining us for today’s webcast, “Practical Thermal Management Solutions: Heat Pipe Design Fundamentals And Product Applications,” sponsored by Advanced Coolant Technologies and Tech Briefs Media Group. I’m Billy Hurley, associate editor with Tech Briefs Media Group, and I’ll be your moderator today.
Our webcast will last approximately 30 minutes and there will be a question and answer period at the end of the presentation. If you have a question you may submit it at any time during the presentation by entering it in the box at the bottom of your screen. Our presenter will answer as many questions as possible at the conclusion of the presentation. Those questions not addressed during the live event will be answered after the webcast. Also, early in the presentation there will be a poll question that we invite you to answer. In order to view the presentation properly please disable any pop-up blockers you may have on your browser.
At this time, I’d like to introduce our speakers, Brian Muzyka, Sales Engineer in ACT’s Electronics Products Group, who graduated with a BS in mechanical engineering from Penn State University. He worked as a research and development engineer for ACT’s aerospace group before taking on his current role. Brian has firsthand experience designing, analyzing and integrating heat pipes into real world applications.
Also on the line is Scott Garner, is vice president of the Electronics Products Group. Scott has 20 years experience in heat transfer and thermal management. This includes both technology and product development with an emphasis on integrating heat pipes and two-phase heat transfer systems into electronics cooling applications. He’s an inventor on 18 patents and has written several technical publications and shared numerous sessions on thermal management for a variety of professional conferences.
At this time, I’d like to hand the program over to our speaker, Brian Muzyka.
Brian: Thanks, Billy and thanks, everyone out there for joining us. My name is Brian Muzyka and I’ll be leading the discussion on heat pipes this afternoon. I’m also joined by Scott Garner, who will be joining me to answer some questions at the end.
I’m going to begin with a quick overview of what we’ll be talking about. The presentation includes when and where to consider heat pipes, some typical applications and reasons to use heat pipes. We’ll discuss heat pipe basic operating principles, the capabilities and limitations of heat pipes, typical uses and system level thermal performance expectations. We’ll discuss how to design and model heat pipe into your system and finally we’ll take a look at an example that compares heat pipe to other common cooling applications.
Here’s a look at some typical applications. Heat pipes are commonly found in electronic assemblies for commercial and military use. These can include power electronics, car chassis, portable electronics and a lot of numerous applications with similar thermal management needs.
Looking at the top figure, this shows a car chassis that originally had a liquid loop running through the middle to cool the electronics. This caused system reliability concerns because any leak would damage the electronics. To increase reliability, heat pipes were embedded into the vertical columns to spread heat to the top and bottom planes. Then the liquid loop was moved to those planes so that any leak would be isolated from the electronics. The bottom figure here shows a finned heat sink with enhanced heat transfer due to implementation of heat pipe.
Billy: Okay. This is Billy Hurley with Tech Briefs Media Group. At this point in the presentation we’d like to present a poll question to you. Have you or your company used heat pipes? If so, have you used them, A, Frequently; B, Occasionally; or C, Not yet? We invite you to answer the poll question at this time. Back to you, Brian.
Brian: Okay. Thank you for those who are participating in the survey. Now we’ll move on to look at when to consider heat pipes. Heat pipes can be used anytime conduction is a limiting thermal path. If you need to transfer heat along a surface, move heat from one point to another or isothermalize components, heat pipes are a practical solution.
The primary benefits of heat pipe include creating more efficient heat transfer and moving heat to more surface area, which allows you to decrease the overall size and weight of your system. Also by allowing you to move heat away from the electronics, you can increase the power input of your electronics while still remaining below the maximum junction or case temperature of your electronic component. Flexibility is also a big benefit because lots of times you can integrate a heat pipe solution without redesigning the system due to the bending and flattening capabilities of the heat pipe.
The figure on top here shows a folded fin heat sink for forced air cooling. Heat pipes were used to isothermalize the base to create greater heat dissipation by keeping all the fans at a similar temperature instead of having much hotter fans directly above the high powered electronics. The bottom picture shows a rotating heat pipe assembly where heat pipes were used to move heat from the top, along the copper fins to the bottom base plate.
Now we’ll discuss some of the basic operating principles of heat pipes. Heat pipes are passive two-phase heat transfer devices. The components of a heat pipe include the envelope material, the wick structure and the working fluid.
