Select Page

## Design Considerations When Using Heat Pipes (Pt. 2)

Now that we’ve determined which heat pipes can effectively carry the required heat load (Pt 1), we need to ensure we select and size the heat exchanger to dissipate the heat into the surrounding air -based on available air flow. As mentioned in the prior post, this article is not intended to provide detailed training/equations on the proper design of heat pipes and heat sinks, but rather to offer guidance on the number and size of heat pipes used as well as to provide some tips for estimating heat sink size and determining attachment methods of the sink to the PCB.

## Types of Heat Sinks

There are numerous choices from zipper pack fins to extruded fin stacks, each with their own cost and performance characteristics.  While the heat sink choice can markedly affect heat dissipation performance, the biggest performance boost for any type of heat exchanger comes with forced convection.

Table 1: Types of Heat Sinks

## Estimating Required Heat Sink Volume

By using the simple equation below, we can estimate the overall volume of the heat sink required to cool a heat source.

V= (Q*Rv)/Delta T

Volume is the outer dimensions of the heat sink. Q is power in watts and Rv is the volumetric thermal resistance (cm3-C/W). Table X provided some well established guidelines for the latter under different air flow rates.  The first thing to notice is the enormous positive affect forced air has on thermal performance. Even with a gentle air flow, volumetric thermal resistance of the heat sink is reduced by two-thirds.

Table 2: Air Flow vs Volumetric Thermal Resistance

Let’s say we have a 50W heat source (Q) who’s specified maximum operating junction temperature (Tjunction) is 100o C. Further, let’s assume we’re designing this device to operate in a max ambient temperature of 70o C – a rugged environment. By subtracting the ambient temperature from the maximum allowable electronics temperature, you arrive at the delta t: in this example it’s 100 – 70 = 30o C delta t.

If we are designing a rugged computer, for instance, we’ll probably have a semi or fully sealed enclosure meaning that our heat sink needs to be designed for natural convection. The resulting heat sink volume would be:

V = (50*500)/30= 833 cm3 = a really large heat sink of 9.4 x 9.4 x 9.4 cm. Obviously, we wouldn’t have a heat sink with 9.4 cm high fins in a natural convection environment so the length and width dimensions would have to grow substantially.  But, we might surmise that this application would probably need to use the chassis lid as the heat sink.

There’s an example of just such an application that I talked about in a prior post on ruggedized electronics and whose heat load, ambient operating temperature, and fanless design match the above example.  Let’s look at just the lower half of the design. The bottom side of the PCB contains the processor and chipset at roughly 50W. A copper plate in conjunction with two heat pipes are used to spread and move heat to the chassis case, whose published dimensions are 25cm (L) X 15cm (W).  We know the overall thickness of the assembled case is 7.5 cm and it looks to me the top and bottom lids are each roughly 2.5 cm thick. This would result in an overall volume of 938 cm3 which is within 15% of the required 833cm3 we calculated earlier.

Figure 1: PerfecTron MIL-810 Computer

Had this same heat source been used in conjunction with forced air, the required heat sink volume would have been reduced by up to ten fold.

Although beyond the scope of this article, the next steps in heat sink design would be some detailed excel modeling followed by CFD analysis. This will enable engineers to fully understand optimal base thickness, fin pitch, fin height, and base/fin material among other things.

## Mating Heat Pipes to the Condenser

Whether we’re dealing with a heat exchanger that is local or remote to the heat source, the options for mating a heat pipe(s) to them are identical.

### Grooved Base

It should go without saying that simply soldering a round pipe to a flat surface is far from optimal.  Circular or semi-circular grooves should be extruded or machined into the heat sink. It’s advisable to size the grooves about 0.1mm larger than the diameter of the heat pipe in order to allow enough room for the solder.

