It's probably been discussed in full somewhere, however, what is the consensus on mounting a PCB directly on top of the outputs on the heatsinnk?
Does the fiberglass hold up?
I am thinking
heatsink/grease/oxide/grease/output/PCB/washer/locknut
Any problems squishing the fiberglass or it losing tension over time?
Gracias
S
Does the fiberglass hold up?
I am thinking
heatsink/grease/oxide/grease/output/PCB/washer/locknut
Any problems squishing the fiberglass or it losing tension over time?
Gracias
S
I cannot say that I have any experience with including the PCB in the stackup that applies clamping force, because it would not have been allowed in any of the companies I designed power supplies for (major computer companies). It was understood that the PCB material could compress over time, particularly in the presense of heat. PCB material is glass fiber in resin, and the resin is considered to soften with temperature. Perhaps we labored under a delusion.
Me too.
My power plates for powered speakers are assembled that way: board directly mounted to aluminum backplate which Is also the heatsink, either flat plate or finned extruded, with TO218/247 transistors sandwiched in between.
Works very well and assembly is very compact.
Use a compression washer (aka a ‘wavy washer’) 😊
I’ve honestly not had a single issue, including on my big 240W amps.
If you want to reduce heat transfer from heatsink to the PCB, remove the anodizing on the back of the heatsink (or add plain Al backplate) so it can't radiate heat, only conduct through the air. Or add a thermal barrier between the heatsink and PCB.
You can arrange that only part of the PCB overlaps the heatsink so part of it can run cooler. In general keeping electrolytics cool is a good thing to extend their life.
You can arrange that only part of the PCB overlaps the heatsink so part of it can run cooler. In general keeping electrolytics cool is a good thing to extend their life.
Won't be able to solder with the board mounted..
In my case, removing the parts will need each heatsink to be taken off
Theyre for the most part welded in between both R and L heatsinks once this is put together.
Not a big deal if it's as reliable as I imagine it will be
The use of a compression washer adds "stored energy" to the stackup which will compensate for flow in the PCB material. Not at all surprising to hear this has worked effectively. I found this drawing in an old standard which I believe is intended to to show the usage of a compression washer. Clearly it is not a flat washer and lock washer combination.
I should also note that most standards now identify "classes" for equipment, with requirements being more stringent for "enterprise" equipment than for "consumer". None of us want to build "consumer" grade equipment, right?
I think I would rather mount the PCB on standoffs on the heat sink and then attach the output devices through holes in the PCB. The holes need to be slightly larger than the heads of the screws (or outer diameter of washers) you intend to use.
I don't recall ever seeing a piece of commercial gear where a semiconductor has been sandwiched between the PCB and the heat sink. Not thereby said it hasn't been done; only that I haven't seen it.
Tom
I don't recall ever seeing a piece of commercial gear where a semiconductor has been sandwiched between the PCB and the heat sink. Not thereby said it hasn't been done; only that I haven't seen it.
Tom