Thermal Epoxy Filler- Calcium Carbonate Experiment

[FighterCat57]

Sigh. Any idea of the thermal conductivity of a wax? Waxes are good insulators, with thermal conductivities typically under 0.25 BTU/(ft hr *F) and thus their use reduces the effectiveness of the whole concept by a considerable magnitude. Add the time needed to remove the mess at the warranty station and it just isn't a good idea for a manufacturer. But like I said, try it.

Keeping an oil inside an ESC has the same problems as keeping water out, except oil usually wicks a lot easier. Transformers never move when in use, ESCs are constantly having their wires tugged on and bent. Who wants a gooey mess inside their boat? And the box makes the ESC that much larger and more awkward to install....

No need to guess, fans are a proven technology and work well for those who have bothered to use them. The larger the hull volume the better they work, and in some cases at least they work better than any water cooling. KISS is one of my main guiding principles with FE boats - reliability is important to me.


.

I don't think it's actually a wax. Like you said wax is a good insulator. The process is referred to as a pre-wax. I don't know the actual material. Anyway, that's moot and beyond the scope of this experiment. Just stating that resin casting is a plausible solution and inspection repair is possible in some processes.

Ignition systems use a resin cast to prevent water damage and help dissipate heat. This is nothing new, just being applied to an ESC in this experiment. Some sandbar resistors also use a resin cast to dissipate heat. SSDA

I've worked with plenty of little gadgets over the years that have been resin casted and just don't see why it's not being used for marine ESC's.

I've hooked a fan up to an ESC and kept it cool. That worked well. In fact I'm using turbo fans with ducted cooling on a pair of CC240's in a Sprint Cat right now.


I <3 you guys.

[FighterCat57]

Hahahaha if only you knew.... I've set several hoods on fire... the worse are the reactions where water causes it to react more.... What are the amounts you are using? I haven't noticed you list what parts for what... (IE 4 parts epoxy to 1 of the CC)

Sorry, missed this reply. 1 part CC, 1 part A mixed in with 2 oz part B
Loctite 1hr Marine

Mix CC & Part A by saturation, 24 hours sealed (vacuum if avail)

Mix CC/A & Part B


My control batch I made on the side in a yogurt container is solid as a rock, so it's definitely a mixing mistake on my part.

[sailr]

There are different types of epoxy. Casting resin and laminating resin. You need casting resin for what you're doing. Laminating, when poured in a thick mass will crack as it heats up.

First point I was going to make was check the mix ratio. Polyester uses a catalyst, so a little bit will still cause the chemical reaction to start, just takes longer to grow the molecules.

Epoxy is a hardener. Part A reacts with Part B, so you need the proper amounts.

As mentioned, heat will help it react. When I water cool my ESC's, I use room temp laminating resin and let it kick off. I then heat the oven until the temp starts reading that it is climbing, maybe 170ish? I then shut it off and THEN put in the coated ESC's. The oven holds the heat and warms the epoxy enough to accelerate the cure. works well, mine run underwater, but the largest is a 90/100.

The best I have seen is cooling in an electrically inert fluid. thier are fluids made just for this or there are oils that will work also. The oil will keep water off of the controller and flow around the ESC while conducting heat. An aluminum plate can act as a heat sink in the oil, or make the case from aluminum and have it mounted in the bottom of the boat. Used to do this on our scale RC combat ships to cool voltage regulators and warm the CO2 lines.

Did some work helping a grad student cool data acquisitions boards for a NASA research plane. the fluid definately cooled the boards well, but it is a fluid and has to be contained.

Brian

[FighterCat57]

I've been working with Polymer Composites and believe we have come up with a potting resin similar to sand bar resistors that will work for this use. (A calcium carbonate mix) Cost is about $8/lb for resin & hardener.

It will be here next week. About the same time as the Suppo ESC's.

I've decided to use some small copper plate and solder copper tubes for water flow. The copper plate is $4 ad the tubing is .89/ft from K&S.

I picked up an Eagle Tree for testing and will setup a hull to test.

[FighterCat57]

Yes but I feel like there is a trick to it or something that needs to be done.... I still haven't found anything but I've just got a nagging feeling there's something.....

Synthesis of materials are very tricky as you probably know....

Yeah, there's a huge problem. It's the wrong formula entirely.

You already have thermally conductive epoxies, designed for this... They are actually thermally dissipative, but will be ok in our applications.

http://uk.rs-online.com/web/search/searchBrowseAction.html?method=searchProducts&searchTerm=552-668+&x=0&y=0

As far as I tested this one it works ok.

The Thermal Conductivity on that formula is only 1.26 (Aluminum is 250 and copper is 400)

Something in the back of my head is SCREAMING that this doesnt work (I'm an organic chemist an work in a lab... I've played with alot of calcium carbonate.) But I can't put my figure on it.... I need to mull this over and get back to you, but its a great idea

A decimal point perhaps?


We just created bubble gum insulation.



WTH is wrong with people these days. They'll tell you anything to sell some *!***!***!***!**. Polymer composites is now saying that I misunderstood them when they told me the mix would yield in the 350's... They meant .35's.

The thermal improvement of CC is in the decimals, from .25 to .35.
Copper is in the 400's.

I really *!***!***!***!**ing hate being mis-informed.

The cool news is that I have some really neat casting resins and stabilizing fillers to use on something later.