The heat sink in Figure 2 uses both a local and remote heat sink.  The extruded heat exchanger is design to accommodate slightly flattened heat pipes, helping to maximize the contact between the copper mounting plate and the heat source. A remote stamped fin pack is used to further increase thermal performance. These types of heat exchanger are particularly useful because the pipes can run directly through the center of the stack, decreasing conduction loss across the fin length.  Because no base plate is required with this fin type, weight and cost can be reduced. Again the holes through which the heat pipes are mounted should be 0.1 mm larger than the pipe diameter.

Had the pipe been completely round at the heat source, a thicker grooved mounting plate would have been required as seen in Figure 3.

Figure 2 (L): Heat Pipes in Grooved Base     Figure 3 (C): Heat Pipes in Grooved Mounting Block     Figure 4 (R): Direct Contact Heat Pipes

### Direct Contact

If conduction losses due to the base plate and extra TIM layer are still unacceptable, further flatting and machining of the heat pipes allows direct contact with the heat source as seen in Figure 4 above. Performance gains from this configuration usually lead to between a 2-8o C reduction in temperature rise.  In cases where direct contact of the heat source to the heat pipes is required a vapor chamber, which can also be mounted directly, should be considered due to its improved heat spreading capacity.

## Attaching the Heat Sink to the Heat Source

The primary reason for considering a heat pipe solution is improved performance. As such, I’m not going to talk about the use of thermal tape or epoxy as the primary means of attaching the heat sink to the die. We generally stick with one of several mechanical attachment methods where we can meet Mil or NEBS shock and vibe requirements. Additionally, the thermal resistance of thermal interface material (TIM) improves as pressure between the die and the heat sink is increased.

An inexpensive attachment method for small (low mass) heat sinks are stamped mounting plate. Although it requires two PCB holes, this method offers better shock and vibe protection thermal tape or epoxy and some TIM compression – although still only 5 PSI.

Figure 5 (L): Stamped Mounting Plate           Figure 6 (C): Spring Loaded Push Pins                Figure 7 (R): Spring loaded Metal Screws

Figure 6 shows spring loaded plastic or steel push pins further increase TIM compression up to around 10 PSI. Installation is fast and simple but removal requires access to the back of the PCB. Push pins should not be considered for anything more than light duty shock and vibe requirements.

Spring loaded screws offer the highest degree of shock and vibe protection (Figure 7) as they are the most secure method of attaching a heat sink to the die and PCB. They offer the highest TIM preload at roughly 75 PSI and work very well with heavy heat sinks.

Celsia is a custom heat sink supplier specializing in heat pipe and vapor chamber thermal assemblies.  We are a U.S. based company, with wholly-owned Taiwan manufacturing facilities, that prides itself on solving difficult cooling challenges under tight deadlines.  Please contact us to schedule a thermal review of your next application.

## Design Considerations When Using Heat Pipes (Pt. 1)

This blog post offers some intermediate level heat sink design guidance when using heat pipes for the most prevalent types of electronics applications: mobile to embedded computing and server type applications with power ranging from 15-150 watts using processor die sizes of between 10x10mm to 30x30mm.  I’m using these constraints because throughout the article I’ll mention some simple tips or guidelines that don’t necessarily apply when talking about power electronics applications.  I’ll also be focusing on the most ubiquitous type of heat pipe – copper tube with sintered copper wick using water as the working fluid.

As you might have read in past blogs, heat pipes should be considered when the thermal design is conduction limited or when non-thermal goals such as weight can’t be achieved with other materials such as solid aluminum and/or copper.  So, if you’re too the point where heat pipes may be an option, keep the following things in mind.

## Thermal Conductivity of Heat Pipes Rarely Reach Published Limits

Regularly published heat pipe thermal conductivity range from 10,000 to 100,000 W/mK. That’s up to about 250 times that of solid copper and 500 times that of solid aluminum.  But don’t count on those figures for typical electronics applications.  Unlike solid metal, the effective thermal conductivity of copper heat pipes varies tremendously with heat pipe length, and to a lesser degree with other factors.