In this presentation, we’ll be talking about copper-water systems but different envelope materials and working fluids can be used in different operating conditions that span from cryogenic temperatures to liquid metal temperatures.
When heat is applied to one section, known as the evaporator, it vaporizes the working fluid and spreads to the cooler region known as the condenser. The vapor spreading is caused by the pressure gradient inside the heat pipe. At the cooler region, the vapor condenses to liquid giving up its latent heat. The liquid is then pumped back to the evaporator by capillary force created by the wick structure. Wick structures are typically groove, screen mesh or sintered powder metal. The choice of which materials is determined by the operating conditions of the heat pipe.
Continuing on with operating principles, heat pipes are vacuum devices. This allows the fluid to vaporize well below its boiling temperature. In a copper-water system, a heat pipe will transfer heat from 20 degrees C to over 200 degrees C. Heat pipes will operate with a temperature difference across their length. This is the driving force of the two-phase heat spreading. Because it is sealed it is essentially a pressure vessel and must resist internal and external pressures. The internal pressure is set by the saturation temperature of the fluid. Because heat pipes are always two-phase, the pressure directly follows the saturation curve.
This slide discusses some of the capabilities and limitations of heat pipes. The largest advantage is the extremely effective heat transfer. Heat pipes utilize two-phase heat transfer, however, it is often easier to look at them as conduction elements. Effective conductivity of a heat pipe can range from 10,000 to 200,000 watts per meter K. This range is dependent on the heat pipe length.
Heat pipes will have a similar delta T across their length. Longer heat pipes will have greater effective thermal conductivities. They also operate passively, which can inherently increase system reliability and lower operating costs.
The limiting factors of heat pipes include power limitations and heat flux limitations. The power limitations are governed by several heat pipe limits. The most notable limit for terrestrial applications is the capillary limit, which is the wick’s ability to return the liquid to the heat source. The maximum heat flux for a typical wick structure is between 50 and 75 watts per centimeter squared. However, custom wick structures have shown operation at much higher heat fluxes.
Here’s a look at heat pipe limits. This graph is a performance curve for a 10-inch long heat pipe. The capillary limit is the limiting factor for this heat pipe. As we discussed in the previous slides the capillary limit is the ability for the wick structure to overcome the pressure drop in the system and return the liquid to the evaporator. The largest pressure drop is typically caused by gravity. Therefore, orientation of the heat pipe will affect the maximum power the heat pipe can transfer. However, it will not affect the thermal resistance of the heat pipe. As long as the system is designed with the correctly sized heat pipe and the correct amount of heat pipe for the worst case-scenario, the thermal performance will not be affected by orientation.
ACT provides a free online heat pipe calculator which allows design engineers to input system conditions and will output the capillary limit for different diameter heat pipe. This application serves as nice, quick check to see if a heat pipe solution is feasible. The web address is listed here, and we encourage users to give it a try and see if a heat pipe could be a feasible solution.
This slide discusses using heat pipes as a way to transport heat. A very common application for heat pipes is to move heat from one spot to another. In typical electronic systems, real estate is valuable and it is often beneficial to keep electronics compact, which doesn’t lend for an easy heat sink design. By using heat pipes in these systems, you can move that heat to a location where there is more room for heat dissipation
Also, a lot of times, a current design that operates perfectly will move to higher-powered electronic components or smaller packaging. Without any thermal alterations these changes could cause overheating. Instead of redesigning the geometry of the entire system the implementation of heat pipes could solve the thermal management problem.
The top figure here shows a concentrated photo-voltaic application where the evaporator block in the center moves heat to the heat pipe that spreads in both directions to directly bonded fins for dissipation This particular heat pipe uses a central evaporator and has two condensers at either end.
Due to the operation of a heat pipe, wherever the hot spot of the heat pipe will cause vaporization and heat spreading to the colder region. It is not uncommon to have multiple condensers multiple evaporators in a single heat pipe.
The bottom figure here shows heat pipe soldered onto an aluminum box, which had electronics mounted on the backside. The heat pipe moved the heat to the top plane, which was liquid cooled, again, moving heat away from the components to a space where it’s easier to dissipate.
