With all due respect, if spacesuits have to be carefully designed and have onboard air conditioners just to compensate for the inability of a space vacuum environment to absorb radiated human body heat. I don’t think it takes a page of equations to understand that running a CPU in space will initially (form the CPUs perspective) be like running it in a Thermos bottle. Space will of course eventually absorb heat but it does so slowly and something as hot as a CPU will cook itself in short order.
Here’s a cite from NASA that space suits are cooled actively:
Except that NASA tends to stick it’s astronauts out in the sun where they can watch them work, so naturally the sunlit side is going to get hot (up to +235 F), without anything to convect that heat to the shaded side (where it’s a nippy -180 F), they’re going to need an onboard A/C unit to keep cool. Would they need those same A/C units to keep cool if they worked exclusively in the shadows? They’re hands dropped down to 17.2 F even when they were working in the sun. From the page I linked to earlier
In reading that, it sounds like simple body heat’s not enough to keep a person in a spacesuit warm if they’re in shadow.
Tuckerfan: you are confusing two different concepts: a cold object and a lack of a heat source. Near Earth space is not cold in the sense of an ice cube being cold. I.e., in space you are not in contact with much that is either hot or cold. Stuff that is getting radiated by the Sun can warm up a bit (depending on surface reflectivity, etc.). Something in shadows just doesn’t become warm. What heat it has is slowly radiated away.
By your logic, the interior of the the vacuum space of a Thermos™ bottle should be extremely frigid. Ergo, it would make it a lousy place to store hot coffee.
There are three basic ways to transfer heat:
- Conduction: Nothing to touch in space. Heatsinks use this. And if you connect it to a large copper tube down to a subsurface wet layer of soil, you can avoid fans!
- Convection: Nothing to convect in space. (But we got air on the surface.)
- Radiation.
So on Earth you have 3 of 3 options. 2 of 3 are generally used. In space you have 1 option.
In short, without convection, your P4 fries in a flash.
I think time is the point at issue here. Fingers in a spacesuit will undoubtedly get cold after being a space a while. An Athlon or P4 CPU kicks up to hundreds of degrees in the space of a few minutes. The insulating vacuum of space doesn’t, by any stretch of the imagination, have enough time to move the radiated heat away from the tiny surface of the CPU before it cooks.
Tuckerfan, NASA has lots of experience putting computers in space, so your question constitutes something of a “known problem.” I work on a satellite program, so I know some (but not all) of the solution. I won’t go into the painful, painful details, but I will try to clear things up for you.
Let’s suppose that you’ve got a spacecraft (aka “the bus”) carrying a sensor, and that spacecraft is in a sun-synchronous polar orbit. This means that its orbital plane is roughly north-to-south, and on each “ascending” pass (northbound) it crosses the equator at the same local (solar) time. And for kicks, let’s put this satellite in a 0600 orbit (which means that it passes the equator at 0600 ascending, and 1800 descending). This means that its left side is always facing the sun (ascending, it’s a sunrise; descending, a sunset).
Because its direction of travel never changes, you will have a “forward” and “backward” direction defined with respect to the satellite body; you will have an “earthward” and a “spaceward” axis; and you will have a “sunny” and a “dark” side of your spacecraft. On the “earthward” side you place your earth-observing instruments. On the sunny side you place solar panels for charging your batteries. And on the dark side, you place radiators (you can also place these on the spaceward side if the geometry is right).
Since your electronics are inside the spacecraft, and wrapped in insulation, they’re safe from Mangetout’s boogie monster, thermal expansion (a very real problem). Since the satellite is exposed to nearly the same sunlight all the time, its structure can be built with the “sunny” side in tension and the “dark” side in compression (so that on orbit, some but not all of the pre-stress is absorbed by thermal stress, and the satellite doesn’t change shape, thereby keeping its center of gravity intact).
If you have electronics that tend to give off lots of heat, you insulate the hell out of them, but place a hole for a heat pipe. Run the heat pipe to the radiator, and cross your fingers. A lot of times, however, things get so cold that you actually need to run heaters to keep your sensitive electronics at a standardized temperature. I’m sure that some clever sonofagun has figured out how to use excess heat from one place on the spacecraft to keep another part warm, but I’m not yet aware of how it’s done.