Figure 1 illustrates the effect of length on heat pipe thermal conductivity. In this example, we used three heat pipes to transport heat from a 75 watt power source. While thermal conductivity of 10,000 W/mK is achieved at just under 100mm heat pipe length, a 200mm length has less than one-third the typically published maximum thermal conductivity of 100,000 w/mK.

Figure 1: Heat Pipe Thermal Conductivity Changes with Length

## Heat Pipes Are Like Automobiles

Performance data for heat pipes are usually of the Honda Accord variety: adequate for most things, but horrible if you have a specific purpose in mind – like track racing or off-roading.  Even when limiting our conversation to copper/water/sintered wick versions, heat pipe customization can markedly affect operational and performance characteristics.  And just like cars, design changes to a heat pipe range from the simple to the exotic.

As we’ll see later, flattening and bending will negatively affect heat pipe performance but are an easy customization that allow engineers to navigate the often crowded route from the evaporator the condenser. If the application requires lower thermal resistance than a typical heat pipe assembly provides, the heat pipes can be machined to create a very flat surface capable of making direct contact with the heat source, eliminating the solid metal base plate and extra TIM layer.  A heat pipe whose diameter is different from one end to the other may be needed so that the evaporator end completely covers the heat source while the rest of it remain small enough for tight bends. These changes allow engineers to physically connect the heat pipe to mating components from the heat source to the condenser.  But what if the heat pipe still can’t meet power handling or orientation requirements?

Changes to the internal structure of the heat pipe, most notably wick porosity and wick thickness, allow heat pipes to be tuned to meet specific operating parameters and performance characteristics.  For instance, when a given diameter heat pipe is required to operate at higher power loads or against gravity, the capillary pressure in the wick needs to increase.  For higher Qmax, this means a larger pore radius. For effectively working against gravity (condenser below evaporator), this means a smaller pore radius and/or increased wick thickness.  Additionally, it is possible to vary both wick thickness and porosity along the length of a single tube. Suppliers who specialize in heat pipe customization will regularly use custom formulated copper powders and/or unique mandrels to ensure the final product meets applications requirements.

## Mechanical Engineering Choices Affect Heat Pipe Performance

Once the internal structure of the heat pipe has been determined, it’s good to know how outward physical characteristics affect performance.  With heat pipes, size generally matters most. However, changes to outward design will degrade performance for any given heat pipe.

Flattening

Figure 2 shows power handing capacities (Qmax) for the most common heat pipe sizes. As noted earlier, Qmax may change amongst vendors for standard (Honda Accord type) heat pipes, but I’m taking some data from recent projects we’ve done – without getting too exotic.

Figure 2: Heat Pipe Qmax is Customizable

Typically, sintered copper heat pipes can be flattened to a maximum of between 30-60% of their original diameter.  I know, some will tell you it’s the lower figure before the centerline starts to collapse but it’s really a function of technique – we regularly flatten our one-piece vapor chambers, which begin life as a very large heat pipe, down to 90%+. I wish I could provide a rule of thumb for how much performance will degrade for every 10% decrease in thickness, but it would be irresponsible.  Why?   The answer comes down to how much excess vapor space is available before the heat pipe is flattened.

Simply put, there are two performance limits important for terrestrial heat pipe applications: the wick limit and the vapor limit. The wick limit is the ability of the wick to transport water from the condenser back to the evaporator. As mentioned, we can tune the porosity and thickness of the wick to specific applications, allowing for changes to Qmax and/or ability to work against gravity.  The vapor limit for a particular application is driven by how much space is available for the vapor to move from the evaporator to the condenser.

The orange and blue lines in Figure 3 plot the respective limits for the various heat pipe sizes shown in Figure 2.  It’s the lesser of these two limits that determine Qmax and as shown the vapor limit is above the wick limit, albeit only slightly for the 3mm heat pipe.  As heat pipes are flattened, the cross sectional area available for vapor to move is gradually reduced, effectively moving the vapor limit down. So long as the vapor limit is above the wick limit, Qmax remains unchanged.

