CodeOfLight said:
Take a good look at the bottom of the shark board. There is a patch of what looks like tiny metal holes under where the IC is mounted. This metal square goes THROUGH the PCB to make thermal contact with the IC. It is a thermal via. When you pot the board in Arctic Alumina from the bottom, some of the AA will migrate into the little holes, and provides a thermal pathway to the sink.
Most of the via holes on my board are full of solder. And there are no vias beneath the zenner diode which will dissipate the most heat.
CodeOfLight said:
I suspect that the IC itself is mounted with thermal compound to the top of this thermal via. It will be the shortest pathway thermally because the IC will get the hottest on the bottom. This is because the actual components inside the IC are very close to the bottom of the DIP package. Heat sinking on the component side of the board will not be nearly as efficient.
when I hold the board up to a light, I can see beneath the diode, it is therefore not heatsinked to the pcb.
I've run 14W through this board (12W to leds, 1.6A input current, leaving 2W dissipated by the converter) while holding my fingers on the diode and linear converter IC. After a few minutes, the components got quite hot. The under side of the pcb was quite cool in comparison. This could have been because I was not using aae in the vias though. I doubt adding aae to the underside would have helped much if most of the holes were filled with solder to begin with. As I've also said, the diode has no vias beneath it at all.
I've heatsinked mine from the component side and it seems to be working fine. I'll do some extended run time tests ans let you know what I find. I've asked for clarification on this from Wayne, but I think he's busy at the moment.
I'm not saying you're wrong CodeOfLight about vias on pcb's. I just wonder how effective they will be on this board given their placement and manufacturing execution.