Help with PC board design

[trident9]

Greetings
Have been reading everything I can find concerning SMT thermal pads and discovered that R(th)JA is difficult to predict because it consists of several parallel paths whose impedance varies as the heat flux changes. I did discover a few rules of 'Thumb' in the process:
1) The lowest R(th)JA is provided by a continuous pad centered around the component, on the same side as the component.
2) Increasing the thickness of the copper pad improves R(th)JA significantly.
3) Pads larger than 1 sq inch do not significantly improve R(th)JA.
4) Multi-layer boards improve R(th)JA.
I have reworked the board with a Dpak MOSFET and moved it to the center line of the board. The positive rail routing in Boncuk's layout gives shorter traces and I've incorporated it as well. Thank you Boncuk
_______
trident

[Boncuk]

Did you read to put thermal drills underneath the transistor you want to cool?

From my understanding thermal drills take care of air movement, hence improving cooling. If they are blocked there won't be any air movement.

Using double layer copper with plated thermal drills both the component side and the solder side serve as cooling pad.

On the other hand you could use a single sided board with 2oz of copper weight as suggested by the manufacturer.

Boncuk

[trident9]

Greetings Boncuk
I did not find any information on using thermal drills (via) for air circulation through the board. Can you point me to an article? In the last design the via are to increase the thermal transfer between the top pad and bottom side pad, so they are close to the hot spot created by the MOSFET, but not inside the solder area for the MOSFET, except for the center via I'm using for manual soldering.
_______
trident