I occasionally see posts where someone comments that they potted a driver using silicone or used silicone for thermal conductivity. I’ve been building computers since the mid 90 so I know that very thin layers are used between chips and heatsinks but potting means significant thickness. Coming into this interest in flashlights having worked in the unrelated field of window manufacturing I frequently worked with silicone when making multi-pane inserts for windows. Silicone was typically used with a butyl rubber strip reinforced with aluminum or an aluminum or steel spacer with a bead of butyl rubber or other material that acted as a thermal break due to its low thermal conductivity and resiliency. For me, use of silicone to conduct heat is a bit confusing.
I’m currently waiting for silicone to cure that has been formed to fill the space between a reflecytor and the walls of the head specifically for its poor thermal conductivity and am brainstorming on a possible future build where titanium and silicone will be used as a thermal break between the emitter housing/head, driver housing, and battery tube. The emitter housing (which would also incorporate heat tubes) and driver housings would each have fins to maximize surface area but be thermally isolated. Too much detail… Anyway, why is it that people use silicone as a medium to transfer heat when its thermal conductivity is so low?
I think some add silicone carbide (?) to the mix for better heat transfer… I have potted with JB weld before and it seems to work well, just don’t plan on taking it apart later.
For a reduced thermal conductivity coating I’ve used jb weld mixed with pumice. Silicon carbide has a high thermal conductivity vs silicone (sealers, rubber, etc). If I pot a driver it will be a jb/silicon carbide mix heavy on the carbide. Based on my experience and logic silicone would be better as a thermal break or insulator.
You cant transfer a lot of heat from a small source like a LED over a big distance, but anything in there is better than air
It helps spread the heat and conducts also some heat
Of course mixing it with aluminum oxide or carbon increases the heat transfer
“Silicone” is just the name for anything in [R2SiO] form, a polymer. What each R happens to be determines its properties. (Just like ROH is an “alcohol”, which can be anything.)
Silicone grease/gel is the thin gel that you can lube O-rings with, insulate sparkplug boots, etc. Add heat-conducting powder to it, and you can make a gel that transfers heat nicely w/o letting the powder work its way loose and blow away.
That’s pretty much it.
Silicone rubber is a rubber made from silicone components, raw/liquid/gelatinous if uncured, solid when cured. Think of “silicone RTV” (room-temp vulcanisation, ie, no blast-furnace required ) But silicone rubber is used for wire insulation, sparkplug wires, etc., and stands up to (relatively) high-temps and holds its shape well when hot.
Anyhoo, silicone makes a nice carrier for those materials which are thermally conductive. Think of conglomerate, solid rocks that are held with a binder of cement. The ratio of rock to cement is quite high, and the cement just fills the gaps, when ideally the individual rocks make actual contact with each other.
Lookit “conductive ink” which can be used to repair PC board traces, the heating element wires on rear-window defoggers, etc. The ink is nonconductive, but the flakes of copper suspended in it are. Sure, in this case the carrier dries and leaves the copper flakes cemented in place, but there’s still plenty of flake-to-flake physical contact which makes electrical contact as well. Same with thermal goop, only with heat instead of electricity.
Thus my reason for filling the airspace with a rtv silicone material to utilize the thermal conductivity of the metal more efficiently. Eliminating the internal airspace or filling with an insulating material should result in more efficient heat dissipation.
I know there are limiting factors. Larger wires = greater current flow = better thermal conduction between the driver and the emitter but I suspect there should be methods possible to insulate the driver from the emitter that will facilitate better heat dissipation from the driver independent of the emitter. Thermal breaks like I’m visualizing might be a good first step.
@Lightbringer I understand what you are saying. In many cases these individuals seem to be using a rtv type silicone with no additions. By adding a thermally conductive material like silicon carbide it can be improved but as the balance tips toward the conductive material particle size, shape, dispersal, and orientation become more important.
Maybe I’m wrong but I suspect that there’s a way to cool the emitter, regulators, fet, and cells while minimizing the thermal interaction.
You mentioned above about “reducing thermal conductivity”, ie, making a thermal insulator. Which is fine if you want to keep a scorching LED from cooking your battery (at the cost of more bottled-up heat total).
And I can understand potting a driver with RTV (’though low-outgassing!!!) if you’re going to mount it on a howitzer, but why would anyone in his right mind want to use RTV as an insulator to “help” thermal management?
You can spray the driver with a conformal coating to insulate the nekkid electronics, then fill the driver with copper sand for all you’d care, as nothing would short out. Miller-Stephenson makes a buncha conformal coatings.
But using a thermal-insulator?!? I gotta be missing something…
You’re thinking in the wrong direction about the Silicone. Compared to lots of other things, it may be an insulator, but compared to air, it is a good conductor! So, as someone said above, it’s better than nothing for thermal conductivity. Silicone is also easily worked in, and a lot easier to remove later than Epoxy, JB Weld, and other substances. Removability might be important if you ever need/want to replace/repair/upgrade the driver or other components.
I use JB-Weld; mix it with isopropyl alcohol to thin it. When it cures, it has a rubbery texture, but more importantly, it doesn’t seem to adhere to anything, so it is removable.
(Actually, I use the cheap Harbor Freight equivalent to JB-Weld. About a thirdof the price, and it seems to be the same product…)
I had some enclosed 110V LED drivers for 12V COB floodlights one time. I decided to tear one down and it was potted with a typical gray compound. I started picking the material out and found that it was full of small ROCKS. I guess rocks are cheaper than potting compound, so they filled the driver housing and then just topped it off with the potting compound.
@ Lightbringer In a typical light the star is attached to the head which provides a path for heat out of the light. As the driver is also attached this shared portion of the light also provides a path for heat from the emitter to the driver. The idea is to create a modular light where the emitter and driver each have their own thermal path insulated from the other. There will still be heat transferred by the wires or contact points but insulating the emitter module from the driver module would reduce heat transferred from the emitter to the driver.
I have grown several strains of turkish and they grow quite nicely here. My space is limited but I typically grow around 200 plants. This year it will probably be white gold, virginia bright leaf, a couple of turkish types, and a burley of some type. I’m totally out of bright leaf right now and have been using a ‘pipe cut’ commercial tobacco and it will be nice to enjoy my own blends again.