Calculating PSI from Axial Force 2

[Chauvet]

We have a heatsink .86 x 1.86 x .125, Al 6061, We are using 440 cap head screws to attach. The screws are .200 from each end, only two. The thermal conductive material is T805 and requires 100 PSI to achieve its thermal conductivity value. Thatcalculates to 160 lbs of force over the 1.6 sq in surface. I can caluclate the axial force of the screws, however I need to determine the PSI distribution over the entire interface, any questions directions or suggestions is appreciated.

Chauvet

 

[Chauvet]

Very familiar with the TO packages and a world of ceramic and plastics both with no prestress and those with prestressed bottoms, some not intentionally. We have a TO-272 Freescale 30W part soldered to the top of this heatsink. Solder has a 48-50w/mK TC. Then we bolt the HS down to an Al housing, the base station housing. Yes, there is a PCB on top of this HS. One of the questions here is the "cumulative" pressure of screws holding something together. I have read in another forum where someone asked if he had 6 screws with axial force X at each one is the force across the surface "cumulative", someone told him yes. That seems incomplete. ie, if I use 4 440 screws torqued to 5in lbs equally placed across a 1.6" rectangle then is the force 223 lbs x 4 T=.2DxF or 5 = .2*.112*223.
Any comments on yes this is true or no and here's why is apprecaited, It just seems that the spacing between the screws is a factor.

 

[CoryPad]

Contact pressure can be cumulative if the screw spacing is small enough for their clamping zones to overlap and interact. There is plenty of data on this in technical journal articles. You would need to use keywords like "contact pressure".

Regards,

Cory

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[sreid]

Since the T805 is electrictly conductive, what's the advantage over bolting Al to Al?

 

[Albigger]

I'm not in the electrical design arena, but I'm pretty sure this is one of the reasons that some heatspreaders on intel processors were slightly convex shaped.

I think if you have enough screws to supply the required force, then your biggest problems will be, as stated previously, bending of the heatsink and manufacturing tolerances on the components.

 

[Chauvet]

SReid, Yes the TC805 is both electircally and thermally conductive. AL to AL is a poor thermal transfer depending on the surface finish, RMS.Air "voids" between the two surfaces are the insulators. The TC805, or grease, fill in those void. Grease is excellent, however it has a dry out factor with heat and time and a Pump out factor, were it moves out of the way and doesn't come back, there are materials called phase change materials that prevent this.
AlBigger, you are correct, many HS manufacturers prestress or dimple the contact area for improved contact. The screw problem is the pressing one for me. The boys that designed this module dont want to change the design and add screws, they want me to tell them what the pressure is in the center of the heat sink with only two outer screws. They assume that we can get there with more torque and thicker matl, that is correct however you have 500 lbs of force on the ends and 150 in the center and the long term reliability of the joints come into question. I found the Handbook of Bolts and Bolted Joints and some software that is suppose to actually solve the very porblem I have, those with the Fuji Film that is suppose to arrive today, sould get me there.....will keep a light in the window until the problem is solved.

C

 

[sreid]

I guess I should have asked what the advantage is of the "Flexible Graphite" over phase change wax between the aluminum pieces. With wax the question of pressure distribution goes away.

 

[MikeHalloran]

Hold on a sec. You've got a package soldered to the back of the HS. Which means that the package and the solder are also stressed as the HS flexes.

Which means it's not a simple beam on elastic foundation problem anymore. It's a composite beam of nonuniform section, and you may not know enough of the package details to analyze it correctly.

My head is starting to hurt.

At this point, I'd be reading Freescale app notes and spec sheets, and doing exactly as they specify.



Mike Halloran
Pembroke Pines, FL, USA

 

[GregLocock]

Ignoring that (it is a good point) here is some messing about with FEA

http://members.optusnet.com.au/greglocock/heatsink/heatsink.html

Until you know the stiffness of the thermal paste all bets are off.

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[Chauvet]

Good Morning, back from a few days of details and cedar fever.....central Texas allergies are the worst.
SREID: The phase change materials are superior for gap filling, they do not have as good a TC, however due to the very thin (.003)and gap filling abilities they have outstanding LOW thermal resistivity. Unfortuately I am told to "stick" with T805 for rework and ease of cleanup.
Mike: WELCOME TO MY WORLD...............I am ex Moto and I have the APP notes pasted to the cubicle wall. They note two 440 screws @ 5in lb offer 450 lbs of force. The device is .2 sqin in size. They use T805 undere the device, however we solder it down. Yes, not only is there a bending moment with the package, then there is a PCB around it and the PCB is soldered to the HS.
GregL: thanks for the info...... (The pictures do not come through, will attempt to contact you via your home page.) They are using Thermal Paste in the lab, everyone is a thermal genius in there, So I let them do what they want, I am focused on the shippable product.
I received a sample of the Prescale FugiFilm, interesting stuff. It works, however like statistics you can make it do what you want, to an extent. The imprints simply do not show the pressure we expected in the center of the HS. I used the sample without the T805 between the two HS blocks, the argument here which is correct? I do get two different, however small,impressions if I use the T805 material or not. Any experience out there on the Fugi material?

 

[GregLocock]

OK, I think I know what is wrong. Try the link again.

http://members.optusnet.com.au/greglocock/heatsink/heatsink.html

Cheers

Greg Locock

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.