Plated aluminum kickplates for Rebels

[Oznog]

I've been wrangling with this problem of how to effectively mount these silly Luxeon rebels. Currently, for what I need in colors, they're pretty much a dead-on match for what I need.

As people have noted, a PCB has somewhat limited thermal conductivity and one of my probs here is that it's gonna be exposed to the environment and the copper will corrode unless I can pay for gold plating.

A plate of aluminum is vastly superior for heat dissipation, but we can't solder the pad; aluminum won't take solder. We can use thermal adhesive, Arctic Silver, but I suspect the thermal connection is not as good due to the thermal resistance being so critical on that tiny thermal backing.

So, here we go with another crazy idea....

You see a "brass kickplate" in the hardware store, nowadays those are usually aluminum plated with a brass finish (or anything other color). Actually it might be a titanium nitride finish, I'm not sure. But... would it take solder? Would it be possible to use this? Can the Rebels take the slow thermal cycle required to heat up such a significant thermal mass? How would the power pins be handled? Could we actually attach a kickplate to a PCB backing and milled into traces?

 

[tvodrd]

Circuit board material can can be had with a wide range of copper thickness. Circuit board houses can conformal coat them after assembly, which will greatly decrease corrosion. Fingernail paint, if you are into DIY.

Aluminum can be nickel-plated, and then gold plated, and both are solderable. I'm not familiar with aluminum kick plates, they're all real brass or stainless at work. Beware of gold-colored anodized aluminum should you elect to try one.

Larry

 

[frenzee]

I don't know the answer but I doubt they are brass coated aluminum. You CAN electroplate aluminum with copper or brass but you need an electroless nickel layer in between. Secondly, aluminum, compared to copper, sucks in both thermal conductivity and thermal mass. Copper is better by a factor of 2:1 and 3:1 respectively. Thirdly, brass's thermal coductivity is miserable. Adding as little as 5-10% zinc to copper reduces the thermal conductivity by about half, which makes it worse than even aluminum, so I would suggest skipping anything made of brass altogether.

I am in the middle of designing a copper core MCPCB for DIY projects, using a super-thin carbon fiber dielectric and thermally conductive epoxy lamination. I looked all over the internet for something like this but couldn't find anything at a reasonable price and in small quantities, so I resorted to making my own. Hopefully when I get it done, I will post the instructions and the materials list in here.

Regarding the corrosion problem, as tvodrd suggested, conformal coatings were designed just for this purpose and they are cheap and easy to apply.

 

[Oznog]

Don't coatings significantly increase the thermal impedance?

 

[65535]

Solder onto your copper piece, then coat the piece, just be careful around the LED>

 

[Oznog]

How would I go about figuring the thermal conductivity here? I'd like to know what thickness of copper is necessary to spread out the heat across the available surface area. Copper which is too thin will increase the total junction-to-ambient whereas there are going to be decreasing marginal returns when the copper gets thicker than necessary to achieve a fairly homogeneous temp across the board surface. Looks like I could have to carry the heat an inch or so around the device.

I tried to get some I think 8 oz boards awhile back, a guy from eBay said he had some but stopped answering my emails (I hadn't given him any money yet). Is there an easily available source of thicker boards?

I saw the design details for Rebel thermal boards. It is not easy to achieve good dissipation. One thing they emphasized was using many plated thruholes (PTHs), not to conduct electricity but to provide a great deal of extra surface area for dissipation. Also probably allows the bottom copper to help conduct heat to the outside but I'm not sure of that.

PTH would not only not be possible to do at home, but conformal coating would probably fill the holes and thus negate the value of the PTHs. So I'd be looking at a fab of thick copper and gold plating which is not cheap at all.

Some applications of what I want to make would clearly benefit from putting a cover on front to protect the devices and rely on the backing to dissipate the heat. This totally blows the PCB idea as far as I can tell because the board is a big thermal resistance and a few PTHs probably won't do nearly enough to carry the heat to the bottom copper layer.

I'm so missing the Lamina BL2000.