[CornelP]

You're right here... but look at the bright side: this resin has one of the highest values of anything that is not metal... These are used to be thermally dissipative: take away the heat from a hot spot on the PCB.
If you take a look at the heatsink pads, the best ones only get to 12W/mK:

http://uk.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=7074774

As for liquids, the Boron Nitride compounds get only to 1.85.

I'll ask around, maybe there is something out there...


PS: sodium is 84...

[FighterCat57]

You're right here... but look at the bright side: this resin has one of the highest values of anything that is not metal... These are used to be thermally dissipative: take away the heat from a hot spot on the PCB.
If you take a look at the heatsink pads, the best ones only get to 12W/mK:

http://uk.rs-online.com/web/search/searchBrowseAction.html?method=getProduct&R=7074774

As for liquids, the Boron Nitride compounds get only to 1.85.

I'll ask around, maybe there is something out there...


PS: sodium is 84...

Yeah, this is where it get's all screwy and real world experimentation comes in to play. IDK WTF I'm doing with any of this, but I'll probably embalm one of the Suppo ESC's in resin and run water tubes through it. Why not, right?

PS, check your email. =)

[befu]

There are different types of epoxy. Casting resin and laminating resin. You need casting resin for what you're doing. Laminating, when poured in a thick mass will crack as it heats up.

fightercat will probably need casting resin for what he is doing as the cast sounds like it will have thick areas. The couple dozen ESC's and voltage regs that I have done all work perfectly, but I am brushing on layers of resin to seal the ESC with the heatsink. In that application, either seems to work well. i also use an epoxy with over an hour of pot life and a full day to cure, kinda slow and doesn't heat up. Many ways to skin a cat....

Brian

[FighterCat57]

fightercat will probably need casting resin for what he is doing as the cast sounds like it will have thick areas. The couple dozen ESC's and voltage regs that I have done all work perfectly, but I am brushing on layers of resin to seal the ESC with the heatsink. In that application, either seems to work well. i also use an epoxy with over an hour of pot life and a full day to cure, kinda slow and doesn't heat up. Many ways to skin a cat....

Brian

Yeah, I've got some actual Casting resin mix and the formula is .35 Wk

I've also got Suppo making a batch of Copper plated ESC's. One plate per layer. Don't know what the price is yet though...

[jac4412]

Bubble gum mold eh? Is there a link to the procedure/formula you have I can view? And I mean maybe... For a thinker (bubble gum) consistency you could have easily mixed the parts up... If you don't use enough of your reaction "driver" everything will get messed up...

From what I've been reading if you add just a bit over the set cc it will disrupt your epoxy, ruining the setting and instead of your nice adhesive... some nice syrup.

What kind of CC are you using? I've also been reading that the recommended size is "nano size" so thats not going to be your standard chemical order either.

http://cat.inist.fr/?aModele=afficheN&cpsidt=21840281

[FighterCat57]

Bubble gum mold eh? Is there a link to the procedure/formula you have I can view? And I mean maybe... For a thinker (bubble gum) consistency you could have easily mixed the parts up... If you don't use enough of your reaction "driver" everything will get messed up...

From what I've been reading if you add just a bit over the set cc it will disrupt your epoxy, ruining the setting and instead of your nice adhesive... some nice syrup.

What kind of CC are you using? I've also been reading that the recommended size is "nano size" so thats not going to be your standard chemical order either.

http://cat.inist.fr/?aModele=afficheN&cpsidt=21840281

Oh, I messed up the mix by mis-calculating 1 part of hardener. It's in an earlier post.

The CC is specifically for epoxy mix from polymer composites in Ontario CA

I have the potting resin I'm going to use (4 hour set time) to coat an esc thoroughly, but just to water proof it and see if it's useful.

[FighterCat57]

I just chatted with them again and they assured me that a mix of CC/Max MCR would be a good thermal dissipater. So... I will donate another ESC to the cause and mix up a new batch of Max MCR and see if it works.

The Max MCR is a Potting Resin. The mix is 25% Calcium Carbonate Powder by weight added to Part A.

I'm still not sure if I'm going to do a mold or if I'm going to just run the coating over the piece until it's fully coated. My main concern is assuring saturation of the middle board.

Just playing around with the mix... Not expecting much beyond the obvious.

[questtek]

Based on my experiments and extensive testing with the HiModel 200 A esc's, I would highly recommend against casting resin. It becomes very brittle when heated and in my opinion is not suited for your application.

I would recommend the two part thermal conductive but electrical non-conductive 2-part black potting compound shown in the attached picture. This does not become brittle, is designed specifically for your application and is easily mixed 50-50 parts A and B.

[FighterCat57]

Based on my experiments and extensive testing with the HiModel 200 A esc's, I would highly recommend against casting resin. It becomes very brittle when heated and in my opinion is not suited for your application.

I would recommend the two part thermal conductive but electrical non-conductive 2-part black potting compound shown in the attached picture. This does not become brittle, is designed specifically for your application and is easily mixed 50-50 parts A and B.

Agreed. Max MCR is a potting resin, not casting.
High temp and the added CC is recommended by the mfr at 25% or increased thermal conductivity and physical stability.

Should be fun!

(the real solution is using copper plates on each FET row at the mfr, which is coming soon)

...or better yet, using the appropriate componentry and not generating heat to begin with.