>Their conclusion is that a “magnetic hose” consisting of concentric tubes of superconducting and ferromagnetic materials ought to do the trick. They say that a

>Their conclusion is that a “magnetic hose” consisting of concentric tubes of superconducting and ferromagnetic materials ought to do the trick. They say that a tube consisting of 20 concentric rings that is about ten times longer than it is wide, should transmit about 90 per cent of a magnetic field at one end to the other. Indeed, a tube of just 2 concentric rings should transmit about 75 per cent. "About 75 percent" is a stretch, at least if the article author and I are looking at the same paper. According to the actual paper, in figure 2e, it looks closer to 68 or 69 percent. Not a huge nit pick though. >These guys have tested this idea with a single superconducting tube 7 cm long (made of BiPbSrCaCuO) and filled with a ferromagnetic alloy (of cobalt and iron). How much does a superconducting tube of BiPbSrCaCuO that is 10 times as long as it is wide cost to manufacture? How much does it cost to keep it cool enough to superconduct? It's billed as a high Tc (108 K) superconductor from the manufacturer (Can), but there is no price list to be found. Figure 4a in the paper seems to indicate that this effect is indeed only observed with a superconducting tube with a ferromagnet core. Regardless, it's still a cool effect to see in action! If there's an actual physicist in the room, I'd love to hear more about Chapter four in the paper. It seems that the discretization to more than just a single Superconductor-Feromagnet pair --figure 2e in the paper is the main graph showing the efficiency gains as the number of pairs goes up-- was done using FET, after a brief justification argument that goes over my head, and was not actually produced and measured in lab. I think it'd be a logical extension (if not an expensive one) to produce some of these higher-paired tubes and take more measurements for another paper.