That's interesting. I'm pinning my hopes on either trace heating or oil submersion to avoid condensation, and feel considerably more optimistic about heating. If you'd critique the following I'd appreciate it.
What I have in mind is a small pipe (<=10/8mm) with cold water running down it, then a layer of cladding, then wire heated electrically, and the entire thing stuffed down another pipe. So it's not exactly small in terms of diameter, and not very flexible either, but it's my best idea so far. The cladding isn't needed, but it'll significantly improve overall efficiency. It needs some means of distributing the heat from the copper layer to avoid cold spots, and in truth I haven't checked whether the conductivity of the outer tube is adequate for the task yet.
There's significant problems with insulating a pump & the cpu block too, which I haven't made much progress with. The tubing idea still needs some calculations, but initial results based on Fourier's law in 1D were promising enough that I've put it to one side for now.
