7 Replies Latest reply on Dec 21, 2010 9:33 AM by Joe Galliera

    How to simulate gap pads

    Erik Steen

      Hi,

       

      I am facing some difficulties regarding simulation of thermal gap pads. In my model I have a CPU cooled towards an aluminium cooling profile. Between the CPU and the cooling profile I have a 1,5 mm thick gap pad compressed to 1 mm. The gap pad is made up of silicon and rubber, and the data sheet gives only thermal conductivity and density of the material but does not address specific heat etc. which Flow Simulation material definition requires.

       

      Here is the material data sheet link:

       

      http://www.bergquistcompany.com/pdfs/dataSheets/PDS_GP_1500_12.08_E.pdf

       

      I had a look at a similar problem in another tread where it was suggested to suppress the pad, rearrange the assembly and use contact resistance instead. The pad is fairly thick in my model (1 mm), so this seems like an wrong approach?

       

      What is the correct way to model/simulate this scenario? According to the help feature in Solidworks it seem like Floworks uses only the thermal conductivity and the thickness of the pad to simulate. If that is the case, how come I have to enter the specific heat in the material definition?

       

      All suggestions are highly appreciated!

       

       

      Best regards

      Erik

        • Re: How to simulate gap pads
          Rich Bayless

          Hello Erik,

           

          Why does Flow Sim need specific heat?  It uses a transient approach to steady state, so having a specific heat makes sense for the solution method.  If you don't care much, you could just pick a number.

           

          Or,

           

          The reference data sheet shows 'Heat Capacity' with units of J/g-K.

           

          From: http://en.wikipedia.org/wiki/Specific_heat_capacity

           

          specific heat has units of J/g-K, so the data sheet's label is misleading, they are actually listing specific heat.

           

          Rich.

            • Re: How to simulate gap pads
              Joe Galliera

              It does not solve as a transient approach to steady state.  Yes, you need to define Specific Heat for a transient problem because it defines how much energy the material can store.

               

              What the issue is here is that Flow Sim requires that you fully define a material and not leave a value blank, in case at some point you do try to use it in a transient problem.  For your case if you are only ever use it for a steady problem (make a note of it for yourself in the comments), then just enter any number.

                • Re: How to simulate gap pads
                  Rich Bayless

                  Hello Joe,

                   

                  In regards to your comment:  "It does not solve as a transient approach to steady state."

                   

                  Here's what the Flow Sim 2010 Online User's Guide states:

                  Solving

                  Flow Simulation discretizes the time-dependent Navier-Stokes equations and solves them on the computational mesh.

                   

                  Since Flow Simulation solves steady-state problems by solving the time-dependent equations, Flow Simulation has to decide when a steady-state solution is obtained (i.e. the solution converges), so that the calculation can be stopped.

                   

                  I interpret this as 'a transient approach'.  It starts with the flow stopped, and slowly speeds up the flow until it reaches steady state.  Is there a better way of describing the solution method, to beginners?

                   

                  Rich.

                    • Re: How to simulate gap pads
                      Joe Galliera

                      The method for a s-s solution is definitely analogous to a time-dependent approach so it appears time-like.  Say you are running a steady external conjugate heat transfer problem, you will see the parts heating up until they reach a stable equilibrium temperature but there is no time element to this. In fact when running a steady free convection problem, I tend to apply an initial temperature (a higher temp than ambient) to the parts that will be heating up so that the flow kicks off in the first step because it is driven by gravity and temperature differentials.

                       

                      Just think of it as an interative approach where the result is updated (n+1) from the previous step (n), and not as time (t) to updated time (t+dt), where there are more concerns with the stability of the solution for the selected time step.  This is why it is advisable not to mess with the automatic time step (too much).

                       

                      I was not trying to correct you from what you wrote, but it could be interpreted in the wrong way.  I just didn't want to give the impression out there that we solve all problems as transient and then when they stop changing that yields the steady result.  That would be a very inefficient approach to solving a steady solution and solution times would be much longer than what we can do now.

                • Re: How to simulate gap pads
                  Bill McEachern

                  the heat capacity is given right on the data sheet at 1000 J/kg*k. Put in the conductivity at the compression level sought. Not too sure what the issue is? You could also model it as a contact resistance and take the model of it out all together.

                  • Re: How to simulate gap pads
                    Erik Steen

                    Thanks! That solved my problem with the gap pad 1500.

                     

                    I also have two insulator materials in my model;

                     

                    http://www.bergquistcompany.com/pdfs/dataSheets/PDS_HF_300P_12.08_E.pdf

                     

                    http://www.bergquistcompany.com/pdfs/dataSheets/PDS_SP_2000_0410.pdf

                     

                     

                    If I understand you right there are three ways to simulate these insulator materials:

                     

                    1. If I'm only interested in the steady state temperature, I only have to enter the correct values for thermal conductivity and thickness. I can just "pick a number" for the other values if they are unknown.

                     

                    2. I can suppress the insulator, rearrange the assembly and add contact resistance between hot component and cooling fin.

                     

                    3. If I'm interested in time dependent temperatures I have to look up the density, specific heat and thermal conductivity in the datasheet/matweb.com and add all the values

                     

                     

                     

                    I'm quite new to this, so I really appreciate your help :-)

                      • Re: How to simulate gap pads
                        Joe Galliera

                        Please note that thermal contact resistance varies by the amount of pressure applied.

                         

                        The new Electronics module for Flow Sim 2011 includes a library of thermal interface materials from Bergquist, Dow Corning, Thermagon and Chomerics.  They include variations for pressure applied.