Figure 3: Heat Pipe Wick & Vapor Performance Limits

In this example, we’ve chosen to flatten the heat pipes to the specifications in Figure 2. As seen by the gray line, the vapor limit is below the wick limit, reducing the Qmax.  Flattening the 3mm by a mere 33% causes the vapor limit to become the determining factor whereas we had to flatten the 8mm by over 60% for this to happen.

Bending

Bending the heat pipe will also affect the maximum power handling capacity.  Here again I’ll offer a couple rules of thumb. First, minimum bend radius is 3x the diameter of the heat pipe. Second, every 45-degree bend will reduce Qmax by about 2.5%. From Figure 2 above, an 8mm heat pipe, when flattened to 2.5mm, has a Qmax of 52 watts. Bending it 90 degrees would result in a further 5% reduction. The new Qmax would be 52-2.55 = 49.45 watts

For more technical information on bending both heat pipes and vapor chambers, please visit a past blog on how bending affect heat pipe performance.

Working Against Gravity

Let me start this section by saying that we’re using the same fairly ‘standard’ heat pipes that we’ve shown in prior examples so they haven’t been optimized to work against gravity (-90o).  Figure 4 illustrates how the relative position of evaporator to condenser can affect both Qmax and heat pipe selection. In each case, Qmax is reduced by roughly 95% from one orientation extreme to the next.

Figure 4: Affect of Orientation on Heat Pipe Performance

In situations where the condenser must be place below the evaporator, we would use a sintered material that allowed for smaller pore radius and/or increase the wick thickness.  For instance, if we optimized the 8mm heat pipe for use against gravity (-90o), we could raise its Qmax from 6W to 25W.

## Heat Pipe Calculations: A Practical Example

Let’s use an example to illustrate how heat pipes might be used to solve a thermal challenge.  Assume the following as know constraints: 70W heat source with dimensions 20 x 20mm and a single 90 degree heat pipe bend required to transport heat from evaporator to condenser. Further, the heat pipes will operate in a horizontal position.

• To be at their most effective, heat pipes need to fully cover the heat source, which in this case is 20mm wide. From Figure 2, it appears we have two choices: three round 6mm pipes or two flattened 8mm pipes. Remember the three 6mm configuration will be placed in a mounting block with 1-2mm between the heat pipes.
• Heat pipes can be used in conjunction to share the heat load. The 6mm configuration has a Qmax of 114W (3x38w), while the flattened 8mm configuration has a Qmax of 104W (2x52w)

Figure 5: Heat Pipe Configuration Choices

• It’s just good design practice to build in a safety margin. You wouldn’t go bungee jumping with a rig only capable of supporting exactly your weight, would you? Same for heat pipes. I typically use 75% of their maximum capacity. So we’ve got 85.5W for the 6mm (75% x 104W) and 78W for the 8mm (75% x 104w)
• Now we need to factor in bending. A 90 degree bend will cause Qmax of each configuration to be reduced by another 5%. The resulting Qmax for the 6mm configuration is now just over 81W and for the 8mm configuration it’s now 74W, both of which are higher than the 70W heat source we’re trying to cool.

As can be seen, both configurations are adequate to move heat from the evaporator to the condenser. So why choose one over the other? From a mechanical perspective it may simply come down to heat sink stack height at the evaporator. As you can see, the 8mm configuration has a much lower profile than does the 6mm configuration.  Conversely, condenser efficiency may be improved by having heat input in three locations versus two locations, necessitating the use of the 6mm configuration.

Please check back for Part 2 of this blog where I’ll be discussing condenser types and sizing as well as methods to attach the heat pipe assembly to the heat source.

## Rugged Electronics: Thermal Challenges and Heat Pipe Solutions

The challenges faced by thermal engineers tasked with cooling devices designed for harsh environments are often extensive. This blog post is meant to provide an overview of ‘rugged’ standards as well as address some thermal management techniques being used.

