oharag11
Mechanical
- Jun 18, 2015
- 42
Hello. I started using SW Flow Sim Analysis - Electronics recently for PCB thermal studies. I understand that simplification is the name of the game for analysis. In SW Flow you can categorize PCBs using their tools - specifying # of dielectric layers - copper weights - etc... You can add 2-R values for components, and if this isn't available Volumetric heat source. All good. My question is this:
- How does everyone model thermal vias? Is there a good source/training/example for modeling thermal vias for PCB thermal analysis?
So, we add solid features for our vias embedded into the PCB. We assign Solid material parameters to the via holes based upon a spreadsheet we use. We can define via weights and fill.
What I noticed in performing thermal studies between - No Vias vs 1 mil via vs 1 mil different fill that results actually got worse between No vs 1mil. I find this strange. There were maybe 1 or 2 instances where temps were improved. I did notice some improvement between 1mil vs 2 mil vs 3 mil. But compared to NO via - again improvements were worse in some cases.
Questions:
- Is our method of modeling thermal vias (solids embedded in PCB) not the correct way to perform PCB thermal analysis?
- Does anyone here actually perform thermal analysis of the PCBs using the copper masks embedded the multi-layered PCB? I believe this would be more accurate, but more complex analysis. Some vias are not thru but connect multiple layers.
Thanks
- How does everyone model thermal vias? Is there a good source/training/example for modeling thermal vias for PCB thermal analysis?
So, we add solid features for our vias embedded into the PCB. We assign Solid material parameters to the via holes based upon a spreadsheet we use. We can define via weights and fill.
What I noticed in performing thermal studies between - No Vias vs 1 mil via vs 1 mil different fill that results actually got worse between No vs 1mil. I find this strange. There were maybe 1 or 2 instances where temps were improved. I did notice some improvement between 1mil vs 2 mil vs 3 mil. But compared to NO via - again improvements were worse in some cases.
Questions:
- Is our method of modeling thermal vias (solids embedded in PCB) not the correct way to perform PCB thermal analysis?
- Does anyone here actually perform thermal analysis of the PCBs using the copper masks embedded the multi-layered PCB? I believe this would be more accurate, but more complex analysis. Some vias are not thru but connect multiple layers.
Thanks