Now we’ll look at high K plates. Enhanced conductivity plates are aluminum conduction plates with embedded heat pipes. A bare aluminum plate typically demonstrates a thermal conductivity of about 180 watts per meter K. By embedding heat pipes, we have demonstrated thermal conductivities from 500 to 1200 watts per meter K. These values are real world values for bulk conductivity of the plate. They were found by modeling and testing the plates with real system inputs. After testing the temperature difference along the plate we went back to model and inserted the bulk conductivity of the plate and adjusted it until we matched the temperature gradient from testing.
Therefore, these thermal conductivity values are a fair estimate and a good way to get a quick check if a high K plate solution might be good for your system. With the increased thermal spreading capabilities, high K plates can help spread heat to the boundary of plates for liquid or air cooled chassis. In addition, high K plates can isothermalize to create higher fin efficiency and lower fin weight in enhanced fin heat sinks.
One of the most favorable benefits of high K plates, though, is that is that it does not harm the structural strength or the weight of your system moving from an aluminum heat spreader. The figure to the top was an IGBT application with natural convection cooling. It had the hotter component mounted to the right side which created hot fins on the right but colder fins on the left. By embedding heat pipes under the fin we were able to effectively double the surface area for heat dissipation.
The lower figure was a shipboard electronics with mounted electronics. The heat pipes were ultimately routed to spread heat to and along the edge of the plate, which was liquid cooled.
This is a model of that plate you saw in the last slide. The model on the left shows the plate as bare aluminum. The three high-powered components were causing hot spots that would cause failure in the system. The heat pipe layout was chosen to move heat to the liquid cooled edge and along the edge to lower the heat flux. This helped reduce the hot spot temperature by nearly 20 degrees, as you can see here.
Now I’ll talk about designing with heat pipes. When designing with heat pipes, there are a couple tricks that may be helpful as a quick way to determine your overall thermal resistance. The two main resistances associated with heat pipes are the radial resistance and the axial resistance.
The radial resistance is the conduction through the wall, the wick and the two-phase heat transfer at the vapor-condensate interface. This value is typically around 0.2 degrees C centimeters square per watt. The axial resistance is the vapor temperature difference across the length due to the temperature across the length due to the internal pressure difference. This is typically a very low number, about 0.02. The easiest way to view heat pipes, though, is the overall temperature gradient, which is typically around 2 to 5 degrees C across its length.
Now I’ll talk a little bit about modeling techniques for heat pipes. The simplest way to model a heat pipe is to use a solid conduction element as your heat pipe. The first thing you want to do is check the power and heat fluxes of your system and choose the correct diameter and amount of heat pipe.
Using ACT’s free online calculator is a good way to check the power limits for copper-water systems. Next you insert a solid rod into your system as the heat pipe and increase the conductivity until you get about a 3-degree temperature difference across its length. The solid conduction technique can also be used for high K plate designs. If you insert a value of around 600 watts per meter K, you’ll get a good idea of the heat spreading for a high K plate design.
Some other modeling techniques include using a complex model where you model the liquid vapor interface, envelope wall, wick structure and vapor space separately. This is usually done for smaller models with high heat flux and custom wick structure.
There is also a mixed model, which is to lump the interface wall and wick material into one material and use a very highly conductive vapor space. All these models can be effective but usually the best thing to do is get a quick sanity check using this simple model and call ACT when considering some more advanced system level modeling.
Here’s a look at heat pipe design guide. This slide takes a look at implementing heat pipes into your system. Typical sizes for heat pipes are three, four, five, six and eight millimeter outside diameter tube and one-eighth, one-fourth, three-eighths and half-inch OD tube.
These pipes can be made out of any size though. These are just the most common because they are the most readily available materials. When routing heat pipes there are a couple of good guides for minimum bend radius and minimum thicknesses. Typically heat pipes are bent to about three times the outside diameter of the heat pipe as the minimum bend radius. This can be exceeded but anything tighter may cause wrinkling in the copper. When flattening heat pipes, a good value to use for minimum thickness is two-thirds the outside diameter of the heat pipe. Once again, this is just a guide. Flattening heat pipes is also a good way to increase thermal contact at the evaporator section as well as embed into thin aluminum plates.
