Aluminum vs Copper vs Brass

[350xfire]

Guys:
I am about to start building some SST 50/90 stuff and have noticed some using copper heat sinks. I have also noticed the big difference in price from copper to aluminum...

Is brass a good alternative to copper or not? Reason I ask is I have some laying around.

Is anyone making off-the-shelf copper heatsinks?

 

[Connor]

The thermal conductivity of brass (~115 W/(m·K)) doesn't even hold up to pure Aluminum (237).
Copper has 401 W/(m·K) and is the best choice for SST 50/90s.

 

[350xfire]

The thermal conductivity of brass (~115 W/(m·K)) doesn't even hold up to pure Aluminum (237).
Copper has 401 W/(m·K) and is the best choice for SST 50/90s.

That's what I needed thanks

 

[PCC]

(Here we go again... :wave

There has been a lot of discussion about the relative merits of copper, brass, and aluminum as used for heatsinks. The conclusion that I came up with is that copper is best used when you are pushing the design of your light or the LED to the extremes. This is because C110 copper has roughly twice the heatsinking ability compared to aluminum. C110 copper (99.9% pure) is roughly half the cost of C101 copper (99.99% pure) but that 0.09% difference accounts for C101 copper being roughly 4% better as a thermal conductor. So, if you have a small heatsink and want more but cannot use more material due to size limitations then use C110 copper instead of aluminum. Likewise, if you are pushing an emitter really hard (or run multiple emitters hard) then copper would be a better choice. Don't forget that copper is much heavier than aluminum, so it will increase the weight of your light. In the 3D DD SSR-90 light that I built for BC the two inches of copper weighs about 1 lb (2.2kg). Brass is pretty close to copper in weigh by volume, I believe, so it, too, will add weight to your project. I'd rather use aluminum over brass any day due to the weight issue and, more importantly, the thermal performance is lacking compared to aluminum.

 

[unterhausen]

if the copper is going inside an aluminum case, then the advantages steady-state are much diminished.

 

[PMM]

Its a daft analogy IMV

Greater importance is the designs ability to radiate the heat away.

If you don't shift the heat be it ALu or Cu whatever slug you use will heatload to the point you will have a temp overload.

Copper is a better transferrer of heat
Aluminium is a better radiator of heat

Neither are any good unless you design to transfer that heat out to the air.

 

[350xfire]

Its a daft analogy IMV

Greater importance is the designs ability to radiate the heat away.

If you don't shift the heat be it ALu or Cu whatever slug you use will heatload to the point you will have a temp overload.

Copper is a better transferrer of heat
Aluminium is a better radiator of heat

Neither are any good unless you design to transfer that heat out to the air.

Ok this will be for a dive light application in maglite head and a single sst 50 or 90.
Thanks

 

[Mick]

For very good heat transfer from the SST90 emitter you can solder the device to a copper heat spreader, then use Arctic Silver compound between the copper and the aluminum housing. I use an Indium solder that melts at 135 C and solder the wires with lead free that melts at 230 C. This way the leads will not fall off when attaching to the copper heat spreader. The Indium solder has a much better thermal conductivity than Bismuth low temp solder. (19 w/mK vs 73 w/mK). Arctic Silver epoxy has a conductivity of 7.5 w/mK so the Indium solder connection is ~10 times better at heat transfer.

 

[wquiles]

For very good heat transfer from the SST90 emitter you can solder the device to a copper heat spreader, then use Arctic Silver compound between the copper and the aluminum housing. I use an Indium solder that melts at 135 C and solder the wires with lead free that melts at 230 C. This way the leads will not fall off when attaching to the copper heat spreader. The Indium solder has a much better thermal conductivity than Bismuth low temp solder. (19 w/mK vs 73 w/mK). Arctic Silver epoxy has a conductivity of 7.5 w/mK so the Indium solder connection is ~10 times better at heat transfer.

Very interesting - thanks. Can you please let us know which specific Indium solder you use (part #, model, etc.) that has the low melting temp?

Will

 

[Mick]

I need to clarify a couple issues on the Indium soldering. I got most of my technical info from: http://www.indium.com/ but I bought the solder from: http://www.micromark.com.
The reason being Indium Corp appeared hostile to small users; they wanted to sell me a sample kit for $350. I think they are used to doing business with the Govt where the price is no object.
The solder I got from Micro-Mark is called TIX and is made by Allied Mfg. Co. in Bozeman, MT. I do not know the alloy composition. I had trouble finding them on the web so I don't know the exact thermal conductivity but guessed from info on the Indiun Corp website.
According to an engineer I emailed at Indium Corp it is not recommended to solder directly to copper with Indium solder as a brittle alloy is formed.
Ref: http://www.indium.com/techlibrary/applicationnotes.php
See: http://www.indium.com/_dynamo/download.php?docid=14
Indium/Copper Intermetallics

I solved this problem with a nice Nickel plating kit from Caswell:
http://www.caswellplating.com/kits/plugnplate.htm
I bought the optional pen wand with the dome point and it work great. It takes about two minutes to plate just the area that the SST90 thermal pad will sit on. Then everything is OK to solder with the TIX Indium solder. I used the TIX flux also.

