Re: Bifilar coil for electromagnet
Posted: Mon Aug 26, 2013 12:28 am
From the discriptions, using two wires in parellel is meaningless overal. Having multiple windings is essentially the same thing. You can have one thick wire, two thinner wires , three more thinner wires, etc. As you have more wires in parellel (looking at only a short segment of the magnet windings like it was independent). More windings means less total current through all of the windings to get the same strength B field, but it is a near tradoff. Each thinner wire can carry proportionatly less current, with the Ohimic heating being the limiting factor. The same applies to superconductor wires, though the limits are somewhat different. A super conductor wire can only carry a certain amount of current through a certain wire thickness, it is not unlimited. There is no Ohmic heating limit but there is a funtional limit due to local magnetic fields inside the wire. A thick copper wire might carry 10,000 A if you can cool it fast enough. A superconductor wire of the same thickness might be limited to a smaller to mildly larger(?) current. No cooling is needed except of course the maintenance of the initial cryogenic temperature. The high temperature superconductors seem to have less capacity in this regard, though things are evolving.
I mentioned a near tradeoff because as the wire diameter is reduced , the number of wires can increase, but the insulting layer thickness becomes progressively more significant as a ratio of wire thickness/ diameter and insulation thickness. If a wire has a diameter of 1 mm with 0.1 mm thick insulation surrounding it then the conductive component of the wire is 80 % of the overall wire diameter (actually different as cross sectional area needs to be considered, but the trend is consistent). If the overall wire thickness is reduced to 0.3 m then the conductive component is only 33%. There all sorts of engineering tradeoffs in the electromagnets. Incidentally the idea of using Bitter electromagnet designs somewhat like those used in the worlds strongest electromagnets have been kicked around because of the engineering concerns of intimately cooling the wires that are carrying high currents.
Splicing a single wire to a run of more wires, then merging the wires back into one might seem to be a solution to increasing the B field in the nubs between magnets, but when thought through the amp turns and thus B field strength would not change. Winding the magnets with a bundle of say 10 wires and then passing one bundle through the nub to the next magnet might help, but it introduces other concerns. Wiring each magnet completely separately with perhaps a bundle of wires feeding through the standoffs may be a simpler engineering solution with better standoff (nub equivalent) magnetic shielding and better cooling margins.
Concerning tests with WB6 and I presume WB7 and possibly WB8.0, the magnets were pulsed in the sense that they could only be run for at most a few seconds before heat buildup necessitated shut down, but compared to other time frames in the machines, this was extremely long steady state B field conditions. The magnets only needed to be on as the switches were thrown and high voltage current and gas was introduced. In WB6 the gas introduction complications limited the tests to only a few milliseconds plus the delays in the crude switching systems used. The magnet heating was completely irrelevant. Once the magnet windings heated up there would be a needed cool down period which was presumably several hours as the magnets was in a very good vacuum thermos bottle. Active cooling would prolong the electromagnet run times a tiny to moderate amount of time depending on the flow rate, but this was not important for the tests, but only for the recovery time between tests. This generalization may be only partially applicable. WB4, which had active water cooling, was able to generate B fields of ~ 3000 Gauss. I don't know how far WB6 could have been pushed because they were running out of time and of course the electrical short fried at least one of the magnets.
WB8.1 tests where they may have tried to increase run times well beyond the expected ion lifetimes (perhaps 20 milliseconds to hundreds of milliseconds), might benefit from more aggressive cooling of the magnets. But I suspect even here control of vacuum pressures, gas puffers and/or ion guns and electron guns were the more challenging issues.