Electronics for both military, marine and industrial applications are exposed to a variety of harsh conditions from extreme temperatures, wet or dusty environments, and situations where components are subject to repeated shock or exposure to corrosive substances.

Thermal engineers designing rugged systems must content with enclosures/chassis that offer little to no airflow and reliability requirements that limit design options. Here are a few of the US standards used to define whether ruggedized devices can stand up to harsh environments.

One the oldest and most widely used US standard is a military specification called MIL-810. It’s been used by all Department of Defense agencies since with 1960’s (with numerous revisions) and is regularly adopted by the private sector for use in commercial/industrial applications.  One of the advantages of MIL-810 is that test requirements, duration and sequences are clearly defined.  There are around thirty test categories ranging from the common like temperature and water resistance to the niche like pyro-shock and acceleration, but equipment is rarely, if ever, designed to pass all of them.

The two most often mentioned ‘rugged’ standards for the industrial and commercial markets are NEMA-250 and IP. Issued by the National Electronics Manufacturers Association (NEMA), specification 250 details a protection rating for electronic enclosures which ranges from limiting access to hazardous system components to degrees of protection from dust, water and corrosives.   The Ingress Protection Ranking (IP Code) rates devices on the degree of solid particle and liquid protection as well as impact protection. Additionally, several ANSI accredited VITA standards specify environmental requirements for different classes of COTS plug-in modules.

While we’ve touched on a few US standards/specifications for harsh environments, it should be noted that that there are many others from the extensive German DIN standards to STANAG standards used by NATO defense forces.

Irrespective of the specific ‘rugged’ standard, thermal engineers need to be creative with how heat is managed in electronic devices designed for harsh conditions. Depending on the required range of operating temperature and level of protection against vibration, shock, dust, sand, and or water rugged electronics chassis and thermal management design can take on many forms.  Let’s use a familiar example: a general purpose computer with Intel Core i7 processor and some level of internal expandability.

The first is a Comark Enduranode PC certified under IEC-60945 for marine use in a protected environment like the bridge of a ship. Test requirements/ranges are beyond that of a consumer PC and include Environmental, Safety and Electromagnetic: dry and damp heat, low temperature, vibration to simulate propeller and wave slamming, corrosion resistance, noise levels, limited access to high voltages within the chassis, and improved protection against electromagnetic interference.

Figure 1: Rugged Marine Computer from Comark

Figure 1 shows the outside of this unit. It uses the small form factor chassis as a heat sink incorporating an external extruded aluminum fin stack with a corrosion resistant coating.  A solid state hard drive was chosen for reliability and resistance to vibration reasons. Additionally, all electronics are conformal coated to protection against moisture, dust, chemicals, and temperature extremes – in this case -20 – +50oC without fan and up to +60oC when the internal system fan is active.

The unit is vented on both sides and incorporates an on-demand fan but because I couldn’t find any internal pictures of this unit, I’ll make some assumptions about how heat is being managed. In addition to the already mentioned conformal coating on all electronics, heat generated from the 50 watt combined CPU and chipset could be spread using an aluminum baseplate, possibly with a low profile fin stack directly above it. Copper heat pipes are almost certainly used to transport heat to the finned chassis lid.

Figure 2: Heat Pipe Heat Sink

Figure 2 is a rendering of a heat sink for a different product (Logic GL-L05) with similar dimensions and thermal design, but could be representative of what’s in the Comark computer. Additionally, the power source may have its own fan and/or be a high efficiency version which generates less heat.

Let’s compare this to a more rugged MIL-810G computer designed to operate in environments from -40oC – +70oC (without CPU throttling) and withstand mechanical shock, temperature shock, and vibration as set forth in the MIL specification. The chassis is fanless and non-vented although not completely sealed due to the I/O ports, however it is tested for dust resistance. The 50 watt combined CPU and chipset are soldered on-board as is the solid state drive for improved shock and vibration capability.  While the electronics and power of this system are comparable to the Comark, the PerfecTron SR100 has an external volume roughly half that of the Comark unit.

