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

 

[GregLocock]

so

what is the thickness of the adhesive?

what is its E?

what are the beam properties of the heatsink?







Cheers

Greg Locock

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

 

[GregLocock]

Oh and...

what is your best guess approach?

which university did you go to?



Cheers

Greg Locock

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

 

[Chauvet]

Greg, Thanks for the response, first the the historicals, UT, Molecular Biology and Genetics, Yea , I know, what the heck am I doing here. Materials and chemistry are an easy extension, I have been in Semiconductor Packaging and doing very well for 20+ yrs. We are a start up and I like others wear several hats, My genius or ignorance is the ability to ask a question and not afriad to admit that someone can answer it quicker than I. I will find an answer, however the forum has a level of competance not often found and I hope to utilize and ACKNOWLEDGE the source. So, again, any directions, or formulas are appreciated. OK......an Al heatsink size, previously noted, attached with two (design flaw, not my design) to a larger heat sink, also Al. The T805 is Thermagon T805 @.005 thickness and a 5w/mK TC. Beam proprerties, I dont have, however I realize that, as a function of Hook's law and th modulus of elasticity is a factor. (I have been in management way to long and lost the math edge) Dont fault me, just help. What else can I offer?

Chauvet

 

[Chauvet]

Greg (automotive) for whom and where? I am ex-Delco, Kokomo, advanced product development.

Regards

 

[Chauvet]

Addtional Data, I contrived a method to 'visually' show the contact pressure (impression) or lack thereof. I had a 250M surface finish put on the HS, as I increase the torque on the screw, you can see the inprint of the finish on the T805, this works well. I have tested to 2x recommended torques and have detemined (.125 thickness, .187 and .250 HS thickness) that I need to use a .250 thick AL HS at 9 in lbs to get an "impression" across the entire surface of the T805. Remember the screws are placed near the edge of the HS. OK that's nice, but I need to guarantee that I have, at a minimum, 100 psi in the middle of the HS on the T805 to assure the 5w/mK or conversely the minimal thermal resistance. It is logically obvious that we need to put two more screws in the HS to get an even distribution, addtionally, until I get a strain gauge to measure the pressure I would still like to calculate the pressure or have a resonable assessment of the pressure. OH, ONE IMPORTANT NOTE, the company is made up of RF ELECTIRCAL ENGINEERS, so tag I'm it for the mechanical, they have their ideas and you would not believe some of them.

 

[unclesyd]

Here is a product we used on very similar problems.

http://www.fujifilm.com/products/prescale/index.html

 

[prost]

That Fujifilm stuff works great. Little spheres of color will explode as you press down harder on it--the harder you press, the more spheres that explode, the more color you get. However, because the film is squishing in between two really stiff objects, the contact pressure distribution with the film in between is not the same as the contact pressure distribution WITHOUT the film (the peaks are lower with the film, also the contact area is bigger with the film sandwiched in between).

But you have to know how to calibrate the film (don't look at me! We got ours done at National Research Council-Canada). As I suppose it would be difficult to calibrate the rather ingenious method (0.25 thick Al) you came up with--you can see there is an impression made on the Al, but what is the pressure level? You can only guess.

I don't understand where the pressure is coming from. Is this magical 100 PSI that you need the pressure you see in the middle between the two objects you've screwed together (the HS and whatever the base is)? If so, what the pressure distribution looks like, the shape of it, depends almost entirely on the shape of the two objects being pressed together. Say your two surfaces are perfectly flat, I would expect the pressure distribution to be flat, or constant. What if one is perfectly flat, and the other is a little bit curved, AWAY from the flat surface, so the two bodies are touching at the edges, but not in the center (the opposite of a ball being pressed onto a flat plate, that is). Then you could have ZERO pressure in the middle if the two objects aren't in contact in the middle.

 

[CoryPad]

If you need to prove that you are developing a certain pressure, then you can use a contact pressure measurement system like the Prescale film from Fuji, or a more capable dynamic system like one from Tekscan:

http://www.tekscan.com/index.html

Yet another physical method is using ultrasound - here is a technical journal article on this method:

http://links.jstor.org/sici?sici=1364-5021(19961208)452:1955<2613:ASOTIB>2.0.CO;2-Y

You also could use computer software to calculate the answer. For example, Finite Element Method with contact elements. This is probably your best bet.

If you want a standard analysis method with equations, I think you will be disappointed. The fastener clamping force creates a ring of pressure around the fastener hole, but the magnitude is a function of so many variables that it isn't modeled with basic equations. Your instincts are leading you in the right direction - relying on two screws at the outer edges isn't going to be a good guarantee of high clamping force in the middle.