Cooling the electromagnet wires to liquid nitrogen temperatures may have been mostly for the improved conductivity (6+8 fold improvement) so that higher currents and thus B fields could be maintained for the tests. The cooling may (pure conjecture) have been much less than that needed to keep the electromagnets at near initial temperatures for many second time frames. If various engineering issues with switch timing, etc. can be managed, then even one second of magnet on time is practically forever at this stage of the research. Things change with greatly increased fusion power intensity/ time conditions. Then you not only need to keep the magnets cool for long times against the Ohmic heating loads but also Bremsstruhlung and neutron heating loads. With P-B11 fusion the neutron heating concerns are almost totally eliminated and this may be a big plus for this fuel choice, despite the increased Bremsstruhlung concerns. Superconductors would eliminate the Ohmic heating, but not the others. You will not be able to cool the super conductor to cryogenic temperatures and then need almost no additional cooling, there will have to be major active cooling to keep the magnet temperatures down, even with superconductors.
Dan Tibbets
I mentioned a near tradeoff because as the wire diameter is reduced , the number of wires can increase, but the insulting layer thickness becomes progressively more significant as a ratio of wire thickness/ diameter and insulation thickness. If a wire has a diameter of 1 mm with 0.1 mm thick insulation surrounding it then the conductive component of the wire is 80 % of the overall wire diameter (actually different as cross sectional area needs to be considered, but the trend is consistent). If the overall wire thickness is reduced to 0.3 m then the conductive component is only 33%. There all sorts of engineering tradeoffs in the electromagnets. Incidentally the idea of using Bitter electromagnet designs somewhat like those used in the worlds strongest electromagnets have been kicked around because of the engineering concerns of intimately cooling the wires that are carrying high currents.
Splicing a single wire to a run of more wires, then merging the wires back into one might seem to be a solution to increasing the B field in the nubs between magnets, but when thought through the amp turns and thus B field strength would not change. Winding the magnets with a bundle of say 10 wires and then passing one bundle through the nub to the next magnet might help, but it introduces other concerns. Wiring each magnet completely separately with perhaps a bundle of wires feeding through the standoffs may be a simpler engineering solution with better standoff (nub equivalent) magnetic shielding and better cooling margins.
Concerning tests with WB6 and I presume WB7 and possibly WB8.0, the magnets were pulsed in the sense that they could only be run for at most a few seconds before heat buildup necessitated shut down, but compared to other time frames in the machines, this was extremely long steady state B field conditions. The magnets only needed to be on as the switches were thrown and high voltage current and gas was introduced. In WB6 the gas introduction complications limited the tests to only a few milliseconds plus the delays in the crude switching systems used. The magnet heating was completely irrelevant. Once the magnet windings heated up there would be a needed cool down period which was presumably several hours as the magnets was in a very good vacuum thermos bottle. Active cooling would prolong the electromagnet run times a tiny to moderate amount of time depending on the flow rate, but this was not important for the tests, but only for the recovery time between tests. This generalization may be only partially applicable. WB4, which had active water cooling, was able to generate B fields of ~ 3000 Gauss. I don't know how far WB6 could have been pushed because they were running out of time and of course the electrical short fried at least one of the magnets.
WB8.1 tests where they may have tried to increase run times well beyond the expected ion lifetimes (perhaps 20 milliseconds to hundreds of milliseconds), might benefit from more aggressive cooling of the magnets. But I suspect even here control of vacuum pressures, gas puffers and/or ion guns and electron guns were the more challenging issues.
Cooling the electromagnet wires to liquid nitrogen temperatures may have been mostly for the improved conductivity (6+8 fold improvement) so that higher currents and thus B fields could be maintained for the tests. The cooling may (pure conjecture) have been much less than that needed to keep the electromagnets at near initial temperatures for many second time frames. If various engineering issues with switch timing, etc. can be managed, then even one second of magnet on time is practically forever at this stage of the research. Things change with greatly increased fusion power intensity/ time conditions. Then you not only need to keep the magnets cool for long times against the Ohmic heating loads but also Bremsstruhlung and neutron heating loads. With P-B11 fusion the neutron heating concerns are almost totally eliminated and this may be a big plus for this fuel choice, despite the increased Bremsstruhlung concerns. Superconductors would eliminate the Ohmic heating, but not the others. You will not be able to cool the super conductor to cryogenic temperatures and then need almost no additional cooling, there will have to be major active cooling to keep the magnet temperatures down, even with superconductors.
Dan Tibbets