One last thing: if you’re in a 1200 orbit (where your spaceward side gets pure sunlight on the ascending node, and pure darkness on the descent) your thermal management gets much trickier. All I can figure is that you come to a steady-state (or at least a sinusoid) temperature pretty quickly, and a clever boffin somewhere has figured out how to either (1) ride out the cold bits using heat absorbed during the “day”, or (2) ride out the hot bits by radiating off all of the excess heat at “night”. Of course, for a sun-synchronous satellite, a full orbit takes about 100 minutes, so you’re not really exposed for all that long.
Different orbits have different problems. YMMV.
ftg, a dewar flask has a mirrored lining to prevent radiated heat from escaping from the center (or block it from getting in), it’s not simply the vacuum which keeps the contents close to the temperature which they were at when they were placed inside of it.
astro, you’re absolutely correct about time being the key factor here. I don’t know how fast things are able to radiate heat in space. You maybe dead right that a processor would fry itself in seconds (unless it were equipped with a heat sink the size of the ISS), I do not know. All I do know, is that without sunlight, objects in space will eventually cool down to near absolute zero. I know that NASA has to use aerogels to keep the Mars rovers insulated from the cold Martian nights, and that on the Apollo missions and on the missions to the planets NASA has had to provide heaters to keep the probes warm.
And on preview, I see that Jurph’s posted that at least some electronics do get cold. I don’t know what type of processors are running onboard the ISS or most satellites, but the Hubble is running a 486 (which isn’t nearly the heat demon that a Pentium class or better processor is).
If you are going to the expense of boosting your PC into orbit, and you are concerned about overheating, spend an exta couple of hundred buck and craft the bigest heat sink the computing world has ever seen for the top of your processor.
a 5 Lb. extruded aluminum heat sink measuring oh say 5 feet on a side with fins about 18" long should keep that processor below the melt down point long enough for you to read all the dope forums.
if the heat loss is slower, you just need more heat sink.
I don’t have a cite for this, but I remember reading that the very low power processors used in satellites have conductive heat pipes leading to large black radiators pointing away from the sun to get rid of heat.
Stefan’s law for black body radiation (radiation from a perfect radiator surface)gives us
P = S . A . T^4
Where
P is the power radiated in Watts
S is Stefan’s constant, 5.67 x 10^-8
A is the surface area of the body in m^2
T is the temperature in Kelvin
If we consider a run of the mill processor, it will need to dissipate say 25W in heat energy. If we say the chip is 4cm by 4cm it will have a surface area of 0.0032 m^2. Factor in some kind of radiator surface and lets say A = 0.032 m^2. We can now calulate the equilibrium running temperature
T_equil = (25 / 0.032 / 5.67 x 10^-8)^0.25
T_equil = 342 K
T_equil = 70 C
Our chip would probably be able to run in space if kept out of the sun, but if we wanted to overclock it, we would certainly need heat pipes and the like.
Rick, you’re deluding yourself (tongue in cheek or not). In a vacuum, the size of your fins doesn’t matter; fins exist to cool a heat sink via convection, which requires an external fluid. You need a radiator. Preferably something matte black, with high thermal conductivity, and lots of cross-sectional area facing deep space.
Design details.
Or you could go the easy way, strap a 50 lb block of ice to the top of your P4 chip. Waterproofing will be left as an exercise for the student.
Here’s a web page dealing with spacecraft thermal control: http://www.coolingzone.com/Content/Library/Papers/Sep%201996/Article%2003/Sep%201996_03.html
Actually, Jurph already covered most of it. Insulation, heatpipes, and a diurnal temperature range of -180°C to 150°C.
And at one point they mention radiators with a “heat rejection capability of up to 350W/m^2.” A Pentium 4 CPU generates abvout 70 watts of heat (cite) so it wouldn’t take a huge radiator to balance that out.