A Practical Guide to Using Two-Phase Heat Sinks

Click here to download the presentation.

Overview: Two-phase devices are incredible heat conductors and significantly boost heat sink performance. We’ll present a practical guide to using both heat pipes and vapor chambers: similarities, differences, operating parameters, mounting options, common mistakes, best uses, examples, and performance modeling.

George MeyerGeorge Meyer
George Meyer is a thermal industry veteran with over three decades experience in electronics thermal management. He currently serves as CEO/CTO of Celsia Technologies. Mr. Meyer has been instrumental in establishing Asian operations, developing new technologies, key customer relationships, managing the product portfolio, and growing sales into the computer, telecommunications, networking, LED lighting, medical, laser and military markets. He holds over 70 patents in heat sinks and heat pipe technologies.

The following are questions presented to the speaker by the attendees during the webinar, along with answers to each.

I do understand that heat pipes transport heat at a phenomenal rate which leads me to a question. I am in the LED industry and I am here in China designing some new compact high power luminaries. I am quite new to the business and may be making some wrong assumptions but some of my simulations indicate that heat pipes make no significant differences to the final SS temps on the test point on the LED COB chip! Why is this? Is it possible that the heat sink design is very efficient and cannot be improved upon very much. Results of simulation show following WITHOUT heat pipes:

Temp on LED chip surface – 78 C
Temp on Base of Chip – 74 C
Max temp on heat sink – 67 C
Min temp on heat sink – 62 C

This gives a max temp differential of 78-62 = 16 C

Heat sink is extruded aluminium about 90 mm dia and 200mm long.

I was trying to lower this by using heat pipes but simulation shows only about 1 C difference lower at SS condition! Am I doing something wrong?
Answer: I think we had this discussion previously. In natural convection applications the conduction losses are quite smaller than in forced convection applications. I believe your results may be correct. If you have more information such as a sketch of the design I can give you further advice.

I had never heard of one piece vapor chambers until this presentation. Are they proprietary to Celsia?
Answer: While a growing number of manufactures offer the one piece design to reduce the cost vs a two-piece version, Celsia was the first to commercialize them in 2006 and still holds a number of patents pertaining to their design.

For a 3mm pipe – vapor limit is higher for a flat heat pipe vs. a round heat pipe? So flat is better in this case?
Answer: No, that is just an issue with the chart. Normally 3mm pipes aren’t flattened. If they are, the vapor limit would be lower than the round one.

Celsia also makes 2mm diameter heat pipes, correct?
Answer: We can make all kinds of shapes and sizes including 2mm if there is a requirement. We’ve also mass produced vapor chambers as thin as 1.5mm.

Are the excel model details proprietary? Can we get a look at it?
Answer: Yes, the details of the excel model are proprietary, but we do share it with our customers.

Please elaborate on the calculations used to calculate the heat pipe thermal conductivity.
Answer: We calculate the thermal resistance of the heat pipe or vapor chamber to get the delta-t in the device. We then, using the cross section and length of the device to calculate the effective conductivity numbers.

How do you model evaporator and condenser resistances for your excel model? And for your CFD model?
Answer: We do this based on values from years of test data.

What method does generally CFD use to model phase change?
Answer: Modeling the phase change is not necessary in a cfd model. Once you have the estimated thermal conductivity you can plug this number into your CFD model.

How reliable are empirical correlations in spreadsheet simulations?
Answer: It is typical to get within 10% of the final test numbers.

How do you mount heat pipe assembly and ensure the processor or ASIC solder joints are not impacted by the additional load?
Answer: The vendor should recommend a maximum load for the ASIC and then design the mounting hardware to the specification. Care must be taken with all thermal solution when mounting to apply the pressure evenly.

If VC is ranged for 50-150W does it mean that it will not work (start up) with 15W?
Answer: No, these are the typical maximum values based on against gravity capacity (50w) and the gravity aided position (150w).

How can I make material & wick structure selection if the heat pipe is under jet engine vibration environment?
Answer: Two types of wicks have been used for these applications, groove and sintered powder. Groove holds up well to shock and vibration because they are part of the wall. While sintered wicks are diffusion bonded to the wall. Mesh wicks would have a tendency to move due to the vibration unless it is also bonded to the wall.

What is a good reference, i.e. book or article available about two phase devices?
Answer: Dunn and Reay is the standard book.

What type of Thermal Interface Material is typically used between heatpipe/vapor-chamber and device? What pressure is required for good thermal contact with vapor chamber?
Answer: The most common are grease and phase change coated graphite pads. A good rule of thumb for interface pressures is greater than 20psi. The server guys with higher power densities use up to 50 or 60 psi while the laptop designers struggle to get more than 5 or 10 psi.

How do VCHS deal with multiple heat sources?
Answer: Very well! We would model it with the full load as one source then get real test data.

Bending reduces the Qmax of heat pipes. Does it also reduce their effective conductivity, or only heat carrying capacity?
Answer: It will affect both. We have a presentation on the effects of bending that we can share.

Is there any min. or max. diameter limit for any heat pipe?
Answer: In theory, no. Large ones up to feet in diameter have been built and micro pipes have been around for some time.

Is there any min. or max. length limit for heat pipes?
Answer: There are particle limits. For short distances copper moves heat pretty good and we have built pipes up to 60 feet.

Is there any min. or max. diameter limit for any thermosiphon?
Answer: Since these are closely related to heat pipes the answer above applies.

Is there any min. or max. length limit for any thermosiphon?
Answer: Since these are closely related to heat pipes the answer above applies.

Which is more effective – ‘4 heat pipes’ or similar size ‘closed loop pulsating heat pipe with 4 pipes’ placed between same evaporator and condenser?
Answer: Properly designed I would order these 1) vapor chamber, 2) heat pipe and 3) pulsating heat pipe.

By how much will delta-t be improved by using vapor chambers with direct contact to heat source versus heat pipes with a mounting plate?
Answer: Direct contact will improve performance by between 3-5 degrees Celsius.

I would like to know the allowable clamping pressure or range based on size of the vapor chamber.
Answer: We typically allow for up to 60 psi clamping pressures, although some designs can support 90 psi.