Oznog
Active Member
I am planning to do a project with 3x Luxeon Rebel LEDs running at 2.5W each. They have a solderable copper thermal pad on bottom. I have about 2 sq in of area for this board. These LEDs are new and remarkable in that they can maintain a long life at 135C junction temp, among other things.
Still it is beneficial to keep this system as cool as possible, for both better output and not burning people.
I have seen the most likely solution seems to be using the PCB area as cooling and perhaps some additional aluminum area, which would be a tricking mating issue.
I have some PCB 0.075in thick, probably 1 oz (maybe 2oz?) copper. I also have some super-thin stuff, 0.016" 1/2 oz. I can get some 5oz copper too but the board is quite thick. I might thermal epoxy the PCB onto some aluminum to get some cooling out of the backside.
Here's my primary question. The thicker boards here would reduce thermal dissipation through the board to an aluminum backing. However the thinner boards come with thinner copper and that would cut down on lateral is thermal conduction. This will result in the copper near the component getting hot while the copper near the edges does not get very hot which reduces the dissipation on the topside as well as the conduction through the board to the backside. I did a test with the first board and a thermal probe, could see the temp gradient moving away from the component, wasn't all that high but it was definitely a different temp right beside the component and 1/4" away.
So how can I determine if the thicker board/thicker copper or thinner board/thinner copper (esp thinner board on aluminum) is the way to get the lowest total thermal resistance? Is cooling through the backside even going to have a significant effect, knowing that PCB material is not a strong thermal conductor?
There are additional mounting concerns which keep leading me to the conclusion that the top surface may need to be covered so it will see limited airflow, back to the idea that the copper needs to spread out the heat so it can effectively leak through to the back.
Still it is beneficial to keep this system as cool as possible, for both better output and not burning people.
I have seen the most likely solution seems to be using the PCB area as cooling and perhaps some additional aluminum area, which would be a tricking mating issue.
I have some PCB 0.075in thick, probably 1 oz (maybe 2oz?) copper. I also have some super-thin stuff, 0.016" 1/2 oz. I can get some 5oz copper too but the board is quite thick. I might thermal epoxy the PCB onto some aluminum to get some cooling out of the backside.
Here's my primary question. The thicker boards here would reduce thermal dissipation through the board to an aluminum backing. However the thinner boards come with thinner copper and that would cut down on lateral is thermal conduction. This will result in the copper near the component getting hot while the copper near the edges does not get very hot which reduces the dissipation on the topside as well as the conduction through the board to the backside. I did a test with the first board and a thermal probe, could see the temp gradient moving away from the component, wasn't all that high but it was definitely a different temp right beside the component and 1/4" away.
So how can I determine if the thicker board/thicker copper or thinner board/thinner copper (esp thinner board on aluminum) is the way to get the lowest total thermal resistance? Is cooling through the backside even going to have a significant effect, knowing that PCB material is not a strong thermal conductor?
There are additional mounting concerns which keep leading me to the conclusion that the top surface may need to be covered so it will see limited airflow, back to the idea that the copper needs to spread out the heat so it can effectively leak through to the back.