When integrating heat pipes into assemblies, the primary attaching methods are solder or thermal epoxy. This allows for a quick and easy integration process. Copper-water heat pipes are a proven technology with ample life test data. ACT has also proven and tested heat pipes in systems that include shock, vibration, acceleration and frozen start-up. Available on our website is a heat pipe product reliability guide with more information on these tasks.
The figure on the top here shows a design where heat pipes are flattened and epoxied in between the folded fins and base plate to isothermalize the fins for better dissipation And the bottom figure shows a look at the bending capabilities of heat pipe.
Now I’ll take a look a quick electronics enclosure example. A user has a power source dissipating 100 watts of power and a resistor dissipating 75 watts of power. The two components are located at opposite ends of the enclosure, which acts as a heat sink. The electronics are overheating at steady state operation. Taking a quick look at the figure, it seems the main cause for overheating is due to a lack of conduction in the enclosure material. Some feasible options may include increasing the enclosure size, adding a liquid cooling element, adding fans to the system or embed heat pipes into the enclosure.
Here’s a look at the pros and cons of each solution. By increasing the enclosure size you add surface area for dissipation and increase the thermal mass. However, you also increase the weight and volume of your system and may [inaudible 00:18:44] hot spot [inaudible 00:18:45] conduction gradients. By adding a liquid cooled cold plate you will get effective cooling that is scalable and can be customized to the location of your electronics.
Again though, you’ll increase the weight and volume and you’ll also move to an active system, which can introduce other failure modes and operating costs. Adding a fan or multiple fans inside the enclosure is an inexpensive way to cool the components. Again though, it adds system reliability concerns and introduces noise to the system.
Finally, when using heat pipes you can move heat to additional surfaces for dissipation, which will reduce the hot spot temperatures of your electronics. You can route the heat pipe custom to where your electronics are in the design and get optimal thermal performance and you don’t incur any additional weight or volume penalties. The outcome is a reliable system, which by the look and feel is no different than the currently designed enclosure.
Hopefully that example demonstrated just how effective heat pipes can be in a system level design. The main points that we’d like to leave you with are that heat pipes are passive, reliable and long lasting. They’re a very good heat transport device when designed correctly. They can be a quick fit without redesigning the system if there are ever power increases or geometric changes that could cause thermal management concerns. They can be manufactured and low and high volume. They’re capable of working in any orientation if designed correctly and can really work in a lot of different systems, even with vibrational or acceleration concerns. They’re also easy to attach and integrate into designs.
We’d like to thank everyone for taking the time to join us today. I think we have about five to 10 minutes left so we’ll be taking some questions. Once again, thank you all for joining us.
Billy: Thanks, Brian. At this time we’d like to begin our Q &. Again, if you have any question you may submit by entering it in the box at the bottom of your screen. Here’s our first question Brian. Can copper-water heat pipes be used in applications where freezing is a concern?
Scott: Hi, this is Scott. I’ll answer the question. Yes, they can be. Water freezes at zero degrees C obviously, so if you’re operating in an environment that goes below zero the heat pipes simply cease functioning then. Usually there is enough secondary path that you don’t need the heat pipe to function during this very cold period. For example, on an outdoor telecommunications enclosure and you’re dissipating heat when it’s very cold outside, secondary paths will cool the electronics efficiently. You really only need the heat pipes when you’re at the high end ambient.
Although they won’t transfer heat when they’re below 0, no damage occurs. When it warms again they start functioning fine and when you go into a hot ambient they’ll be transferring heat and cooling your electronics.
Billy: Here’s a question from one of our attendees. Can you confirm that performance is not affected by orientation relative to gravity?
Scott: Sure. The orientation of the heat pipe relative to gravity is a determining factor in how much power that heat pipe can carry. If you have a 6-inch long, quarter-inch diameter heat pipe that’s fully gravity aided, and when we say gravity aided the condenser is located above the evaporator. So when you input heat and boil the water it vaporizes and travels up to the condenser and condenses. In that case, gravity is helping to return the fluid in addition to the pumping capability of the wick.
A quarter-inch diameter gravity aided heat pipe might be capable of carrying a 100 watts whereas if it was fully against gravity and it had to pump the fluid back, in that case it might only be able to carry 60 to 75 watts. So it is a factor in how much power the heat pipe can carry but if you design your system for worst-case orientation you will always have excess capacity in any other orientations it may see.