Figure 3: PerfecTron MIL-810 Computer

For this example, I won’t have to speculate on the thermal solution as an exploded diagram was available from the PerfecTron website (Figure 3). Residing on the top side of the PCB are I/O functionality and the power source to which a copper heat spreader is attached. The spreader mates to two copper heat pipes embedded into the base of the upper heat sink lid. The CPU and chipset, located on the opposite side, are similarly attached to a copper spreader with heat pipes embedded in a lower heat sink lid.

While I’ve chosen rugged designs that both use heat pipes, other options are available to thermal engineers depending on the ambient temperature as well as the size, weight and power goals (SWaP) of the rugged electronic device. Here’s a concise article from Military Embedded Systems Magazine on Airflow Through, Liquid Flow Through & Spray Cooling.

## Understanding 3D Vertical Vapor Chambers

Until the mid-2000’s heat pipes were really the only option when designing a high heat-flux heat sink whose condenser was oriented vertically above a horizontal heat source. That’s because they’re easily bent post-production into a variety of shapes including “L” or “U” configurations. Despite the advantages of vapor chambers, better spreading and less thermal resistance, these devices were for all practical purposes non-manufacturable in these shapes. Why?

A traditional vapor chamber, what we at Celsia call a two-piece design, is manufactured by stamping an upper and lower plate (thus the two-piece moniker), sintering copper powder to each side, and diffusion bonding each plate to one another on all sides.

Typical two-piece vapor chamber designs look like this:

As seen in the images above, this type of vapor chamber is customizable into virtually any shape along the X&Y planes (length and width). Additionally steps or pedestals can designed into the stamped upper and/or lower plates, making them great for recessed heat sources or those that need to conform to a surface of varying heights. But, how could a two-piece production method create a 3D vertical vapor chamber like an “L” or “U” shape? Well, it wouldn’t be easy, fast, or very cost effective.

•  First, the upper and lower plates would need to be stamped into that shape or stamped flat and bent.

•  Second, specialty L/U shaped fixtures would need to be fabricated to hold the copper powder to the walls of each plate as they’re being oven sintered, and don’t forget these fixtures will take longer to heat up increasing the overall sintering cycle time. Another potential problem here is that the fixtures need to be made such that they insure a very even distribution of copper powder across the surface and also at the bend. Variations in sintered copper can lead to lower yield rates and/or finished parts that don’t perform according to specification.

•  Third, the two plates need to be attached together either through the traditional diffusion bonding process (back into an oven) or through a four sided TIG welding process. The former is not known to work well on complex, especially bent, shapes while the latter would require an operator to manually weld all sides together.

In the 1980’s, I remember working on some military applications that overcame these issues by combining vapor chambers for the horizontal portion of the sink with integrated copper tubes for the vertical portion (see my crude drawing below L). Here we cooled two 2kW traveling wave tubes for a radar application. The second image incorporated a 75x75mm vapor chamber with a one meter fin stack for a 3kW application.

Vapor Chamber with Copper Tubes – 4kW Application (L), 3kW Application (R)

Although technically not a 3D vertical vapor chamber, this solution is excellent for very high heat-flux applications provided adequate structural support is done at the condenser/ tube areas to address shock and vibe concerns. Structural issues arise because a lot of leverage can be placed on the vertical pipes/condenser causing the copper tube braze joints to fail or cause damage to the tubes or vapor chamber itself.

So, is there a cost effective, process-optimized method of producing a true 3D vertical vapor chamber? About ten years ago, we began work on a design (manufacturing process) that in addition to cost savings provided designers with some unique options; most relevant to this article, the ability to bend it post-production to create L/U shapes.

The design we created was fairly straightforward. We call it a one-piece vapor chamber and it’s available from a growing number of suppliers.

