EML3005 - Supplemental Lecture 1 - Thermal Management: Heat Sink Design I

This is the first installment in a short supplemental lecture series discussing thermal management topics in mechanical engineering for the UF Senior Design course (EML3005). This lecture gives an introduction to forced convection heat sink design for microprocessors, focusing on optimizing fin spacing. Familiarity with basic fluid mechanics and heat transfer concepts is recommended.

[@mcinman87]

Hi Kunal, In this presentation it is assumed that: (1) the fins are at uniform temperature and... (2) the fins are very thin with respect to the channel length (L>>t) These 2 conditions allow for fin thickness to factor out of the final steady state heat transfer equation entirely (see final 3 slides). Lecture 2 covers the same analysis, only this time the fin temperatures vary along the height. This requires the solution to a second order homogeneous ODE with homogeneous boundary conditions. Once solved the maximum fin efficiency can be isolated and you can develop relationships for the optimum values of "t" and "b", as well as the other parameters of note. Hopefully Lecture 2 answers your question fully. Best, Matt

[@kunallamba698]

ohhh thanks a lot for this video I like this video. :) one more question while calculating b(optimum height)in numerical,you put the value of w=5cm, why you take value of w in centimeter as all the other parameter given in meter. sorry for this silly question but I am weak in units. please explain me ASAP.

[@kunallamba698]

you have not reply me yet :(

[@mcinman87]

I chose to write 5 cm instead of 0.05 m because it is a more intuitive and easy-to-visualize length. It is just personal preference, all of the arithmetic is correct and the units all balance out. As for the delayed response, I've been a little sleep deprived and suffered from some baby-induced tunnel vision lately. If you check out my profile pic you'll see I'm a new dad, and the last 2 months have been pretty amazingly busy. I will reply as promptly as I can, but unfortunately exploding diapers and crying babies come first at the moment.

[@sudheernorway]

#16:09 --- I can see z3 in the equation for Q Where as at #22:27 ---- I can see z2 in the equation for Q1 Please clarify how it become z3--->z2

[@sudheernorway]

Can you please make a video on designing a common heat sink (forced cooling) for 4 IGBT modules?

[@Fun_Dealer]

Hello, this presentation is awesome! Is there anywhere I can download it to make my notes directly onto it?

[@mcinman87]

This link leads to a PDF providing the abbreviated analysis for determining the optimum fin spacing for an isothermal, linear array of rectangular fins via Natural Convection. It is recommended that you watch lecture 1 first, as the analysis here is very similar. Good luck. drive.google.com/file/d/0B2sp53zkUVeUNG9CQzl2amthZ2s/view?usp=sharing

[@kunallamba698]

hi matthew imman, In this lecture you optimize z(spacing) and b(height) by assuming (w). I want to know how you are optimizing t(thickness)? reply as soon as possible.

[@ankitkumarparsurampuria7682]

Hi Matthew, must say, this was a good and thorough lecture. Could you also share theory/lecture on 'natural convection' in heat sinks and how to oprimize fin spacing for natural convection?

[@mcinman87]

Ankit Kumar Parsurampuria Hi Ankit, Sorry for the late reply. My family and I took a mini-vacation over the long Memorial Day weekend and I was away from my computer and old texts. I recorded this lecture as a graduate student at UF and have since graduated and moved away, so I no longer have access to the recording studio. I can howeverpost a PDF outlining the procedure, and I will begin working on that today. I will send you another reply once I have posted it. The outline will follow the analysis found in "Thermal Design: Heat Sinks, Thermoelectrics, Heat Pipes, Compact Heat Exchangers, and Solar Cells" by HoSung Lee (Wiley & Sons, 2010), and I would highly recommend this text if you have interest in the subject. E-copies are available for free to current students of many Universities through the library system.

[@mcinman87]

Ankit Kumar Parsurampuria Ijust posted the link. I hope you find this helpful. drive.google.com/file/d/0B2sp53zkUVeUNG9CQzl2amthZ2s/view?usp=sharing

[@ankitkumarparsurampuria7682]

Matthew Inman Thank you for the prompt response. I am surprised and at the same time happy to see that you could spare your time to summarize the theory and upload it.

[@marwanzaid9764]

i honestly had no idea how you got the Zopt= 4.6mm. could you tell me what are your values that you used. That's what i got -Dynamic viscosity 1.8616*10^-5 (kg/ms) -thermal coff of air under T=25C, K= 0.026(w/m.K) -density p=1.184 (kg/m^3) -cp= 1.00(KJ/Kg.K) - to find the thermal diffusivity "alpha"= K/density*cp (m^2/s) -length L=0.05m i used your Zopt equation but i keep in getting the answer = 0.8668mm where is my mistake! I've been going back and forth into derivatives and constants and i still couldn't get your value (Z=4.6mm!!) thanks for your help

[@hanyuzhi9864]

the calculation of the thermal diffusivity "alpha" is OK?

[@ankitkumarparsurampuria7682]

Hi Matthew, Using the theory in this video I have been trying to design a heat sink for an application. There seems to be a flaw in the approach to reach the optimum value. Towards the end of the video you use the fan’s efficiency curve to select a certain pressure drop (delta_P) and use it to calculate z_opt and fin height. Now suppose the required chip temperature changes - then at the same pressure drop you will get a new value of b (fin height). But since the fin geometry has changed (due to change in fin height) the fan will NOT be operating at the same delta_P which conflicts with the initial value of delta_P I have explained this here in more detail - drive.google.com/file/d/0B8-la1FomNr3ejZfSHJqbjQzbkU/view?usp=sharing My question is does operating the fan at its max efficiency ensure minimum heat sink temperature (or maximum heat dissipation)? If not, then what is the significance of q_max?

[@mcinman87]

+Ankit Kumar Parsurampuria Hi Ankit, To answer your question at the end first, the max fan efficiency point does not relate directly to optimum heat dissipation or min. fin temperatures. It is a design point for the fan only, and relates the power transferred to the flow to the power used to turn the blades. We are using that design point to minimize the parasitic load on the overall system and that is all. Then, as you note in your comment, that design point corresponds to a specific delta_p, and that informs the rest of the analysis. In terms of the required chip temperature changing, I'm having trouble envisioning a scenario where this would happen. The chip temperature is taken directly from the manufacturer's guidelines, so that should remain constant. That said, if the ambient temperature changes, then yes, you will need to go through the whole analysis again. Ambient conditions are assumed, and changes in your assumptions in a design scheme usually mean you need to go back to square 1. This is fairly simplified analysis, and like you have shown, is not very robust when considering changes to the design points. But in a controlled environment (say for a personal computer housed in a air conditioned room at sea level) it should be sufficient. I hope this has answered your questions. Best of luck. Matt

[@chandraprakash934]

Dear Sir I am student in India. How can I have access to HoSung Lee ebook ?

[@mcinman87]

Below I have provided a link to a PDF version of the slides with all of the hand written portions typed in for clarity. drive.google.com/file/d/0B2sp53zkUVeUa21QRDhfcWhmUXc/view?usp=sharing