The good thing is once you design for that condition, orientation will not affect the thermal resistance or the gradient of a heat pipe. If you design it and it’s operating fully against, it will have the same delta T in any of the other orientations.
Billy: What is the effective conductivity?
Scott: It really varies. I think Brian had a slide in here that showed effective conductivity from 10,000 to 200,000 watts. If the question is, “What is the term effective conductivity,” if you’re modeling this in Solid Works or one of your finite elements or CSD packages, it is the conductivity you can apply to the heat pipe to get the effective gradient across the heat pipe. Since heat pipes are two-phase heat transfer devices, it’s not a very realistic design guide but it’s very helpful in modeling. It really depends on the length of the heat pipe. As we said, it can range from about 10,000 watts per meter K to about 200,000.
What you need to do is pick a value, maybe about 50,000 watts per meter K, to put in your model and then run your model and you look for a temperature gradient across the heat pipe. A well-designed heat pipe operating within its capabilities will typically have about a 2 to 4 degree C delta T across its length. Pick a number in that range and run it. If you’re getting a 10-degree gradient then you need to increase that effective conductivity in your model. If you’re only getting a 1-degree gradient you may want to decrease it a little bit.
When you run your models put a value in to simulate the effective conductivity of the heat pipe. When you’re done go back and look at your results and you want that gradient across the heat pipe to be about 2 to 4 degrees C. You can adjust that accordingly.
With our high K plates, as Brian talked about, they range in value from about 600 watts per meter degrees C to 1000 watts per meter degrees C. A real quick check to an existing heat sink or spreader plate and, say, it’s aluminum with conductivity of 180 watts per meter degrees C, you can just click on that, change the material property to somewhere around 700 to 800 watts per meter degrees C. That’s the simplest check to determine if integrating heat pipes will help you in your assembly.
Billy: What could happen if heat addition exceeds 100 watts per centimeter squared?
Scott: Typical heat pipes with standard wick structures will handle about 75 watts per square centimeter heat flux. If you start driving one higher than that you’ll get a localized dry out at the evaporator, you’ll get a higher gradient. When you decrease the power again it will re-start and re-wet and run fine. We do custom design wick structures that we’ve tested up to over 500 watts per square centimeter, very high heat fluxes. If you know it’s a condition that’s going to happen often and it’s a design condition, we can design a wick structure to handle that. If you’re doing with just a traditional heat pipe or standard wick heat pipe, it will dry out temporarily when you’re inputting that higher heat flux into the assembly.
Billy: Here’s another question from an attendee. While shaping or bending or flattening heat pipes, do you damage the wick structures? If not, to what extent can one flatten while soldering into chassis?
Scott: Almost everything we do is bent or flattened in some configuration. There is a design guide on our website. If you go to our website and click on “References,” you’ll get that design guide. We typically try and keep the bending to about three times the tube diameter and flattening to about two-thirds of the original tube diameter.
It does not damage the wick structure. The wick structure is a soft, conformal material inside the pipe. It goes along for the ride when you bend and flatten. We can exceed those values listed in our design guides but it’s more challenging. If you stick with the values listed there you should be fine with any designs you’re working on.
Billy: What’s the material option for the heat pipe enclosure?
Scott: Heat pipes need to be compatible with the envelope, the wick structure and the working fluid, that’s the materials system. Primarily, for most terrestrial electronics cooling, copper-water is the preferred material system, where we have copper tube, copper wick structure and pure water as the working fluid.
For some satellite applications that is typically aluminum-ammonia. However, ACT makes a variety of heat pipes, from very low temperature cryogenic working fluids up to high temperature where we’re using some super alloy stainless steel with liquid metals as the working fluid, operating around 1000 degrees C. It really depends on the operating temperature you’re trying to run at, but for electronics cooling it’s primarily copper-water systems.
Billy: Okay. We are out of time. Just a reminder, those questions that were not addressed during the live event will be answered after the webcast. That will conclude today’s webcast and thanks, Brian and thanks, Scott and thanks to everyone for joining us. Just a reminder, this webcast will be available on demand at www.TechBriefs.com for the next 12 months. Have a great day. Thanks.