One-Piece Vapor Chamber Bent Into Various Shapes

•  Rather than using the traditional two-piece approach beginning with upper and lower stamped plates, we start with a very large diameter copper tube to create a heat pipe of between 15-75mm OD.

•  A simple cigar shaped mandrel is inserted and copper power is layered in between it and the inside wall of the tube. This insures a very even sintering process and keeps oven cycle times to a minimum.

•  Next, the tube is flattened to the desired thickness, usually between 2-3.5mm, and a perforated wave-shaped spacer is inserted into the tube. This spacer allows the device to withstand clamping pressures up to 90psi and provides internal support for bending post-production.

•  Deionized water is added and the tube is vacuum sealed and TIG welded – only on the two narrow sides of the device. So now we have something that looks like an extremely wide flat heat pipe, but with an internal support structure.

•  Lastly, the device can be bent to form a variety of shapes including L/U configurations.

Here’s an example of a high power LED application we did in 2008. The horizontal portion of the vapor chamber makes direct contact with the heat source, eliminating a mounting block and TIM layer as would be typical with a heat pipe implementation. The vertical portions of the vapor chamber run directly into the condenser with a continuous and evenly distributed sintered copper layer along the inside wall. For the this application, vapor chamber width was only 30mm or so, but one-piece vapor chambers can be created as wide as 110mm.

One-Piece 3D Vertical Vapor Chamber – HBLED Application

In summary, a traditional two-piece design is really limited to steps or pedestals of 4-5mm in height. A flat vapor chamber with vertical copper tubes have some structural issues that need to be addressed yet are great for very high heat-flux applications. The newer one-piece vapor chamber is probably the truest form of a 3D vertical vapor chamber and offer cost savings over other designs.

## Should I Use Heat Pipes or Vapor Chambers to Cool My Electronics Application

I’m often asked which two-phase cooling technology, heat pipes or vapor chambers, is best suited for a particular electronics application. While there are some good rules of thumb the answer isn’t always simple. Nonetheless, this blog post will cover the basics with additional information covered in ann online webinar found in the Resources section of this site.

Let’s start by saying that a transition from traditional heat sinks to either type of two-phase technologies should only be considered when the design is conduction limited and/or when non-thermal goals such as weight or size can’t be achieved with other materials such as solid aluminum or copper.

Once it’s determined that traditional heat sinks just won’t meet thermal requirements, how do mechanical engineers begin to determine the next best option?

## Are You Moving or Spreading the Heat?

While there’s no hard line of distinction, think of the difference like this.

Space constrained applications and/or those with two or more heat sources sharing a common condenser (fin stack) are good candidates for moving the heat to a remote location of more than around 40mm distance. In these cases, the design flexibility of heat pipes, by virtue of their bendability in all directions, are the default choice. And this is especially true if only a single heat pipe is needed. For reference, a 3mm sintered metal heat pipe using water as the working fluid is adequate for cooling 5-20 watts of power while an 8mm pipe can carry 20-90 watts. The range is affected by the porosity and thickness of the wick structure, the amount of working fluid, the degree of bending required and the amount of flattening done to the round pipe.

In my experience heat pipes are used in 95%+ of applications where heat needs to be moved to a remote condenser. However, when conduction loss needs to be reduced further, such as in the laser diode example below, vapor chambers become a viable option. Here three heat sources share a common condenser, reducing temperature rise by roughly 15% over a similar heat pipe alternative.

450W RGB Laser Diode Solution for a 3D Projector

When it comes to spreading heat to a local condenser, the choice between heat pipes and vapor chambers becomes a lot more complicated….

Heat pipes are probably a best choice if:

• There’s plenty of air flow
• You’ve got lots of room for fins
• Nominal power densities are <25 w/cm2
• Ambient temperatures are normal – let’s call this below 45oC
• Cost is a key consideration – every penny counts!