Regards,

Cory

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

 

[GregLocock]

Sorry for the hardass attitude, it looked like a dopey engineering student question, not a "help I'm out of my depth" one.

I like the idea of the pressure film. I'd probably use FEA.

So far as equations go, what you have approximately is a simply supported beam on an elastic foundation.

The good news is that that is a solvable case, for uniform beams.

The bad news is that there is no simple general analytical solution for beams of varying width, which is what you have.

If you measure the contact pressure then you could use the standard beam equation to work out how much to pre-bend the heat sink before installing it.






Cheers

Greg Locock

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

 

[CoryPad]

I forgot to include one "rule of thumb" calculation. You can graphically determine clamping zones by using a sectional view of the parts. Extend lines at 45[&deg;] away from both edges of both screws into the clamped part. You can assume essentially zero pressure in the region between the clamping zones. The attached image should be helpful.

 

[TheTick]

Any reason Pressure=Force/Area won't work for this?

 

[Chauvet]

Gents
Sincerely appreicate the responses. Bare with me as I respond to each. TheTick, yes it would if I had a uniform pressure over the entire surface, I need 100 psi, I have 1.6 sq in therefore I would need 160 lbs of force, however I have only the two screws to apply force. CoryPad, thank you, may I ask the source?. The rule appears to be the same, and is resonable, as with Thermal distribution in a fixed ambient temperature. I will use your pictorial. It is the math that I am out of the league with. Will investigate the journal. GregL, yup, this is a simple real world problem, I will investigate your recommendations and I totally understand the pre-stressed beam example, however that is too elegant a solution for this bunch. Prost and UncleSyd, familiar with the Fuji material, trying to get some, however time is my nemesis. Prost, the object is a rectangular Al HS, .86 x 1.86 x .125 (original design) the graphite film is a TC film that transfers heat, the HS has a 32 RMS finish with Ni plate and is .0004/in across the HS. The 100psi is specified by the Thermagon folks to achieve the 5w/mK (Thermal conductivity of the graphite film). The issue is that any pressure less than the 100 psi, the thermal resistance goes up (inverse of thermal conductivity). The T805 material is compressible under the screws so the HS bows and there is no contact in the middle, (many flatness experiments)between the AL and T805, and air is a poor thermal transfer medium. I progressed up to .250 thick Al and can get a contact impression (see previous notes)across the entire film surface, Although this does not guarantee the, all important, pressure in the middle of the HS, therefore, Strain gauges and Fuji film are my next shots. My further recommendation will be to put two screws nearer the center, total of 4 screws across the length of the HS, the complication is there is a 50W power amplifier mounted in the middle of the HS. Will send pics of the WHOLE problem as time permits tomorrow. I genuiely appreicate all of the responses. Anu more thoughts are appreicated.
Regards
Chauvet (not a handle, the real name)

 

[MikeHalloran]

That's a tiny and exceptionally thin heatsink for a 50W _anything_. It's also an exceptionally thin object on which to 'mount' anything.

If space limitations dictate that it can't be thicker, perhaps a dab of thermal grease and no spacer would do beter?



Mike Halloran
Pembroke Pines, FL, USA

 

[Chauvet]

Mike H, I cannot disagree, this was the previous specified dimension, will probably go to .250 and at least 2x area, this will complicate the contact pressure further. This heat sink is then bolted to a larger BASE Station HEAT SINK.
That is what the two screws were for, however we will have to add screws with increased area, the problem still lies wit the circuit board layout on top of the heat sink and room for the addtional screws.........

 

[MikeHalloran]

I think you've got enough screws already; you just need the heat sink to be a bit stiffer in bending.



Mike Halloran
Pembroke Pines, FL, USA

 

[CoryPad]

I created the graphic myself. It is based on several information sources, although the best one for this is Chapter 6 Designing Fastening Systems written by Christoph Friedrich from the book Modeling and Simulation for Material Selection and Mechanical Design published by CRC Press.

Regards,

Cory

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

 

[Chauvet]

Thanks Cory, I will get a copy.

Regards

CP

 

[GregLocock]

Have you considered bending the heatshrink to even out the pressure, or profiling it or adding a shim in the centre?



Cheers

Greg Locock

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

 

[sreid]

This has been a problem forever with power semiconductor packages. The pressure distribution under a TO-3 package was terrible. In modern TO-227 packages, the bottom has a slight bow so that when the outboard screws are snugged up, there is pressure in the center of the package (which is where the die is mounted).

I'm with Greg Locock, machine a profile.