[cfd_user_pune]

Rep Power: 4
cfd_user_pune is on a distinguished road
Dear Thermal Designers,

I am trying to validate CFD results of Junction temperature using theoretical calculations.

I have simple PCB of thickness 1.5 mm with K(in plane) and K(through) as 35 and 0.9 W/mK resp. In middle of PCB I have single IC component with power load of 0.88W. Rjb= 2 C/W, Rjc = 1C/W. System is supposed to have natural cooling without any heat sink so whatever cooling is achieved is through PCB board.
h = 0.955*(delT/L) ^0.25

I am modeling PCB heat transfer assuming it as circular disc fin with r1 and r2 as inner and outer radius. Temperature profile can be represented by 2nd order differential equation having complex number. Please refer
https://www.google.com/url?sa=t&sour...kCmH6JJq0dENJh



https://www.google.com/url?sa=t&sour...=1614486348298

Solving it requires bessel calculator (METHOD 1)and which is available online.

Another way (METHOD 2) I am doing heat balance by assuming (heat conduction through disc fin) = (heat connected + heat radiated through disc fin) = heat load of IC

Unknown in method 1 is outer radius of disc fin r2 or effective area of cooling.

In second method unknowns are r2 and temperature T(r2). Here I am assuming that conduction and (convection+radiation) are in series.
To evaluate r2 I am assuming T(r2) = 1.6 *T(ambient) based on judgment. Using goal seek I'm funding to and T(r1).

Both methods give fair enough results which matches with CFD Tj value. But for small IC size only (like < 10x10mm)

Personally I feel bottle neck is to find the effective cooling area for conduction convection.

Is anybody aware of how to decide effective cooling area nearby IC? Is there any thumb rule?

Thank you in advance.
cfd_user_pune is online now Add to cfd_user_pune's Reputation Report Post