A vapor chamber should be considered if:

• Power densities are high – certainly by the time they hit 50 w/cm2
• Reducing hot spot across the die is a key concern
• Z-direction (height) is constrained, yet fin area needs to be increased
• Atypical ambient temperatures and/or low air flow
• Performance is a key consideration – every degree counts!

Before I review a couple of examples, it’s important to note some of the reasons vapor chambers can perform better than heat pipes for spreading heat. First, vapor chambers make direct contact with the heat source, whereas heat pipes usually require a solid base plate increasing conduction loss –from the base plate itself and from an additional TIM layer between the base plate and the heat pipes. Second, heat spreading via vapor chambers is multi-directional while with heat pipes it’s linear. Third, a vapor chamber solution often allows for additional fin area. These advantages typically allow vapor chamber solutions to have between 10-30% better performance than their heat pipe counterparts. This translates to around 3-9oC for most applications.

Despite the performance difference, vapor chambers were regarded as a very niche solution due in large part to their cost premium over heat pipes, especially for non-consumer (low volume) applications. But manufacturing innovations from an increasing number of suppliers helped narrow the gap.

Let’s take a look at a couple of examples. The first is for a telecom infrastructure manufacturer that wanted to understand both the cost and performance difference between competing heat pipe and vapor chamber designs. The heat source was in the 85 watt range with ambient temperature of 55 degrees Celsius. We compared a four 8mm heat pipe solution to a single vapor chamber design. As the image below shows, the former required a base plate as secondary machining of the heat pipes was not possible due to the wall thickness of the pipes. Testing showed the vapor chamber design to outperform the alternative by a full 4oC, allowing the assembly to meet specifications.

The second example is from one of the first two graphic card solutions to use vapor chambers. Increasing GPU power required that the current heat pipe based solution be redesigned while maintaining the heat sink stack height. Here two 8mm pipes were replaced with a vapor chamber. Elimination of the base plate allowed us to increase fin height and overall fin area while the better heat spreading of the vapor chamber all contributed to a 6oC better solution.

## Heat Pipe Reliability – What You Know is Wrong!

Last week I gave an IMAPS presentation on heat pipe reliability with Ross Wilcoxon from Rockwell Collins. Thought I’d give you the highlights in graphical form along with some comments.

Product reliability is determined by understanding the reliability metrics for each of the components and or sub-assemblies within the final product. But, does the current thinking about heat pipe reliability make sense?

Neither Ross nor I have never seen or heard about heat pipes failing… for reasons other than stupidity or planned obsolescence.

There’s a lot of conflicting market and research data about the mean time between failure (MTBF) of heat pipes.

Here’s the conventional reliability assurance approach with heat pipes.

Temperature and failure are correlated; chemists and semiconductor folks believe it, but does a much adhered to rule of thumb apply to heat pipes. Namely, every 10 degree Celsius temperature rise will double the failure rate.

In examining past research, it appears that some erroneous assumptions were made.

Ross plotted the findings from a heat pipe failure study and showed that they did follow an Arrhenius curve. Then he estimated the activation energy; a key portion of the Arrhenius equation.

He then used the estimated activation energy of -4507 to determine failure rate predictions at various temperatures. He found that our much believed rule of thumb (double the failure for every 10 deg. C rise) was only true for one point on the graph.

As you can see below, the observations made in the Mochizuki study (that heat pipes follow an Arrhenius curve) should not have lead to the assumption that it followed the half-life rule of thumb. While the blue curve, below, seems to be a better approximation of heat pipe reliability in the field, it’s based on an erroneous assumption.

To determine a more accurate predictive test, we need to find a suitable activation energy. The problem is we’ve yet to get the most common type of heat pipes (water, sintered copper, copper) to fail in the lab.

Celsia is continuing its elevated heat pipe testing until we can determine a new activation energy. In the mean time, companies should remember so long as heat pipes are properly manufactured and tested for the most common early types of failures heat pipes should last in excess of 30 years. Here are some simple guidelines.

Here’s a specific example of how we test at Celsia.

To summarize..