how much copper...
...do you think i would need as a heat sink with a Lux-RC L333 running full power?
I have the facility to use wire erosion and could make whatever dimensions I needed, also would have the capability to make the fins super thin so the area would be big while the weight would not.
Just wondering what you reckon you'd need in terms of square mm surface area roughly.
list of conductivity
I'm running 15 square inches of exposed AL, and when running 10 watts in still air, the case hits 50ºC in just over one minute.
With 8 to 15 mph of airflow, it doesn't hit the 50º trip point.
There is one piece of data for you.
The only issue I see with 'super thin' is that you will lose some heat path out to the fins; I believe the point of diminishing returns in aluminum heat spreaders is something like 5:1 or 10:1, depending upon who you ask (i.e. 10mm wide spreader @ 1mm thick.) I think the super thin fins will not be a problem, but I would just copy a known working heatsink fin spacing and thickness; let those engineers figure it out.
Later Edit: Two XPG run parallel, driven by an LFlex.
Last edited by P220C; 03-21-2011 at 01:37 PM.
Another pair of data points but NOT the module:
Three XPG's with Bflex (@1 A, assuming 80% heat from LEDs, 90% efficiency on driver = 8.8 W of heat) in a fully copper lined anodized Al housing of 7.5 sq inches. It did not reach 50 C cutoff even after an hour and a half, The entire body was wihin 2.5 *C of the temp at the MCPCB's, hovering just below 50 *C when protected from overhead fan. With the fan draft it dropped to about 40 *C. On the bike it barely felt warm even on a hot summer night.
An XP-E and two XP-Gs B2Flex, Full copper lining of a different design, slightly less effective, about 9.5 W of heat, same area. Hit 50 C cutodff in two minutes no fan. Hits 70 * C cutoff in 5 minutes under a slow ceiling fan, 20 minutes with oscillating fan, never hit it in a ride at about 7 * C and a 13 kph wind riding about 24-26 kph. This light may be a little over the equilibrium point for staying under 70 C on a hot summer's night. Fortunately 0.75 A running cooler is only a 10% lower output than 1 A output running hot.
These lights are pushing the limits hard using lapping, polishing, and thin AA to get it done. I would suggest as a minimum 1 square inch per watt of heat maybe per watt of light output plus driver heat to keep it cool, a bit more would give some leeway.
Thin fins have a diminishing return with height. If close together, you also create a boundary layer issue where you need higher air speeds to strip the warmer air away. So don't use sinks designed for fans as a model.
Then there are OTBs to consider. A wider shorter fin is more robust to tree limbs or just loading mishaps. Copper is 60% higher in conductance than Al but a bit lower in air transfer rate, so a sophisticated guess is you can use about a 2/3 as thick a fin for the same height as far as thermal transfer up the fin issues go. Since copper is easily bent, you may not want to go tall.
Hope this helps.
thanks for the replies guys, think i have some research to do....
Brian is there somewhere on here i can see photos of your housing? i'm intrigued by the copper lined aluminium.
Not hijacking thread. I think this saves a lot of back and forth.
I used EL34's Marwi bullet shaped housing. It is anodized Al, supposedly that is slightly higher in moving heat to the air than bare Al which is superior by a small amount to Cu. It comes with glass (if you prise it carefully off the halogen bulb), O-ring, threaded bezel, mounting point and Trailtech male connector and avaialble mounts With silicon trreatment it can be water resistant NOT water proof. To bad it isn't MR16 size.
The problem is even with bezel included, surface area is maybe 7.8 sq in. Three XP-G's @ 1 A are about 8 Watts of heat and another 0,7 for the driver. That means every inch needs to do its part, or the light will run too hot needing a step down to 0.75 watts, or a short life.
The Marwi bullet bodies from front to back have a 25.4 mm deep 32.2 mm ID section, then a cone section that is 18.3mm deep and 23.8 mm ID. A section of the botton of the cone part is open to allow a plastic mount and tailcone to attach.
The two copper lining designs are here.
Version 1 of the lights is here.
Version two of the low beam ended up here after running on batteries and a Bflex for a year. It and was cool @ 1 A even as a DRL in the summer at close to 40 * C and sun beating down on the black housing. warmed up if stopped but took its time doing so.
The second light is being rebuilt and evaluated and a deeper liner was needed. I made a new one thinned to 1/16" at the three lens barrel mounts but 1/32 between. Cut the new liner from the same type of pipe union that made the other liner:
A copper ring from house wiring takes the space of the flange on the MR 11 bulbs allowing the bezel to contact the liner as it bottoms out on the threads.
I will say no more on it here as I wish to post the revised version as a thread, other than it runs a bit warm with about 0.4 more watts to deal with. Can fix without external heat sink/fins
Hope this helps, Brian.