Cooling used to be a problem, back in the 60's and 70's. But then someone (I've personally met, BTW) invented the 3-axis stabilised bus and mounted a black body radiator on the opposite to the sun side. Problem solved.
So, if one takes one of the Avalon modules aluminium radiators (ask nicely, admit that they design is sufficient), chemically blackens it (increase surface), and then mounts it on the opposite to the sun side of the cooler-box contraption, it should work.
But still, the more difficult thing to do is to convince Elon Muck to donate +/- 50kg payload space...
Even with an emissivity of 1, do you know how hot that Avalon heatsink would have to get to dissipate 150W radiatively?
Stock Avalon radiator will perform quite poorly, although it's shiny and looks cool and high-tech in silver. It must be anodised (or processed otherwise) to increase it's surface area at least 30-40-maybe-even-50 fold. Then it would need to dissipate only +/- 70W (my estimate).
Yea, I think it should work - but can you imagine the delta-T on that config? BFL, have you tested the lowest temperature of your ASICs, can they perform reliably at 20K?
BTW, in my previous life I've sent couple of my design and manufacture modules in space, and for my surprise they've worked for over 10 years without overheating. Admittedly, that was over 25 years ago, and technology progressed since that time...
No offense, but that makes no sense to me. Regardless of what you do to the heatsink the power draw of the chips isn't going to be significantly reduced unless you undervolt them and downclock them. They might run a little less power at cooler temperatures, but temperature will almost certainly be higher than it would be on Earth. For reference, the heatsink looks very similar to a 1ft length of this profile.
http://www.aavid.com/products/extrusion-heatsinks/76620The 12" would have a total surface area of 406 square inches, or 0.262m^2. However it would be a horribly inefficient design as a spacecraft radiator because the majority of the surface area is perpendicular to other parts of the heatsink. Even if you chemically treated it to increase the surface area it won't overcome the fact that the geometry of that heatsink is terrible and most of the heat that is radiated would just be reabsorbed elsewhere in the heatsink. The effective surface area would be much lower.
None taken, I've playfully suggested Avalon radiator, because many has seen the picture of it and get the idea of what size more or less one would need to get 70W dissipated. But I do not believe the ribs will be a problem - properly prepared surface is selective in nature anyway. Also, the surface is etched to achieve fractal shape, which increases the size considerably. The Avalon radiators will be overkill for sure, that's the reason for the 20k joke!
The designs of the radiators I have worked with were more of a cavity radiators, flat plane covered with micro-conical shape, and they were very efficient in dissipating heat - those days our power budget was less than 200W and the radiators were small - around 20cm x 40cm. The temperature (I'm struggling to remember the exact figures now) was in the order of 100K on the surface of the radiator, not much higher inside, so the components were super-cooled all the time. The tricky part was to manage to move the heat without use of heat-pipes (no-gravity, remember), and plane silver rods were too inefficient - so we ended up sticking the components to the shaded wall in direct contact. That worked well. The only risk in that config is the micro-meteorites and the high-velocity dust - they go through the metal like butter - but are stopped very effectively by the Styrofoam in any outdoor-type cooler. Hence me suggesting that, again half-jokingly, half-seriously
. Perhaps one can design a gap-trap in a shape of black body radiator. Hmmm, that could work, I have to check what Boeing is doing on their latest and greatest.
But in this example, I think it will be very easy - remove the stock BFL Single SC's cooler, affix the ASICs (via the thermal pad supplied) directly to the radiator, install on the shaded side. The rest of the boards I would not worry, but nothing stops from installing on the same radiator - there will be plenty of real estate there. And if it gets destroyed by the dust in couple of years, so be it - at least for the time being everyone will have fun.
@Puppet
If you have to use motors to orient the satellite in 3D, you have to make provision for the fuel - and that will limit the life of the satellite, plus will make it very expensive. The small LEOs do not have any orientation control - they rely on internal gyros or spin the whole bus as a gyro, to keep them in one direction and correct the antennas if needed. In your example, it will be more efficient to stabilise the bus in 3D and then correct the PV panels and the radiator to point in the desired direction, then use omni-directional antenna for comms with the Iridium constellation and that one you would not care which direction it is pointing.