I will try to do a nice "Will" type tutorial on my method in the near future.

 

[wquiles]

I will try to do a nice "Will" type tutorial on my method in the near future.

Thanks - that would be awesome!

I recently did solder an SST-50 directly into a custom copper heatsink, and although it took a couple tries, I was able to solder the two power wires from the sides (while the emiter was in place!):

So that you are doing is of a lot of interest to me as that would have made my life easier for this custom project I am working on :thumbsup:

Will

 

[Mick]

Here is a photo extracted from the Alibre CAD of a SST90 sitting on my copper heat spreader. The wires get soldered first then the LED gets placed in a jig with the copper disk on top. The wires are in the grooves, This unit is then flipped over and placed on a hot plate to re-flow the Indium solder. Pardon the salmon colored copper. LOL

assembly:

 

[350xfire]

Cool 3D stuff!!!
So question becomes- if I use a copper heatsink, the heat will initially be dissipated quick until it hits the aluminum, right? Once that happens the aluminum's thermal characteristics will determine how fast total heat is dissipated... So, is it even worth using copper on an aluminum based head? Especially on a single LED light?

Thanks for all the discussion guys!

 

[Mick]

The copper conducts the heat rapidly away from the small thermal pad on the led then spreads the heat out over a large surface area into the aluminum. This improves the thermal path to the outside of the aluminum body where it can be removed by radiation and convection. A better explanation would require math and an understanding of thermodynamics. If your really interested a good book is Heat Transfer Engineering by HilbertSchenck,Jr.

 

[precisionworks]

The P7 Mag from Mac had the emitter bonded to a huge copper sink, and the SST-50 Catapult V2 uses a "2 oz brass PCB". Gene Malkoff has always used a thick brass shell (although I don't know how his thermal path from emitter to shell works).

is it even worth using copper on an aluminum based head? Especially on a single LED light?

Way back when, meaning two or three years ago, there were few power LED's available. Most emitters were running in the 1 watt or 2 watt range. Then Cree brought out the X-Lamp and everything changed.

I can't remember the amp draw on the P7, but the SST-50 Catapult V2 measures 1700 mA @ 8 volts. That's 13.6 watts, from an emitter driven at a moderate level of 3500 mA. I spoke recently to a modder that's running an SST-50 at 5000 mA drive

My best guess is that a good thermal path is critical to long emitter life when the light level exceeds around 50 lumens OTF ( roughly 75 bulb lumens). Most everything today pumps out way more than that

 

[UnderTheWeepingMoon]

Way back when, meaning two or three years ago, there were few power LED's available. Most emitters were running in the 1 watt or 2 watt range. Then Cree brought out the X-Lamp and everything changed.

I think that heatsinking has been a concern for longer than the the introduction of the X-Lamp. The old Luxeons had prodigious cooling requirements. If anything, the X-Lamp reduced the need for heatsinking because of its improved efficiency, although it's true that we drive LEDs much harder these days than we used to, with better electronics available.

I think you're right in that we saw many improvements in heatsinking about the time that the X-Lamp came out and a movement away from the use of plastic in LED lights. Some old Luxeon lights still had pretty impressive heatsinks by today's standards, though, such as this one by Icarus.

 

[precisionworks]

4xUWOJ LuxIII emitters sit on a custom copper heatsink with individual McR-20 reflectors

Yow

I had no idea the Luxeons put out that much heat.

 

[Mick]

This is a Lumen / Watt definition from the dept. of Physics, University of Georgia.

Luminous Flux

The radiant power is the total radiated power in watts, also called radiant flux. This power must be factored by the sensitivity of the human eye to determine luminous flux in lumens. The standard definition is as follows:

As you can see a incandescent bulb is horribly inefficient at 2.5%.
A LED that produces 100 Lumen / Watt would be about 15% efficient. So 85% of the applied power is lost as heat. The incandescent bulb radiates much of this lost energy as infrared radiation. A LED does not emit infrared so the heat must be removed by the heat sink / light body.

 

[archer6817j]

We are just talking about the ability to dissipate a given amount of heat over a given amount of time right? So, wouldn't an ice cold aluminum heatsink (maintained at that temperature) potentially dissipate more heat than a copper heat sink at room temperature?

 

[65535]

Well yes, but only because of the high temperature gradient in the material.

You'll notice the units for thermal conductivity contains, temperature, distance, and energy.