Heatsink Question

Thread Starter

Ddunham5

Joined Nov 9, 2018
6
Ok so I have designed a circuit and drawn up its PCB and I have a slight question about the heatsink I'm going to use.
Now you'll see in the images below the place where I generally want to place my heatsink. (The yellowish green area.)
BACKPLANE MASK VIA OVERLAY.jpg
My questions are:
A) Is it okay to utilize a monolithic heatsink to service multiple parts (provided you use proper electrical isolation.)
B) You will see in the second photo a highlighted area a trace falling under where the heatsink will be placed. Will this be a problem (Again assume I will be using some sort of electrical isolation.)
BACKPLANE MASK VIA OVERLAY HIGHLIGHT.jpg
I already have done calculations based off of thermal resistance to to test my idea I'm just not 100% if it will work. The last image is an example of my basic modeling. I am basing it off of Dave Jones' thermal resistance tutorial videos. I know he only ever did it for one part to a heatsink, but since he said the system could be modeled like an electrical circuit (Replacing current for Power and Voltage for temps) I figured this would work. The node where every thing meets is the Sil-pad or thermal compound touching all of the components before it hits the heatsink.
BASIC THERMAL MODELING.PNG
 

crutschow

Joined Mar 14, 2008
27,477
A) Is it okay to utilize a monolithic heatsink to service multiple parts (provided you use proper electrical isolation.)
Yes.
B) You will see in the second photo a highlighted area a trace falling under where the heatsink will be placed. Will this be a problem (Again assume I will be using some sort of electrical isolation.)
I see no problem with that.
since he said the system could be modeled like an electrical circuit (Replacing current for Power and Voltage for temps) I figured this would work.
It should, but can you explain what the various resistors and the voltage source represent?
If those voltages are temperatures, than some of your components appear to be overheating.
 

Thread Starter

Ddunham5

Joined Nov 9, 2018
6
The resistors are supposed to be the thermal resistance at the junction. Power I assumed 10W above normal operation (No better kill than over kill) and each of the parts maximum are rated for at least 20 degrees higher than what you see shown. Voltage source is ambient temp.
 

andrewmm

Joined Feb 25, 2011
1,645
Easy first approximation.
Find the degree C per watt of the heat sink, then find the watts your using on the board.
That gives you the uniform temperature of the heat sink,
from which you can work out the junction temperature for each part.

A big spread sheet I find is the best if you do not have real simulation software.
 

Thread Starter

Ddunham5

Joined Nov 9, 2018
6
@crutschow you do have a point there I'm actually starting to wonder if separate heatsinks would be better.
@andrewmm the way your recommending isn't that essentially what I'm doing by uniting all of those parts at the same node?
 

andrewmm

Joined Feb 25, 2011
1,645
@dbunham5,
Yes, using LT spice to solve thermal questions is an interesting method, I have not seen before,
I have always used a spread sheet is all I was saying.

The other thing to take from this is that thermal is very variable.
so its IMHO designed with big assumptions and big spare capacity.
 
Top