Keeping electronics cool in the Antarctic: harder than you think
Power-related design and management challenges often fall into two broad categories: those where the design is severely power-source limited (think energy harvesting/data logger); and those where thermal dissipation and heat buildup are the primary challenge (a large motor drive). Certainly, many designs have a combination of both power sourcing and dissipation constraints to resolve, but the two groupings are quite different, as the first has minimal thermal concerns, while the second is largely about thermal issues.
But there are cases where you assume that cooling will be little or no problem, despite the amount of heat that has to be dissipated - yet you would be wrong. That's why a reader's forum entry - "Keeping equipment cool can be a challenge at Antarctic ground station" - about the Marisat-GOES Terminal (SPMGT) and GOES backup antenna (Figure 1) in the latest issue of Physics Today, about the challenge of keeping electronics cool enough in the Antarctic, was fascinating. That entry also called out a longer previous article - "The care and feeding of an Antarctic telescope" - which had details on another project, the High Elevation Antarctic Terahertz (HEAT) telescope, with a more-complicated power situation and more difficult thermal issues (Figure 2).
My first thought when seeing that provocative first title was along the lines of "Really, what's the problem? Just open a window!" Of course, it's not that easy. You can't let the electronics get too hot or too cool, and there is other equipment in the enclosure that also has to be thermally managed. But the real thermal problem in the Antarctic is that the air is extremely dry, and so loses much of its convection capacity (whether natural-convection unforced or fan-driven forced type).
In the case of the Antarctic designs, the thermal assessment has to take the poor convection capacity of the air into account - and the problem is aggravated by the difference in this parameter's value versus altitude, as the air is much thinner. Thus, a thermal design that is fine at ground level may not work at higher altitudes. The telescope article notes that air at 4000 meters has about half the density it has at sea level - that's quite a difference.
This density-variation problem is not unique to the Antarctic, of course. Aircraft and spacecraft designers are well aware of it, and know that in the vacuum of space, there is no convection cooling; there is only radiation cooling.
But for nominally ground-based systems, the change in convention-cooling potential is something designers have to take into their modeling and design, but may overlook. Fortunately, one of the referenced links in the Physics Today reader forum entry is to a very readable paper which analyzes the situation; see "Forced Air Cooling at High Altitude."
Have you ever had cooling problems where the variation from minimum to maximum for some ambient parameter or operational specification was the biggest challenge, much more so than the nominal situation?