Slashdot Mirror
New Chip-cooling Technology
BillOfThePecosKind writes "Researchers have demonstrated a new technology using tiny "ionic wind engines" that might dramatically improve computer chip cooling, possibly addressing a looming threat to future advances in computers and electronics. Purdue researchers funded by Intel have improved the "heat-transfer coefficient" by some 250%. I never liked water cooled systems, and this sounds promising. However I wonder how much ozone one of these things produces."
Cool!
The problem is the power consumption on this thing. If you assume that they want to move all the air in a small region around the wire even once per second, say 10mm x 1mm x 1cm, to use the dimension quoted in TFA and nominal orders of magnitude for chip size and wire thickness, that corresponds to something ~ 10^-5 moles of air. Since Nitrogen has an ionization energy of 1402.3 kJ/mol (Wikipedia), that means if you want to move that quantity of air every second, you need at least something around 15W. That's even assuming you perfectly convert electrical energy into removing electrons from air molecules, and it's just to ionize the air, neglecting the extra energy it then takes to get the ions moving (we'll pretend the fan does all that, even though that would mean that our device isn't doing jack).
I don't know how much energy my laptop uses, but my power adapter is 65W, so 15 seems non-negligible.
You are quite right. The AC has no idea what he is talking about. If only his grasp of "simple heat transfer" matched his arrogance. This is not a sealed chamber. The ions impart momentum to a near wall flow and destroy the boundary layer. Good mixing at the wall = good heat transfer! (The article says as much) These Purdue dudes have a lot of neat electronics cooling stuff going on. I had the pleasure of getting the whole delivery at a seminar last Fall.